DOI: 10.1021/la8042796 8663Langmuir 2009, 25(15), 8663–8671 Published on Web 04/24/2009

pubs.acs.org/Langmuir

© 2009 American Chemical Society

Self-Assembly of Esters of Arjunolic Acid into Fibrous Networks

and the Properties of their Organogels†

Braja G. Bag,*,‡ Shaishab K. Dinda,‡ Partha P. Dey,‡ V. Ajay Mallia,§ and Richard G. Weiss*,§

‡Department of Chemistry and Chemical Technology, Vidyasagar University, Midnapore 721 102, WB, India,and §Department of Chemistry, Georgetown University, Washington, D.C. 20057-1227

Received December 30, 2008. Revised Manuscript Received March 2, 2009

Nine esters of a naturally occurring triterpenoid, arjunolic acid (from Terminalia arjuna), with alkyl chains have beensynthesized, and their self-assembly has been studied in organic liquids. All of the esters examined were found to beexcellent gelators. No birefringence was detected in optical micrographs of the transparent toluene gels with 5% (w/w)ethyl arjunolate or 5% (w/w) p-nitrobenzyl arjunolate as the gelator, but a spherulitic-type pattern was seen for a gel of1.2% (w/w) p-nitrobenzyl arjunolate in 1/1 (w/w) chloroform/cyclohexane. Electron microscope images revealed self-assembled fibrillar network (SAFIN) structures with right-handed helical ribbons in some gels. With increasingconcentration of the gelators, the gel-to-sol transition temperature (Tgel) increased and then approached plateau values.Differential scanning thermograms demonstrated that the heats for transition from transparent gels to sols of ethylarjunolate or p-nitrobenzyl arjunolate in toluene are very small. Powder X-ray diffractograms revealed that themolecular packing in the SAFIN of the 5% (w/w) ethyl arjunoate in the toluene gel was amorphous and similar to thediffractogram recorded for the neat gelator. Although the diffractogram of neat p-nitrobenzyl arjunolate consisted ofbroad peaks, suggesting disordered packing, the low-angle peaks of the corresponding toluene gel were much sharper;these results indicate more crystalline packing in the SAFIN than in the neat gelator. The kinetics and growth of thetransformation of sols of p-nitrobenzyl arjunolate in 1/1 (w/w) chloroform/cyclohexane to their gels have beeninvestigated at different incubation temperatures by circular dichroism spectroscopy. The data have been analyzed toprobe the mechanism of SAFIN formation and the relationship between the molecular structures of the esters ofarjunolic acid and their abilities to function as gelators of a wide variety of organic liquids.

Introduction

The utilization of plant metabolites as renewable resources hasbecome a significant area of investigation in recent years becauseit is an approach to the development of sustainable chemical feed-stocks.1,2 Triterpenes, the C30 subset of the major plant secondarymetabolite terpenes, are promising in this respect because of theirabundant skeletal types and included functional groups.3 More-over, the extended lengths of triterpenoids, whether acyclic, mono-cyclic, or rigidpentacyclic structures, are in the nanometer range.Assuch, theymay be useful building blocks for nanoscience and nano-technology applications.4 A major problem in exploiting triterpe-noid compounds has been the difficulty in obtaining them in pureform. We have recently reported a simple and efficient method forthe isolation and purification of a 6,6,6,6,6-pentacyclic triterpenoid,arjunolic acid (1), from Terminalia arjuna5 and discovered that anarjunolic acid-derived molecule forms thermochromic organogels.6

The location of the hydroxyl and carboxyl functional groups of 1, atopposite ends of the triterpenoid framework, constitute awonderfulframework on which to build a diverse set of potential low-

molecular-mass organic gelators (LMOGs) and to examine in detailthe relationshipsbetweenmolecular structures andgelatingabilities.

By esterifying the carboxyl end of 1, it is possible tomake a seriesof derivatives that are less polar than the parent acid, but eachretains three differing structural/polarity regions. Thus, we felt thatit should be possible to modify systematically both the interactionsamong the groups that are separated by nanometric distances andthe specific aggregation modes in this way. Herein, we report thesyntheses and gelation abilities of very low concentrations of 8n-alkyl esters (2-9) and the (p-nitrophenyl)methyl ester (10) of 1(Scheme1) in awidevariety of organic liquids.Thesemolecules con-stitute a class of LMOGs analogous to the aromatic-linker-steroid(ALS) molecules that have been examined extensively and employcholesterol frequently as the steroidal part,7 but the structures of thederivatives of1are “truncated” in that theyneednot contain anaro-matic moiety (in all examples except one); they resemble more theless-studiedLS-type molecules.7a They offer interesting possibilitiesto study how primarily steroidal-like species are able to gel organicliquids; there are relatively few such studies in the literature.8 In fact,

† Part of the Molecular and Polymer Gels; Materials with Self-AssembledFibrillar Networks special issue.*Corresponding authors. E-mail: bgopalbag@yahoo.co.in, weissr@

georgetown.edu.(1) (a) Anastas, P. T.; Kirchhoff, M. M. Acc. Chem. Res. 2002, 35, 686. (b)

Sheldon, R. A. Green Chem. 2008, 10, 359. (c) Anastas, P. T.; Warner, J. C. GreenChemistry: Theory and Practice; Oxford University Press: New York, 1998; p 30.(2) Bag, B. G.; Dinda, S. K. Pure Appl. Chem. 2007, 79, 2031.(3) Shibuya, M.; Xiang, T.; Katsube, Y.; Otsuka, M.; Zhang, H.; Ebizuka, Y. J.

Am. Chem. Soc. 2007, 129, 1450.(4) Bag, B. G.; Garai, C. Unpublished results.(5) Bag, B. G.; Dey, P. P.; Dinda, S..K.; Sheldrick, W. S.; Oppel, I. M. Beilstein

J. Org. Chem. 2008, 4, 24.(6) (a) Bag, B. G.;Maity, G. C.; Dinda, S. K.Org. Lett. 2006, 8, 5457. (b) Bag, B.

G.; Maity, G. C.; Pramanik, S. R. Supramol. Chem. 2005, 17, 383. (c) Bag, B. G.;Maity, G. C.; Pramanik, S. R. Pramana 2005, 925.

(7) (a) George, M.; Weiss, R. G. Acc. Chem. Res. 2006, 39, 489. (b) Huang, X.;Terech, P.; Raghavan, S. R.; Weiss, R. G. J. Am. Chem. Soc. 2005, 127, 4336. (c)George,M.;Weiss,R.G. J.Am.Chem.Soc. 2001, 123, 10393. (d)George,M.; Snyder,S. L.; Terech, P.; Glinka, C. J.; Weiss, R. G. J. Am. Chem. Soc. 2003, 125, 10275.(e)Weiss, R. G., Terech, P., Eds. Molecular Gels: Materials with Self-AssembledFibrillar Networks; Springer: New York, 2006. (f ) Terech, P.; Weiss, R. G. Chem. Rev.1997, 97, 3133. (g) Abdallah, D. J.; Weiss, R. G. Adv. Mater. 2000, 12, 1237.

(8) (a)Technical Bulletin, Nikko Chemicals Co., Ltd., 1-4-8 Nihonbashi-Bakur-ocho, Japan, 1989. (b) Martin-Borret, O.; Ramasseul, R.; Rassat, A. Bull. Soc. Chim.Fr. 1979, 401. (c) Hishikawa, Y.; Sada, K.; Watanabe, R.; Miyata, M.; Hanabusa, K.Chem. Lett. 1998, 795. (d) Valkonen, A.; Lahtinen, M.; Virtanen, E.; Kaikkonen, S.;Kolehmainen, E. Biosens. Bioelectron. 2004, 20, 1233. (e) Nakano, K.; Hishikawa, Y.;Sada, K.; Miyata, M.; Hanabusa, K.Chem. Lett. 2000, 29, 1170. (f ) Lawrence, A. S. C.In Liquid Crystals and Ordered Fluids; Johnson, J. F., Porter, R. S., Eds.; Proceedingsof an American Chemical Society Symposium on Ordered Fluids and Liquid Crystals,September 10-12, 1969, New York; Plenum Press: New York, 1970, pp 289-291. (g)Terech, P. Mol. Cryst. Liq. Cryst. 1989, 166, 29.

8664 DOI: 10.1021/la8042796 Langmuir 2009, 25(15), 8663–8671

Article Bag et al.

the distribution of functional groups about the fused cyclic frame-work of 1 is more similar to that of cholic acid in that both containone carboxylic and three hydroxyl groups. However, the lipophilicplatform of 1 is longer than that of cholic acid (or a steroid), onlyone of the three hydroxyls around the fused ring part of cholic acidis located at the end opposite its carboxyl group, and onlysubstituted cholic acid and other bile salts with fewer hydroxylgroups are reported to form hydrogels.9 In general, gels formed bythe self-assembly of small concentrations of LMOGs in a liquidhave attracted considerable attention in recent years by us7 andothers10 to gain a better understanding of self-assembly processes indifferent media and because of their potential applications innanomaterials, sensor devices, cosmetics, nutrition, pharmacology,and so forth. A wide variety of techniques has been used tocharacterize the structures and kinetics of formation of the self-assembled fibrillar networks (SAFINs) from sol phases of 2-10 andthe thermal and temporal stabilities of their gels over awide rangeofLMOG concentrations and liquid types.

Experimental Section

Methods. A Leitz 585 SM-LUX-POL microscope equippedwith crossed polarizers, aLeitz 350 heating stage, a PhotometricsCCD camera interfaced to a computer, and an Omega HH503microprocessor thermometer connected to a J-K-T thermocou-ple has been used to record optical micrographs (OMs) ofsamples flame-sealed in 0.5 mm-pathlength, flattened, Pyrexcapillary tubes (VitroCom). UV-vis absorption spectra wereobtained with a Cary 300 BIOUV-vis spectrophotometer using0.1 mm quartz cells. CD spectra were recorded on samples inwater-jacketed quartz CD cells (Hellma Inc., 0.1 mm pathlength) using a Jasco-700 CD spectrometer. The temperaturewas maintained by a thermostatted, circulating bath. Differen-tial scanning calorimetry (DSC) was performed on a DSC Q200differential scanning calorimeter (TA Instruments) connectedto a refrigerated cooling system (TA Instruments) under a50 mL/min stream of nitrogen flowing through the cell. Sampleswere in hermetically sealed aluminum pans (Tzero hermetic lidand Tzero pan). Heating and cooling rates were 10 �C/min.Powder X-ray diffraction (XRD) patterns of samples wereobtained on a Rigaku R-AXIS image-plate system with CuKR X-rays (λ = 1.54 A) generated by a Rigaku generator

operating at 46 kV and 40 mA with the collimator at 0.5 mm(to obtain 0.5-mm-diameter beams of X-rays11). Data proces-sing and analyses were performed using Materials Data JADE(version 5.0.35) XRD pattern-processing software. Sampleswere sealed in 0.5 mm glass capillaries (W. M

::uller, Sch

::onwalde,

Germany), and diffraction data were collected for 1 (neatpowders) or 10 h (gels). Gel melting temperatures (Tgel) weredetermined by the temperature at which a gel fell under theinfluence of gravity when inverted in a sealed glass tube that wasplaced in a water bath that was heated from room temperatureat ca. 1.5 �C/min.12 Samples for SEM were prepared by placinga thin aliquot of a gel on a glass microscope coverslip and dryingit at room temperature in the air for 3 to 4 days and then underreduced pressure for several hours. The resulting xerogels weresputter-coated with platinum before FESEM analyses in a JeolJSM-6700F field emission scanning electron microscope. Trans-mission electronmicroscopy images of the gels were recorded onsamples of gel that were dried (at ambient temperature in the airfor 24 h and then under reduced pressure for 24 h) on 300-meshcarbon grids.

Syntheses of Esters with Alkyl Chains.Arjunolic acid usedfor the syntheses of the alkyl-chained esters and gelation testswas extracted from Terminalia arjuna and purified following theprocedure as described previously.5 All of the solvents usedfor purification purposes were commercial-grade solvents andwere distilled before use. The liquids used for gelation studiesand DBU and DMF were laboratory-grade reagents and weredry distilled using standard literature methods before use. All ofthe alkyl tosylates used for the esterifications were synthesizedfrom the corresponding alcohols by tosylation (pTsCl/pyridine/5 �C). Methyl arjunolate 2 was prepared from arjunolic acid onreaction with diazomethane as reported earlier.13 Two generalprocedures (A andB) for synthesizing 3-10 are described below.The characterization data for these gelators are collected in theSupporting Information.

General Procedure (A) for the Synthesis of Esters 8 and 9:Synthesis of n-hexadecylArjunolate (8). Amixture of crystal-line arjunolic acid5 (0.200 g, 0.409 mmol) and n-hexadecyltosylate (0.195 g, 0.491 mmol) contained in a dry 5 mL round-bottomed flask was suspended in dry benzene (0.80 mL) andtreated with 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) (0.075mL, 0.491 mmol). The mixture was refluxed for 4 h withcontinuous stirring. The reaction mixture was treated with anadditional amount of n-hexadecyl tosylate (0.03 g, 0.076 mmol)andDBU (0.015mL, 0.098mmol), and the reactionmixture wasrefluxed for another 2 h. The volatiles were removed underreduced pressure, and the crude product was purified by columnchromatography (Si-gel, 100-200 mesh, 1.0 � 12.5 cm) using70% ethyl acetate/petroleum ether as the eluant. The productwas obtained as a foamy solid (0.21 g, 72% yield), mp 48-51 �C.1H NMR (600 MHz, CDCl3) δ: 5.28 (app.t, 1H), 4.03-3.97 (m,-CO-O-CH2-, 2H), 3.75 (m, 1H), 3.65 (d, J= 10.8 Hz, 1H),3.56 (s,-OH, 1H), 3.42 (t, 2H), 2.87 (dd, J1= 13.8 Hz, J2= 4.2Hz, 1H), 2.70 (d, -OH, 2H), 1.99-0.86 (m, terpenoid protons,20H, and long-alkyl-chain protons, 31H), 1.13 (s, 3H), 1.01z(s, 3H), 0.93 (s, 3H), 0.90 (s, 3H), 0.86 (s, 3H), 0.73 (s, 3H). 13CNMR (150MHz, CDCl3) δ: 177.81, 143.87, 122.03, 80.32, 70.31,68.69, 64.29, 48.99, 47.50, 46.60, 46.00, 45.83, 42.43, 41.72,41.21, 39.31, 38.17, 33.86, 33.10, 32.44, 32.33, 31.91, 30.69,29.70, 29.67, 29.65, 29.59, 29.35, 29.19, 28.58, 27.56, 26.05,25.93, 23.59, 23.43, 22.93, 22.68, 18.31, 17.05, 16.99, 14.12,12.75. FTIR: vmax (neat, cm-1) 3378 (s), 2925 (s), 2854 (s),1725 (s), 1464 (s), 1387 (m), 1260 (w), 1175 (w), 1048 (s). [R]D298

+43.0� (c 0.166, CHCl3). HRMS (ESI): m/z calcd(C46H81O5) 713.6084, found 713.6096 [M + H]+.

Scheme 1. Arjunolic Acid 1 and Its Esters 2-10

(9) Bhattacharya, , S.; Maitra, , U.; Mukhopadhyay, S.; Srivastava, , A. InMolecular Gels: Materials with Self-Assembled Fibrillar Networks; Weiss, R. G.,Terech, P., Eds.; Springer: New York, 2006; Chapter 17.(10) (a) Ajayaghosh, A.; Praveen, V. K. Acc. Chem. Res. 2007, 40, 644. (b) van

Esch, J. H.; Feringa, B. L. Angew. Chem., Int. Ed. 2000, 39, 263. (c) Friggeri, A.;Gronwald, O.; van Bommel, K. J. C.; Shinkai, S.; Reinhoudt, D. N. J. Am. Chem.Soc. 2002, 124, 10754. (d) Messmore, B. W.; Hulvat, J. F.; Sone, E. D.; Stupp, S. I.Am. Chem. Soc. 2004, 126, 14452. (e) Jung, J. H.; John, G.; Masuda, M.; Yoshida,K.; Shinkai, S.; Shimizu, T. Langmuir 2001, 17, 7229. (f) Fujita, N.; Shinkai, S., InMolecular Gels: Materials with Self-Assembled Fibrillar Networks; Weiss, R. G.,Terech, P., Eds.; Springer: New York, 2006; Chapter 15.

(11) Alexander, L. E. X-ray Diffraction Method in Polymer Science; JohnWiley& Sons: New York, 1969, pp 102ff.

(12) Takahashi, A.; Sakai, M.; Kato, T. Polym. J. 1980, 12, 335.(13) Bag, B. G.; Pramanik, S. R.; Maity, G. C. Supramol. Chem. 2005, 17, 297.

DOI: 10.1021/la8042796 8665Langmuir 2009, 25(15), 8663–8671

Bag et al. Article

General Procedure (B) for the Syntheis of Esters 3-7 and

10: Synthesis of (4-Nitrophenyl)methyl Arjunolate (10). Amixture of crystalline arjunolic acid5 (0.101 g, 0.208 mmol) and4-nitrobenzyl bromide (0.055 g, 0.258 mmol) contained in adry 5 mL round-bottomed flask was suspended in dry DMF(0.10 mL) and treated with DBU (0.040 mL, 0.267 mmol). Themixture was heated over an oil bath at 60 �C for 24 h withcontinuous magnetic stirring. The reaction mixture was treatedwith an additional amount of 4-nitrobenzyl bromide (0.056 g,0.258 mmol), and the reaction mixture was heated for another7 h. The volatiles were removed under reduced pressure, and thecrude product was purified by column chromatography (Si-gel,100-200 mesh, 1.0 � 12.5 cm) using 50% ethyl acetate/chloro-form as the eluant. The product was obtained as a white solid(0.109 g, 84% yield), mp 226 �C. 1H NMR (600 MHz, CDCl3)δ: 8.22 (d, J=9.0 Hz, 2H), 7.51 (d, J=9.0 Hz, 2H), 5.30 (t, 3.6Hz, 1H, 12-H), 5.19 (d, J = 13.8 Hz, 1H, Ar-CHa-), 5.11 (d,J= 13.8 Hz, 1H, Ar-CHb-), 3.81-3.69 (m, 1H, 2-H), 3.60 (d,J = 10.8 Hz, 1H), 3.39 (d, J = 10.2 Hz, 1H), 3.38 (d, J = 10.8Hz, 1H), 2.89 (dd, J1 = 13.8 Hz, J2 = 4.2 Hz, 1H, 18-H), 2.07-0.85 (m, terpenoid H’s, 20H), 1.13 (s, 3H), 0.96 (s, 3H), 0.93 (S,3H), 0.91 (s, 3H), 0.81 (s, 3H), 0.54 (s, 3H). 13CNMR (150MHz,CDCl3) δ: 177.18, 147.55, 143.64, 128.41, 123.70, 122.35, 79.63,69.31, 68.63, 64.53, 48.52, 47.37, 46.80, 46.00, 45.70, 42.54,41.72, 41.34, 39.25, 38.06, 33.72, 33.02, 32.37, 32.21, 30.66,27.49, 25.95, 23.57, 23.41, 23.02, 18.17, 16.94, 16.86, 12.85.FTIR: νmax (neat, cm-1) 3387 (s), 2946 (s), 2880 (s), 1726 (s),1666 (w), 1606 (w), 1524 (s), 1460 (m), 1454 (m), 1388 (m), 1364(m), 1347 (s). UV (MeOH, log ε) λmax 266 nm (4.02). [R]D298

+33.89� (c 0.226, MeOH). HRMS (ESI): m/z calcd(C37H53NO7Na) 646.37197, found 646.3750 [M + Na]+.

Results and Discussion

Syntheses of compounds 3-10were accomplished in 70-90%isolated yields by heating 1 with alkyl halides or alkyl tosylates inbenzene or DMF in the presence of DBU as a base.

Gelation tests consisted of allowing a hot solution/sol of 5 mgof gelator and 0.5 mL of liquid in a closed vial (capacity 4 mL;1 cm (id)) to cool to room temperature in the air and thenobserving whether the sample flowed when the vial was inverted.Among the 21 liquids tested, gels were obtained in 19; those inwhich no gelation was observed in 2-10 were methanol andethanol, and arjunolic acid did not gel any of the liquids, it

remained insoluble in most of them at elevated temperatures.Methyl arjunolate (2), having the shortest alkyl chain, gelatedonly five liquids; it was insoluble in most of the others (Table 1).Ethyl and n-butyl arjunolate (3 and 4) gelated most of the liquids,but with increasing chain length and solubility, gelators 5-9 didnot precipitate as their sols cooled and gels were obtained mostlyin alkane liquids. 10 was an excellent gelator of various liquids,especially those that are aromatic.

Among the aromatic liquids tested, benzene could be gelatedonly by ethyl arjunolate 3 and (4-nitrophenyl)methyl arjunolate10 (Table 1), but toluene could be gelated by methyl arjunolate 2,ethyl arjunolate 3, n-butyl arjunolate 4, and (4-nitrophenyl)methyl arjunolate 10. Gelator 2 had a comparatively highercritical gelation concentration (cgc; the minimum concentrationof gelator necessary to obtain a gel in a liquid at room tempera-ture) in toluene than did 3, 4, or 10, the LMOGs with longer ormore complex ester groups. n-Decane was gelated by 7 of the 10gelators; the cgc values of the esters having comparatively shorteralkyl chain lengths, 3, 4, 5, 6, and 7, were lower than the LMOGshaving longer ones, 8 and 9. The period necessary for gels to formwas usually 2-4 h at room temperature when the gelatorconsisted of one of the shorter-chained esters (N.B., methyl,ethyl, and butyl) and 10; the esters with longer chains (N.B.,hexadecyl and octadecyl) required 2-4 days. Once formed, thegels were stable for several weeks, at least when stored at roomtemperature in closed containers.Tgel as a Function of 10 Concentration. The gel-to-sol

melting temperature, Tgel, was plotted versus w/v % of 10

to determine the thermal stability of the gels.10 The increase inTgel with increasing gelator concentration indicates that self-assembly is driven by strong intermolecular interactions. Theword “strong” is used here in a qualitative sense to indicatethat gelator-gelator pair interactions are at least several kThigher than for liquid-gelator pair interactions at temperaturesbelow Tgel.

In most cases, the gels obtained in o-, m-, and p-xylenes havehigher Tgel values than the corresponding gel in benzene(Figure 1A). Although the Tgel in benzene and m-xylene aresame at 0.83 w/v %, the melting temperature increases muchmore rapidly with increasing concentration of 10 in them-xylene

Table 1. Gelation Tests with 1-10 in a Variety of Liquidsa

liquid 1 2 3 4 5 6 7 8 9 10

benzene I FC G/5 S S S S S S G/1toluene I G/5b G/0.4b G/1.3b S S S S S G/0.5b

o-xylene I G/4 G/3 G/7 G/10 S S S S G/2m-xylene I G/10 G/3 G/9 S S S S S G/1p-xylene I G/10 G/6 G/9 S S S S S G/1mesitylene I S G/3 G/9 S S S S S G/1chlorobenzene I S G/3 G/9 S S S S S G/3bromobenzene I S G/3 G/8 S S S S S G/3nitrobenzene I G/10 G/3 G/8 G/10 G/5 G/5 S S G/10b

o-dichlorobenzene I S G/10 G/9 S S S S S G/2n-hexane I I I G/2 G/0.5 G/1 PG/3 S S In-heptane I I I PG/2 G/2 PG/3 G/5 PG/5 G/5 In-octane I I G/0.2b G/2 G/2 PG/2 G/5 PG/6 G/6 In-decane I I G/0.2b G/0.8b G/0.4b G/3b G/1.7b G/6b G/5b In-dodecane I I G/0.2b PG/2 G/2 G/3b G/5 G/6 G/6 PG/1c

cyclohexane I I VL G/3 G/6 S S S S G/1c

methylcyclohexane I S G/0.2b G/2 G/2 S S S S G/1c

CH3CN I P/5 C G/6 G/5 P/1.7 P/1.7 S S P/1.6CCl4 I S PG/0.7 S S S S S S G/0.2methanol S S S S S S C S S Sethanol S S C S S S S S S S

aGelation tests were carried out with a maximum concentration of 10% w/v. G= gel, PG= partial gel, VL= viscous liquid, FC= fibrous crystal,C= crystal, S= solution, I= insoluble, and P=precipitate. bCritical gelator concentration (cgc). cGels were prepared in amixture of chloroform andalkanes to overcome insolubility problems; see Table S2 in Supporting Information for details.

8666 DOI: 10.1021/la8042796 Langmuir 2009, 25(15), 8663–8671

Article Bag et al.

gel than in the benzene gel. A steeper increase is indicative of a

stronger dependence of the solubility of 10 on temperature.

Additional curves in Figure 1A-C demonstrate further the

dependence of Tgel on liquid structure when 10 is the LMOG.

Figure 1D explores the dependence of Tgel on the length of the

alkyl chains of theLMOGs of 2 and 4 in o-xylene; theTgel values

of 4 are higher than those of 2 in the same solvent. A steeper

increase in Tgel values with concentration was observed for 6 in

n-decane.Self-Assembled Fibrillar Networks by Transmission

Electron Microscopy (TEM). The morphologies of the xero-gels prepared by slowly evaporating the liquid on a grid at roomtemperature were examined by TEM. Right-handed helical rib-bons of 16-32nmwidthwere observed (Figure 2a,b) froma gel ofmethyl arjunolate 2 in o-xylene (5% w/v). Helical ribbons werealso observed in a gel of 10 in toluene (2.1%w/v) (Figure 2d), butno uniformity was observed in their diameters. A fibrous networkwas observed in a gel of ethyl arjunolate 3 in methylcyclohexane(Figure 2c). In each of these, the fiber width is at least 10 timeslarger than the extended molecular length of the constituentmolecules. These results suggest that the fibers may be bundlesofmuch thinner ribbons;14 the basic unit of the gelator aggregatescannot be discerned from the resolution of the TEMs. CD spectraobtained for a gel containing compound 10 at different tempera-tures and after different incubation periods (as discussed in detailin the kinetics of gelation section) also support the supramole-cular chirality observed by the TEM images.SAFIN Structures by Scanning Electron Microscopy

(SEM). Fibrillar network structures having intertwined fibersof 40-50 nm diameter, and micrometer lengths were obtainedfrom a gel of methyl arjunolate 2 in o-xylene (4% w/v)(Figure 3a,b). Virtually the same-shape fibrillar network wasobserved from a gel of n-butyl arjunolate 4 in o-xylene (6.6%w/v). The diameters of the fibers from gels of 2.5% (w/v) 10depended on the liquid: 30-40 nm in chlorobenzene (Figure 3d),50-150 nm in bromobenzene (Figure 3e), and 100-300 nm intoluene (Figure 3f ); the lengths of the fibers in all cases were inthe micrometer range. As mentioned above, the cross sections ofeach of these fibers must contain many molecules of gelator.

Differential Scanning Calorimetric Studies. At 24 �C, the1-5% (w/w) gels of 3 in toluene were transparent in appearance.TheirDSC thermograms for the second heating cycles are shownin Supporting Information, Figure S1. Sol-gel transitions arecontinuous in the sense that they occur in many cases overrelatively broad temperature ranges. Thus, to measure the heatsinvolved using DSC, the areas under the peaks must be inte-grated over the whole temperature ranges. Although that hasbeen done here, it was not possible to calculate with reasonableprecision the heats of transition of these gels because the heatchanges were very small and occurred over broad temperatureranges, necessitating rather subjective measurements of peakareas. On a molecular (but not rheological) basis, the gel-to-soltransition corresponds to two important processes, fiber melt-ing and dissolution of the fiber molecules into the liquidcomponent. Because these two processes occur in tandem, it isfrequently difficult to separate the contribution of each to theheat measured even when the heat of melting of the neat gelatoris known. For these reasons, only the temperatures of themaximum heat (Tm) are reported in Table 2.

The mole fraction of 3 in each of the toluene gels was plottedversus Tm

-1 according to the Schr::oder-van Laar equation (eq 1;

Figure 4).15 ΔH and Tm are the enthalpies of melting and thetransition temperature of the gel-to-sol tranition, respectively, asmeasured by heating endotherms. In this analysis, it is assumedthat the clear gels melt to an ideal solution and that a knownamount of gelator is involved in the transition. The firstassumption is not strictly correct but is probably approximatelyso; the second is reasonably correct because the amounts ofgelator in the SAFINs have been corrected for the amountdissolved in the toluene liquid (i.e., the critical gealtor con-centration, cgc). From the plot, the enthalpy is calculated to be190 ( 54 kJ/mol.

Figure S2 in the Supporting Information shows the DSCthermogram for neat 3 during its first and second heating andcooling cycles. The DSC thermogram of neat 3 exhibits asecond-order transition at 95 �C on cooling, and the subsequentheating curve confirms the formation of a glassy phase. Severalpeaks (endotherm at 84.4 �C, ΔH = 1.5 kJ/mol; exotherm at

Figure 1. Plots of Tgel vs w/v % of 10 in different liquids: (A) p-xylene (-0-), m-xylene (-O-), o-xylene (-4-), and benzene (-3-); (B)bromobenzene (-0-), chlorobenzene (-o-), and dichlorobenzene (-4-); and (C) benzene (-0-), toluene (-O-), andmesitylene (-4-). (D) Plots ofTgel vs concentration for methyl arjunolate 2 in o-xylene, n-butyl arjunolate 4 in o-xylene, and n-decyl arjunolate 6 in n-decane.

(14) Aggeli, A.; Nyrkova, I. A.; Bell, M.; Harding, R.; Carrick, L.; McLeish, T.C. B.; Semenov, A. N.; Boden, N. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 11857.

(15) (a) Atkins, P. W. Physical Chemistry, 5th ed.; Freeman, New York, 1994, p227. (b) Moore, W. J. Physical Chemistry, 4th ed.; Prentice Hall: Englewood Cliffs, NJ,1972; p 249.

DOI: 10.1021/la8042796 8667Langmuir 2009, 25(15), 8663–8671

Bag et al. Article

140.5 �C, ΔH = 19.5 kJ/mol) observed during the first heatingonly may be due to melting of a different solid morph and theconversion of glassy to crystalline material catalyzed by thepresence of some crystal in the material obtained by precipita-tion from ethyl acetate. Clearly, these results and those fromXRD (vide infra) indicate that 3 is not a very crystalline sub-stance as handled by us. Finally, 3 became an isotropic melt(from optical microscopy observations; vide infra) at 179.3 �C(ΔH = 1.8 kJ/mol). Surprisingly, the enthalpy of transition forthe gel-to-sol transition calculated for the toluene gel of 3 fromthe Schr

::oder-van Laar plot is much larger than the heat

measured by DSC for neat 3. There appears to be a precedencefor such an observation although the reasons are still notcompletely understood.16

ln½gelator� ¼ ΔH=RTm ð1ÞGels of 10 in toluene (1-5%w/w) were transparent. Attempts tocalculate their gel melting temperatures using DSC were un-successful because the low heats of transition did not yield largeendotherms that could be distinguished easily from the base-lines. Supporting Information Figures S3 and S4 show DSCthermograms for the first heating and cooling cycles of a gel of5.1% (w/w) of 10 in toluene and that of neat 10.X-ray Diffraction (XRD) Studies. We have compared the

XRD diffractograms of neat powders of 3 and 10 as well as their∼5% (w/w) gels in toluene. The diffraction peaks of the gels were

identified by subtracting the “amorphous” scatter of the liquidtoluene from the total gel diffractogram.17 Figure 5 show theX-ray diffractograms obtained for neat 3 and a 5.1% (w/w) 3in toluene gel. The calculated length of a conformationallyextended molecule of 3 is 16.7 A (Supporting Information,Figure S5).18 The absence of sharp peaks in the diffractogramssuggests that the molecules of 3 both within the neat matrix andin the SAFINs of the gel are highly disordered (Figure 5). Thedata are not definitive, but they suggest that the same morph ispresent in the SAFINs of the gel and in the neat powder. Thatconclusion is supported by the nearly identical IR spectra of a4.5 wt % 3 in toluene gel and of neat 3. If subtle differences arepresent in the O-H stretching region, they must be masked hereby the vicinal nature of the three hydroxyl groups. In other gels,IR spectroscopy has been shown to be a very useful tool indeciphering the mode of aggregation of gelator molecules.19

The broad peaks in the diffractogram of neat 10 suggest itsdisordered packing aswell (Figure 6). The low-angle peaks of thecorresponding toluene gel are much sharper and contain aprogression indicative of better ordering in the SAFINs thanin neat 10. A comparison of lattice spacings (d, A) of 3 and 10 intheir neat powders and toluene gels (from XRD data at 24 �C)and calculated extendedmolecular lengths (L, A)18 are summar-ized in Table 3. The calculated molecular length in the fully

Figure 2. Transmission electron micrographs of xerogels of (a, b) methyl arjunolate 2 in o-xylene (5% w/v), (c) ethyl arjunolate 3 inmethylcyclohexane (0.2% w/v), and (d) 10 in toluene (2.1% w/v).

(16) Murata, K.; Aoki, M.; Suzuki, T.; Harada, T.; Kawabata, H.; Komori, T.;Ohseto, F.; Ueda, K.; Shinkai, S. J. Am. Chem. Soc. 1994, 116, 6664.

(17) Ostuni, E.; Kamaras, P.; Weiss, R. G. Angew. Chem., Int. Ed. 1996, 35,1324.

(18) Using Chem 3 D Ultra 8 software (Cambridge Soft Corporation) andadding the van der Waals radii of the terminal atoms according to Bondi, A. J.Phys. Chem. 1964, 68, 441.

(19) Pal, A.; Ghosh, Y. K.; Bhattacharya, S. Tetrahedron 2007, 63, 7334.

8668 DOI: 10.1021/la8042796 Langmuir 2009, 25(15), 8663–8671

Article Bag et al.

extended conformation of 10 is 17.2 A (Supporting Information,Figure S5). The low-angle reflection in the diffractogram of neat10 corresponding to a Bragg distance of 15.8 A is consistent with

Figure 3. SEMmicrographs of xerogels: (a, b) 2 in o-xylene (4.0%w/v), (c) 4 in o-xylene (6.6%w/v), (d) 10 in chlorobenzene (2.5%w/v), (e)10 in bromobenzeme (2.5% w/v), and (f ) 10 in toluene (2.5% w/v).

Table 2. Tm (Temperature at Maximum Heat Flow, �C) of TolueneGels at Different Concentrations of 3

%, w/w Tm

1.0 56.11.7 57.72.6 58.13.2 62.33.8 61.04.5 62.35.1 61.3

Figure 4. Semilog plot of the mole fraction of 3 in toluene gelsvs the inverse of the gel-to-sol transition temperature (tempera-ture at maximum heat flow during the second heating DSCthermogram).

DOI: 10.1021/la8042796 8669Langmuir 2009, 25(15), 8663–8671

Bag et al. Article

(but does not require) the molecules being organized in a layeredarrangement. The diffractogram of the toluene gel of 10 is verydifferent and exhibits a number of sharp, low-angle peaks. Onecorresponding to 37.6 A is suggestive of a bilayered packingarrangement, but it is a bit too long and it was not possible toindex the other peaks. Thus, no definitive statement about themolecular packing arrangements in the neat solid and (espe-cially) the toluene gel of 10 can be made.Kinetics of Gelation.Gels were prepared by dissolving 10 in

chloroform and adding an equal weight of cyclohexane so thatthe total gelator concentration was 1.2% (w/w). The mixturewas heated to∼80 �C in a water bath and cooled to 24 �C for 15min to obtain a transparent gel whose Tgel was 60-64 �Caccording to the falling drop technique.12

The intensities of CD spectra of this 1.2% (w/w) 10 gel wereinfluenced by linear dichroic effects that arose as the SAFIN

formed in the cell; rotating the cell resulted in large changes inthe intensities.20 However, the dichroic spectra are qualitativelyas expected on the basis of the UV-vis absorption spectra whereλmax = 264 nm (Supporting Information, Figure S6). To over-come this problem, the concentration of 10 in the liquid mixturewas decreased (Supporting Information, Figure S7), and at0.4% (w/w) 10, the orientational effects were no longer dis-cernible. The SAFIN of the gel at 24 �C consists of spherulites(Figure 7a), and its birefringence is reduced at higher tempera-tures as the gelator molecules are increasingly dissolved in theliquid (Figure 7b).

Figure 8A is a compilation of CD spectra that follow theformation of the SAFIN for 0.4% (w/w) 10 at different timesafter it was incubated in 1/1 (w/w) chloroform/cyclohexane at25 �C. Change in the intensities and spectral shapes can be

related to the degree of aggregation and, thereby indirectly, tothe amount of the gel network present at various times if oneassumes that no aggregates of 10 are large enough at time= 0 toinduce macrochirality and that the aggregation and extent offibrillar network formation are at their equilibrium values whenthe CD intensities no longer increase with time. The kineticsof gelation of the sol phases of other optically active gelatorshave been analyzed from CD data according to eqs 2 and 3.7b

The fraction of the gel phase at time = t, φgt (φg

t is the volumefraction of the gel phase), can be expressed in terms of theCD intensities at time = 0, t, and ¥ (I(0), I(t), and I(¥), respec-tively): φtg{[I(t) - I(0)]/[I(¥) - I(0)]}. Then, values of K (K is atemperature-dependent parameter, analogous to a rate con-stant) are obtained (in this case) from single-exponential decayfits of the data using eq 2.

1-φtg ¼ expð-KtnÞ ð2Þ

ln½-lnð1-φtgÞ� ¼ nðln K þ ln tÞ ð3Þ

From the treatment of the data in Figure 8B according to eq 3, asshown in Figure 8C, the value of the Avrami exponent (n), whichis near unity (1.37( 0.02), is calculated. It indicates that 1D-typenucleation and growth is occurring.22 Then, the value of K wasobtained from Figure 8D using eq 2. The Avrami n and K valuesfrom experiments at different incubation temperatures are col-lected in Table 4. From a plot of the logarithm of the K valuesversus the inverse temperature (Figure 9), the thermodynamicdriving force for nucleation and growth (as a result of super-saturation of 10 in the liquid) ofSAFINwas calculated to be-5.0( 0.8 kcal/mol. This value compares favorably with the heat ofmelting of neat 10 obtained from DSC, +7.5 kcal/mol, andsuggests that, at least here, the rate-limiting step in the formationof the SAFIN in the gel is the nucleation and growth of solid 10.

There are very few numbers for comparison in the literature.A much lower value, ca. 95-135 cal/mol, has been reportedfor the nucleation and growth of fibers by 5-R-cholestan-3-β-ylN-(2-napthyl)carbamate (CNC) in n-alkane gels.7b The energyof activation for the nucleation of 12-hydroxystearic acid in avegetable oil gel, 2.9 kcal/mol,23 is slightly lower, and thecrystallization of a triacyl glycerol (consisting of hydrogenatedpalm oil, partially hydrogenated palm oil, and a blend ofpartially hydrogenated palm oil and palm stearine) from its melt(a related, but phenomenologically different, process from fiberformation in a gel) has a nucleation energy of 5.8 kcal/mol.24

Heats of transition from gel-to-sol phases calculated from theSchr

::oder-van Laar equation for cholesteric LMOGs having

p-alkoxyazobenzene (where alkoxy is methoxy, ethoxy, pro-poxy, butoxy, pentoxy, and decoxy) groups were also higher(8-39 kcal/mol) than in the present system.16

Qualitatively, the rate of gelation increases with decreasingtemperature. Quantitative analyses based on eq 3 yield Avrami

Figure 5. Vertically offset XRD diffractograms at 24 �C of (a) agel consisting of 5.1% (w/w) 3 in toluene after (approximate)empirical subtraction of solvent diffraction and (b) neat 3 powder.

Figure 6. Vertically offset XRD diffractograms at 24 �C of (a) agel consisting of 5.1% (w/w) 10 in toluene after (approximate)empirical subtraction of solvent diffractions and (b) neat 10.

Table 3. Comparison of Bragg Spacings (d, A) of 3 and 10 in Their

Neat Powders and Toluene Gels (from XRD data at 24 �C) and TheirCalculated Extended Molecular Lengths (L, A).

L15 d (powder) d (gel)

3 16.7 23.2, 6.5 23.2, 6.510 17.2 15.8, 6.2, 4.7, 4.1 37.6, 28.5, 24.5, 19.9,

14.8, 10.8, 6.8, 4.9, 4.1

(20) Gottarelli, G.; Spada, G. P; Castiglioni, E. In Molecular Gels: Materialswith Self-Assembled Fibrillar Networks; Weiss, R. G.; Terech, P., Eds. Springer:Dordrecht, The Netherlands, 2006; pp 431ff.(21) (a) Avrami,M. J. Chem. Phys. 1939, 7, 1103. (b) Avrami,M. J. Chem. Phys.

1940, 8, 212.

(22) Liu, X. Y.; Sawant, P. D. Adv. Mater. 2002, 14, 421.(23) Rogers, M. A.; Marangoni, A. G. Cryst. Growth Des. 2008, 8, 4596.(24) Marangoni, A. G.; Tang, D.; Singh, A. P.Chem. Phys. Lett. 2006, 419, 259.

8670 DOI: 10.1021/la8042796 Langmuir 2009, 25(15), 8663–8671

Article Bag et al.

exponents that are slightly larger than 1 at all temperaturesinvestigated (Table 4). The K values from analyses of the datausing eq 2 decrease with increasing temperature, as expected of aprocess in which the nucleation and growth of fibers is con-trolled by thermodynamic considerations.25

Conclusions

Wehave demonstrated that low concentrations of several alkyland (4-nitrophenyl)methyl esters of a plant-derived 6,6,6,6,6-pentacyclic triterpenic acid, arjunolic acid, self-assemble in a widerange of organic liquids, forming gels (i.e., fibrillar networks withconcomitant immobilization of the liquid components). By ester-ifying the carboxyl end of arjunolic acid, we have made a series of

derivatives that are less polar than the parent acid, but eachretains three differing structural/polarity regions. Whereas theparent acid, arjunolic acid, was insoluble in all of the organicliquids examined except the polar, protic ones, its esters can bedissolved, at least at elevated temperatures, in many of them. Themorphology of the gels at several length scales from optical andelectron micrographs reveals spherulitic-type organizations com-posed of very long fibers with nanometer-range widths. Several ofthe fibers were chiral, exhibiting right-handed helicities. Thus, themolecular chirality is expressed at much longer length scales.

On the basis of data from a combination of XRD and DSCexperiments, solvent-precipitated 3 and 10 are not very crystal-line, and the very low heats of their gel-to-sol transitions intoluene are consistent with their SAFIN structures being dis-ordered as well. The growth kinetics of 1/1 (w/w) chloroform/

Figure 7. Polarizing optical micrographs of a 1.2% (w/w) 10 in 1/1 (w/w) chloroform/cyclohexane gel at (a) 24 �C and (b) 40 �C. The blackobject in the micrographs is an instrumental artifact.

Figure 8. (A)CDspectra of a 0.4%(w/w) 10 in 1/1 (w/w) chloroform/cyclohexane sample thatwas cooled fromthe sol at 68 to 25 �Cover 30 sand incubated there.The spectrawere recorded at (a) 0, (b) 75, and (c) 120 s after reaching 25 �C.The timeneeded to record each spectrumwas12 s. (B) CD intensity at 264 nmvs time for a 0.4% (w/w) 10 in 1/1 (w/w) chloroform/cyclohexane sample thatwas cooled from the sol at 67 to6.1 �C over 30 s and incubated there; intensities are recorded after the final temperature was reached. (C) CD spectral intensities at 264 nmfromspectra inpartBplotted according to theAvrami equation (eq 3, see the text).21 (D)CDspectral intensities at 264nm frompartBplottedaccording toAvrami equation (eq2).21The red line is thebest fit of thedata toa single-exponential decay. InbothCandD, time=0 is taken tobe the start of the inflection of the data in part B (i.e., after the induction period).

(25) Liu, X. Y.; Sawant, P. D. Appl. Phys. Lett. 2001, 79, 3518.

DOI: 10.1021/la8042796 8671Langmuir 2009, 25(15), 8663–8671

Bag et al. Article

cylcohexane gels of 10, as monitored by circular dichroism and

analyzed by the Avrami equation, suggest 1D growth of fibers

and the thermodynamic driving force for phase separation as the

major factor dominating the rates of nucleation and fiber growth.The ability of 2-10 to gelate a wide variety of organic liquids

and the correlations explored here between the gel properties andester chain length and type of the gelators provide a comprehen-sive data set for understanding at least qualitatively the factorsnecessary to form this class of gels. In some respects, they areunique because all but one lacks an aromatic ring and the

positions of their three hydroxyl groups, localized at one end ofthe fused ring system constituting the molecular frame, are verydifferent from those of the analogous bile acids.9 As such, they fitinto a category of steroidal-like gelators lacking an aromatic or alinker group. The relatively simple structures of 2-10, withhydroxyl groups at one end of a stiff backbone and the estergroups at the other, suggest microphase functional-group segre-gation analogous to that proposed to understand the packing ofthe ALS LMOGs.

Thus, theSAFINs obtained from the esters of the plant-derivedpentacyclic triterpenoid, arjunolic acid, open interesting possibi-lities for exploring other plant metabolites that might be sustain-able feedstocks for LMOGs and other precursors of newmaterials. They join a select number of naturally occurringmolecules in which the basic structures are amenable to a widevariety of structural modifications, each potentially an LMOG,and in which the sets provide valuable information about the linkbetweenmolecular structure and self-assembly leading toSAFINsand gels. In principle, it should be possible tomodify the structureof the lipophic platformand,more easily, the end of arjunolic acidwhere the three hydroxyl groups reside to make other series ofLMOGs in which the specific aggregation modes are variedsystematically in different ways. Efforts to do so are ongoing.

Acknowledgment. Financial assistance from DRDO, NewDelhi and AvH Foundation, Germany are gratefully acknowl-edged. S.K.D. and P.P.D. thank CSIR, New Delhi for researchfellowships. A.M. and R.G.W. thank the U.S. National ScienceFoundation for its support of the part of this research conductedat Georgetown.

Supporting Information Available: Spectroscopic andphysical characterization of 2-7 and 9, DSC thermograms(second heating) of toluene gels at different concentrations of3, DSC thermogram of the first and second heating andcooling cycles of neat 3 and 10, DSC thermogram of the firstheating and cooling cycle of a 5.1% (w/w) gel of 10 intoluene, lengths and structures of fully extended conforma-tions of 3 and 10 simulated by molecular mechanics MM2calculations, UV-visible absorption spectrum of a 1.2%(w/w) 10 in 1/1 (w/w) chloroform/cyclohexane gel at 24 �C,Tgel values of 1/1 (w/w) chloroform/cyclohexane gels as afunction of concentration of 10, mps of compounds 1-10,and additional gelation test results. This material is availablefree of charge via the Internet at http://pubs.acs.org.

Figure 9. ln (K) vs 1/T plot for 0.4% (w/w) 10 in 1/1 (w/w)chloroform/cyclohexane.

Table 4. Summary of K and n Values from Analyses of CD Data for

Isothermal Gelation of 0.4% (w/w) 10 in 1/1 (w/w) Chloroform/

Cyclohexane at Different Temperatures

temp (�C) na,b K ((sn)-1)c

1.8 1.14 ( 0.02 0.0826.1 1.37 ( 0.02 0.0937.3 1.29 ( 0.02 0.08512.1 1.15 ( 0.01 0.08318.1 1.30 ( 0.02 0.06525.0 1.16 ( 0.01 0.04530.3 1.25 ( 0.01 0.039aFrom eq 3. bAllR2 correlation values were 0.99. cFrom eq 2, taking

n as the calculated values from eq 3.