DissertationFormation and Stabilization ofVesi les in Mixed Surfa tant Systems

Nina Vla hy

University of RegensburgNatural S ien es Fa ulty IVChemistry & Pharma yPh.D. Supervisor: Prof. Dr. Werner KunzAdjudi ators: Prof. Dr. Werner KunzProf. Dr. Ksenija KogejProf. Dr. Jörg HeilmannChair: Prof. em. Dr. Dr. Josef BarthelJuly 2008

Promotionsgesu ht eingerei ht am: 20. April 2008Promotionskolloquium am: 4. Juli 2008

For Katko: my sister, my toughest ompetitor, my biggest fan

iv

ContentsContents vPrefa e xi1 Introdu tion 11.1 Surfa e A tive Agents . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.2 Surfa tant Self-Assembly: Vesi les . . . . . . . . . . . . . . . . 31.1.3 Catanioni Surfa tant Mixtures . . . . . . . . . . . . . . . . . 41.1.4 Appli ation of Catanioni Vesi les in Cosmeti Formulation . . 71.2 Ion E�e ts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81.2.1 Ions in Water . . . . . . . . . . . . . . . . . . . . . . . . . . . 81.2.2 Hofmeister E�e t . . . . . . . . . . . . . . . . . . . . . . . . . 111.2.3 Ion Pairing in Water . . . . . . . . . . . . . . . . . . . . . . . 121.2.4 Collins' Theory of Mat hing Water A�nities . . . . . . . . . . 131.2.5 Ion-Spe i� E�e ts in Colloidal Systems . . . . . . . . . . . . 17I Salt-Indu ed Morphologi al Transitions in Non-EquimolarCatanioni Systems 192 Blastulae Aggregates: Spontaneous Formation of New Catanioni Superstru tures 212.1 Abstra t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.2 Introdu tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22v

vi CONTENTS2.3 Experimental Pro edures . . . . . . . . . . . . . . . . . . . . . . . . . 242.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262.4.1 Chara terization of SDS / DTAB Mi ellar Solution . . . . . . 262.4.2 Salt-Indu ed Mi elle-to-Vesi le Transition . . . . . . . . . . . 272.5 Dis ussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322.5.1 Models of the Mi elle-to-Vesi le Transition . . . . . . . . . . . 322.5.2 Blastulae Vesi les . . . . . . . . . . . . . . . . . . . . . . . . . 352.5.3 The O urren e of Convex-Con ave Patterns in Biologi al Sys-tems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372.5.4 Raspberry Vesi les . . . . . . . . . . . . . . . . . . . . . . . . 382.5.5 Blastulae Vesi les: A General Trend in Catanioni Systems? . 382.6 Con lusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393 Spe i� Alkali Cation E�e ts in the Transition from Mi elles toVesi les Through Salt Addition 413.1 Abstra t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413.2 Introdu tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423.3 Experimental Pro edures . . . . . . . . . . . . . . . . . . . . . . . . . 433.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443.4.1 Phase Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . 443.4.2 Counterion E�e ts . . . . . . . . . . . . . . . . . . . . . . . . 453.4.3 Co-ion E�e ts . . . . . . . . . . . . . . . . . . . . . . . . . . . 513.4.4 Nonioni E�e ts . . . . . . . . . . . . . . . . . . . . . . . . . . 513.4.5 E�e ts of `Hydrophobi Ions' . . . . . . . . . . . . . . . . . . 523.5 Dis ussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543.5.1 Aggregation Behavior of Catanioni Systems . . . . . . . . . . 543.5.2 Counterion Properties . . . . . . . . . . . . . . . . . . . . . . 553.5.3 Collins' `Law of Mat hing Water A�nities' . . . . . . . . . . . 563.5.4 Counterion Sele tivity of Alkyl Sulfates . . . . . . . . . . . . . 563.5.5 Counterion Sele tivity of Alkyl Carboxylates . . . . . . . . . . 573.5.6 Alkyl Sulfates vs. Alkyl Carboxylates . . . . . . . . . . . . . . 57

CONTENTS vii3.5.7 Mole ular Dynami s Simulations . . . . . . . . . . . . . . . . 583.5.8 Generalization of the Con ept: Hofmeister Series of Headgroups 603.5.9 The Anioni E�e t . . . . . . . . . . . . . . . . . . . . . . . . 603.5.10 The Non-Ioni E�e t . . . . . . . . . . . . . . . . . . . . . . . 603.5.11 E�e ts of `Hydrophobi Ions' . . . . . . . . . . . . . . . . . . 623.6 Con lusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 634 Anion Spe i� ity In�uen ing Morphology in Catanioni Surfa tantMixtures with an Ex ess of Cationi Surfa tant 654.1 Abstra t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 654.2 Introdu tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 664.3 Experimental Pro edures . . . . . . . . . . . . . . . . . . . . . . . . . 674.4 Results and Dis ussion . . . . . . . . . . . . . . . . . . . . . . . . . . 684.4.1 Ion Binding in Catanioni Surfa tant Mixtures . . . . . . . . . 684.4.2 Phase Behavior upon Salt Addition . . . . . . . . . . . . . . . 704.4.3 Anion Spe i� ity in Physi o-Chemi al Properties of Alkyl-trimethylammonium Systems . . . . . . . . . . . . . . . . . . 724.4.4 In�uen e of Salt on the Aggregation Behavior of Surfa tants . 734.4.5 Explaining Counterion Spe i� ity in Surfa tant Systems . . . 764.4.6 Di�erent self-aggregation behavior of atanioni systems inthe atanioni - and anioni -ri h regions . . . . . . . . . . . . . 774.5 Con lusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78II In reasing the Stability of Catanioni Systems 795 In�uen e of Additives and Cation Chain Length on the Kineti Stability of Supersaturated Catanioni Systems 815.1 Abstra t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 815.2 Introdu tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 825.3 Experimental Pro edures . . . . . . . . . . . . . . . . . . . . . . . . . 835.4 Results and Dis ussion . . . . . . . . . . . . . . . . . . . . . . . . . . 855.4.1 Shift of Solubility Temperature with Time . . . . . . . . . . . 85

viii CONTENTS5.4.2 Behavior of the Anioni -Ri h Region of the Phase DiagramsWithout Additives . . . . . . . . . . . . . . . . . . . . . . . . 885.4.3 E�e t of Additives on the Stability and the `Solubility Tem-perature Depression' . . . . . . . . . . . . . . . . . . . . . . . 925.5 Con lusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101III Toward Appli ation 1036 Use of Surfa tants in Cosmeti Appli ation: Determining the Cy-totoxi ity of Catanioni Surfa tant Mixtures on HeLa Cells 1056.1 Abstra t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1056.2 Introdu tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1066.3 Experimental Pro edures . . . . . . . . . . . . . . . . . . . . . . . . . 1076.3.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1076.3.2 Growing HeLa Cell Cultures . . . . . . . . . . . . . . . . . . . 1086.3.3 HeLa Toxi ity Test . . . . . . . . . . . . . . . . . . . . . . . . 1096.3.4 Dete tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1116.3.5 Evaluation of spe tros opi al data . . . . . . . . . . . . . . . . 1116.4 Results and Dis ussion . . . . . . . . . . . . . . . . . . . . . . . . . . 1116.4.1 Single-Chain Surfa tants . . . . . . . . . . . . . . . . . . . . . 1116.4.2 Catanioni Surfa tant Systems . . . . . . . . . . . . . . . . . . 1136.5 Con lusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1167 Spontaneous Formation of Bilayers and Vesi les in Mixtures ofSingle-Chain Alkyl Carboxylates: E�e t of pH and Aging 1177.1 Abstra t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1177.2 Introdu tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1177.3 Experimental Pro edures . . . . . . . . . . . . . . . . . . . . . . . . . 1207.4 Results and Dis ussion . . . . . . . . . . . . . . . . . . . . . . . . . . 1227.4.1 Lowering of the Solubility Temperature of Fatty A ids . . . . 1227.4.2 The E�e t of pH on Vesi le Formation . . . . . . . . . . . . . 1237.4.3 Cryo-TEM Study of Time-Dependent Vesi le Formation . . . 127

7.4.4 Cytotoxi ity Potential of Surfa tant Mixtures . . . . . . . . . 1307.5 Con lusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131Con lusion 133Bibliography 137A knowledgements 161List of Publi ations 163List of Oral and Poster Presentations 165

ix

x

Prefa eSurfa tants an self-assemble in dilute aqueous solutions into a variety of mi rostru -tures, in luding mi elles, vesi les, and bilayers. In re ent years, there has been anin reasing interest in unilamellar vesi les, whi h are omposed of a urved bilayerthat separates an inner aqueous ompartment from the outer aqueous environment.This interest is motivated by their potential to be applied as vehi les for a tiveagents in osmeti s and pharma y. A tive mole ules an be en apsulated in thebilayer membrane if they are lipophili or in the ore of the vesi le if they arehydrophili . Furthermore atanioni systems an be used as models for biologi almembranes.When oppositely harged surfa tants are mixed, new properties appear due to thestrong ele trostati intera tions between the harged headgroups. These so- alled atanioni mixtures exhibit low riti al mi elle on entration ( m ) values and anon-monotoni hange in the surfa tant pa king parameter (P ) as the mixing ratiois varied. One advantage of atanioni systems, as ompared with more robust gen-uinely double hained surfa tants, are their greater sensitivity to parameters su has temperature or the presen e of salts. To optimize the appli ations it is importantto have a general understanding of the interplay of intera tions between the surfa -tants and of the fa tors in�uen ing the phase diagram of a mixed system. In thisthesis the formation and stabilization of atanioni vesi les was studied. The e�e tof al ohol and salt on the morphologi al transitions is des ribed, and the potentialof using su h systems for osmeti al and pharma euti al purposes was explored.A short overview of the phase behavior of atanioni systems is given in the Intro-du tion.The morphologi al transitions o urring in mixed surfa tant systems upon the in-xi

rease of the ioni strength are reported in Chapter 2. A mi elle-to-vesi le transitionis investigated and an intriguing new intermediate stru ture, named �blastulae vesi- le�, is des ribed.Many phenomena in olloid s ien e that involve ele trolytes show pronoun ed ionspe i� ity. In order to elu idate the spe i� ion e�e ts in atanioni surfa tant sys-tems with di�erent surfa tant headgroups were ompared. Using various experimen-tal te hniques and MD simulations, and employing a general on ept of `mat hingwater a�nities', a detailed study of ion spe i� ity in atanioni systems is des ribedin Chapters 3 and 4. A Hofmeister like series for lassifying surfa tant headgroupsis established.Considering that ontrolling the pre ipitation phenomena is useful for a vast numberof industrial appli ations, the e�e t of various additives on the stability of atanioni systems was probed and is reported in Chapter 5.Chapter 6 introdu es an important parameter that should be onsidered when for-mulating new en apsulating systems for either osmeti or medi al use: the toxi ityof the parti ipating surfa tant mole ules. We evaluated the ytotoxi ity of a numberof ommonly used surfa tants, as well as that of atanioni surfa tant mixtures.Seeing that the presen e of only a small amount of a ationi surfa tant in the mix-ture results in a large in rease in its toxi ity, we fo used on a mixture of two anioni and in osmeti formulation already ommonly used surfa tants. In Chapter 7 wepresent a way to form vesi les in su h mixtures at room temperature and at physi-ologi al pH.This thesis omprises di�erent studies ranging from phase diagram determination to ytotoxi ity studies. For this reason, the thesis was written so that ea h hapter lay-out follows the usual onvention: Abstra t, Introdu tion, Experimental Pro edures,Results, Dis ussion and Con lusions. The bibliography is given at the end of thethesis in order to avoid repeating. The di�erent studies led to several publi ationswhi h are already published, a epted or submitted, summarized in the followingtable. A omplete list of publi ations and a list of oral and poster presentations,whi h were presented at international ongresses, are also given at the end.xii

Chapter Publi ation2 N. Vla hy, A. Renon ourt, M. Dre hsler, J.-M. Verbavatz, D. Touraud,W. Kunz, Blastulae aggregates: New intermediate stru tures in themi elle-to-vesi le transition of atanioni systems. J. Colloid Interf. S i.2008 320, 360-363.3 A. Renon ourt, N. Vla hy, P. Bauduin, M. Dre hsler, D. Touraud, J.-M.Verbavatz, M. Dubois, W. Kunz, B. W. Ninham, Spe i� alkali atione�e ts in the transition from mi elles to vesi les through salt addition.Langmuir 2007 23, 2376-2381.3 N. Vla hy, M. Dre hsler, J.-M. Verbavatz, D. Touraud, W. Kunz, Roleof the surfa tant headgroup on the ounterion spe i� ity in the mi elle-to-vesi le transition through salt addition. J. Colloid Interf. S i. 2008319, 542-548.3 N. Vla hy, B. Jagoda-Cwiklik, R. Vá ha, D. Touraud, P. Jungwirth,and W. Kunz, Hofmeister series of headgroups and spe i� intera tion of harged headgroups with ions. J. Phys. Chem. B 2008 (Submitted).4 N. Vla hy, M. Dre hsler, D. Touraud, W. Kunz, Anion spe i� ity in-�uen ing morphology in atanioni surfa tant mixtures with an ex essof ationi surfa tant. Comptes rendus Chimie A adémie des s ien es2008 (Submitted).5 N. Vla hy, A. F. Arteaga, A. Klaus, D. Touraud, M. Dre hsler, W. Kunz,In�uen e of additives and ation hain length on the kineti stability ofsupersaturated atanioni systems. Colloids Surf., A 2008 (A epted).6 N. Vla hy, D. Touraud, J. Heilmann, W. Kunz, Determining the de-layed ytotoxi ity of atanioni surfa tant mixtures on HeLa ells.Colloids Surf., B 2008 (Submitted).7 N. Vla hy, C. Merle, D. Touraud, J. S hmidt, Y. Talmon, J. Heilmann,W. Kunz, Determining the delayed ytotoxi ity of atanioni surfa tantmixtures on HeLa ells. Langmuir 2008 (Submitted).xiii

xiv

Chapter 1Introdu tion1.1 Surfa e A tive Agents1.1.1 GeneralSurfa e a tive agents (a.k.a. surfa tants) are mole ules with a hemi al stru turethat makes them parti ularly favorable to reside at interfa es. All lassi al sur-fa tant mole ules onsist of at least two parts, one whi h is soluble in water (thehydrophili part) and the other whi h is insoluble in water (the hydrophobi part).The hydrophili part (a polar or ioni group) is referred to as the head group and thehydrophobi part (a long hydro arbon hain) as the tail (Figure 1.1). The hydropho-bi part of a surfa tant may be bran hed or linear. The degree of hain bran hing,the position of the polar head group and the length of the hain are parametersthat a�e t the physi o- hemi al properties of the surfa tant1,2. The primary lassi-� ation of surfa tants is made on the basis of the harge of the polar head group:anioni s, ationi s, non-ioni s and zwitterioni s.Surfa tant mole ules adsorb at interfa es, thereby redu ing the free energy of thesystem3. When surfa tants are dissolved in water, the hydrophobi group disruptsthe hydrogen-bonded stru ture of water and therefore in reases the free energy ofthe system. Surfa tant mole ules therefore on entrate at interfa es, so that theirhydrophobi groups are removed or dire ted away from the water and the free energyof the solution is minimized. The distortion of the water stru ture an also be de-1

Figure 1.1: S hemati representation of a surfa tant mole ule (with all the parameters involvedin the theory of the pa king parameter) and a unilamellar vesi le. reased (and the free energy redu ed) by the aggregation of surfa e-a tive mole ulesinto lusters (mi elles) with their hydrophobi groups dire ted toward the interiorof the luster and their hydrophili groups dire ted toward the water. The pro essof surfa tant lustering or mi ellization is primarily an entropy-driven pro ess1,3.However, the surfa tant mole ules transferred from the bulk solution to the mi ellemay experien e some loss of freedom from being on�ned to the mi elle. In addi-tion, they may experien e an ele trostati repulsion from other similarly hargedsurfa tant mole ules in the ase of surfa tants with ioni head groups. These for esin rease the free energy of the system and oppose mi ellization. Hen e, mi elleformation depends on the for e balan e between the fa tors favoring mi ellization(van der Waals and hydrophobi for es) and those opposing it (kineti energy of themole ules and ele trostati repulsion).The on entration at whi h mi elles �rst appear in solution is alled the riti almi ellar on entration, abbreviated CMC, and an be determined from the dis on-tinuity of the in�e tion point in the plot of a physi al property of the solution as afun tion of the surfa tant on entration1,3. Above this on entration, almost all ofthe added surfa tant mole ules are onsumed in mi elle formation and the monomer on entration does not in rease appre iably, regardless of the amount of surfa tantadded to the solution. Sin e only the surfa tant monomers adsorb at the interfa e,the surfa e tension remains essentially onstant above the CMC. Thus, the surfa etension an be dire tly related to the a tivity of monomers in the solution3. TheCMC is strongly a�e ted by the hemi al stru ture of the surfa tant4,5, by the tem-2

perature6 and by the presen e of osolutes su h as ele trolytes7 or al ohols6. Mi elleformation, or mi ellization, an be viewed as an alternative me hanism to adsorp-tion at interfa es. In both ases, the driving for e of the pro ess is the tenden y ofthe surfa tant to remove their hydrophobi groups from the onta t with water8.1.1.2 Surfa tant Self-Assembly: Vesi lesAs was mentioned previously, surfa tants an self-assemble in dilute aqueous solu-tions into a variety of mi rostru tures, in luding mi elles, vesi les, and bilayers. Inre ent years, there has been an in reasing interest in unilamellar vesi les, whi h are omposed of a urved bilayer that separates an inner aqueous ompartment fromthe outer aqueous environment (Figure 1.1). This is mainly be ause of their wideappli ation in biology and medi ine as model ell membranes, as well as their strongpotential as drug arriers and other en apsulating agents of industrial relevan e9.Two major theoreti al approa hes have been pursued in the modeling of surfa tantself-assembly: the urvature-elasti ity approa h and the mole ular approa h.The urvature-elasti ity approa h des ribes the vesi le bilayer as a ontinuousmembrane hara terized by the spontaneous urvature and the elasti bending mod-ulus10,11. In this approa h, the formation of �nite-sized vesi les thus depends on theinterplay between these two quantities12,13. The theory provides an elegant, simpleway to des ribe the formation of vesi les, however, be ause this approa h is basedon a urvature expansion of the free energy of a membrane, it breaks down for smallvesi les, for whi h the urvature is quite pronoun ed.The mole ular approa h was pioneered by Israela hvili, Mit hell, and Ninham14�16who developed a geometri pa king argument that allows one to predi t the shapeof self-assembling mi rostru tures, in luding spheroidal, ylindri al, or dis oidal mi- elles, vesi les, and bilayers. Whi h aggregates form, is determined primarily bygeometri pa king of amphiphiles, hydro arbon hain sti�ness, and the hydrophili -hydrophobi balan e. For dilute solutions in whi h intera tions between aggregatesare not important, the ne essary (geometri ) onditions for formation of an aggre-gate an be des ribed by a a surfa tant pa king parameter P = v/(lmaxa)14�16, wherev is the volume per hydro arbon hain, or the hydrophobi region of the surfa tant,3

a is the a tual headgroup area in the �lm, and lmax is an optimal hydro arbon hainlength related to about 90% of the maximum extended length (see Figure 1.1). Theoptimal stability of the di�erent aggregates o urs as follows: (1) spheri al mi ellesP ≤ 1/3; (2) globular or ylindri al mi elles 1/3 < P ≤ 1/2 (3) vesi les or bilayers1/2 < P ≤ 1.Low pa king parameters (around 1/3) are found for single hained surfa tantswith a strongly polar head group. An in rease in the pa king parameter anbe obtained by adding a se ond hain, therefore doubling the hydro arbon vol-ume. To rea h this value, double hain surfa tants17�22, two surfa tants of opposite harge23�26, or the asso iation of a surfa tant and a o-surfa tant27�33 an be used.In the latter two ases, a pseudo-double hain surfa tant is obtained by either an ion-pair formation between the anioni and ationi surfa tant or due to an asso iationof the two di�erent mole ules via hydrogen bounds.In many ases, the formation of vesi les requires the input of some form of en-ergy, for example, soni ation, inje tion or extrusion34�39. However, vesi les havebeen found to form spontaneously in some aqueous surfa tant systems, in lud-ing solutions ontaining (i) mixtures of le ithin and lysole ithin40, (ii) mixturesof long- and short- hain le ithins41, (iii) mixtures of AOT and holine hloride42,(iv) dialkyldimethylammonium hydroxide surfa tants20,21,43�45, (v) ationi siloxanesurfa tants46, and (vi) mixtures of ationi and anioni surfa tants24,47�50. Thesespontaneously-forming vesi les o�er advantages over the more traditional phospho-lipid vesi les in being easier to generate and more stable, thus making them moreattra tive as en apsulating agents in diverse pra ti al appli ations, in luding the ontrolled delivery of drugs, a tive substan es in osmeti s, and fun tional foodingredients su h as enzymes51,52.1.1.3 Catanioni Surfa tant MixturesThe main thermodynami driving for e for the asso iation of a ationi and ananioni surfa tant mixture is the release of ounterions from the aggregate surfa es.This results in a large entropy in rease. Sin e the two surfa tants are single- hained,the resulting atanioni surfa tant an be onsidered as a pseudodouble- hained4

surfa tant, in the sense that the two hains are not ovalently bound to the sameheadgroup. For these systems a non-monotoni hange in P is observed, with apronoun ed maximum as the mixing ratio is varied1,3. Due to the strong ele trostati intera tion between the oppositely harged headgroups, atanioni mixtures exhibitCMC values mu h lower than those of the surfa tants involved. The CMC is dire tly orrelated to the leaning e� ien y of surfa tants, pointing to yet another advantageof su h systems.Mixtures of anioni and ationi surfa tants exhibit ri h mi rostru ture phasebehavior in aqueous solutions. Aggregate stru tures su h as spheri al and rod-likemi elles, vesi les, lamellar phases, and pre ipitates have all been observed depend-ing on the on entration and the ratio of the surfa tants in solutions24,26,48,53�57.Around equimolarity a zone of pre ipitation is observed. However, when one of thesurfa tants is present in a small ex ess, the ationi -anioni surfa tant bilayers usu-ally spontaneously form losed vesi les. The phase behavior of atanioni mixturesis represented in Figure1.2.

Figure 1.2: S hemati phase behavior en ountered in atanioni surfa tant systems. Phasenotations: V − and V +: regions of negatively and positively harged vesi les; 2φ: two-phase regions,i.e. mostly demixing of phases between a vesi ular and a lamellar phase or a vesi le and a mi ellarphase; L− and L+: lamellar phase with an ex ess of respe tively anioni and ationi surfa tants;P: pre ipitate region; I− and I+: mixed mi ellar solutions with an ex ess of respe tively anioni and ationi surfa tants (reprodu ed from Khan58).5

Yuet and Blanks htein59 presented a detailed mole ular thermodynami theoryto des ribe the formation of atanioni vesi les. Their theory reveals that (i) the dis-tribution of surfa tant mole ules between the two vesi le lea�ets plays an essentialrole in minimizing the vesi ulation free energies of �nite-sized vesi les, (ii) the om-position of mixed ationi /anioni vesi les is mainly determined by three fa tors: thetransfer free energy of the surfa tant tails, the ele trostati intera tions between the harged surfa tant heads, and the entropi penalty asso iated with the lo alizationof the surfa tant mole ules upon aggregation, and (iii) the entropy asso iated withmixing �nite-sized vesi les an be an important me hanism of stabilizing vesi les insolution. The present mole ular-thermodynami theory also has the ability to overthe entire range of vesi le sizes (or urvatures), thus enabling a des ription of small,energeti ally stabilized, vesi les.The free energy of vesi ulation, gves, an be viewed as omposed of the follow-ing �ve ontributions: (1) the transfer free energy, gtr, (2) the pa king free energy,gpack, (3) the interfa ial free energy, gσ, (4) the steri free energy, gsteric, and (5) theele trostati free energy, gelec. These �ve free-energy ontributions a ount for theessential features that di�erentiate a surfa tant mole ule in the vesi le and in themonomeri state. The transfer free energy, gtr, re�e ts the so- alled hydrophobi e�e t60, whi h onstitutes the major driving for e for surfa tant self-assembly in wa-ter. Indeed, the transfer free energy is the only favorable free-energy ontribution tomole ular aggregation, with the other four free-energy ontributions working againstthis pro ess. The hydrophobi region in a vesi le is di�erent from bulk hydro arbon.In a vesi le, the surfa tant tails are an hored at one end on either the outer or innerinterfa es, whi h restri ts the number of onformations that ea h surfa tant tail anadopt while still maintaining a uniform liquid hydro arbon density in the vesi lehydrophobi region. This subtle di�eren e between a bulk hydro arbon phase andthe hydrophobi region in a vesi le is aptured by the pa king free energy, gpack. Inaddition, free-energy penalties are imposed, upon aggregation, by the reation ofthe outer and inner hydro arbon/water interfa es, aptured in gσ, and by the steri repulsions and ele trostati intera tions between the surfa tant heads, aptured ingsteric and gelec, respe tively. 6

A distin tion is made in the literature between two types of atanioni sys-tems: (1) In the `simple mixtures' of ationi and anioni surfa tants or atanioni surfa tant systems with ex ess salt, both surfa tants still behold their own ounteri-ons. (2) The `true atanioni s' (ion pair amphiphiles) onsist of surfa tant systemswhere the original ounterions have been removed and repla ed by hydroxide andhydronium ions. The ombination of the ounterions at equimolarity thus formswater mole ules. Ea h surfa tant stands as ounterion for the surfa tant of opposite harge. The present work will fo us on the �rst type of atanioni systems (with ounterions).1.1.4 Appli ation of Catanioni Vesi les in Cosmeti Formu-lationVesi les are ommonly used in osmeti s and pharma y as vehi les for a tive agents.A tive mole ules an thus be en apsulated in the bilayer membrane if they arelipophili or in the ore of the vesi le if they are hydrophili . En apsulation is usefulto prote t a tives in preventing any undesired rea tion. Vesi les an thus be usedas ve tors to deliver drugs to a spe i� pla e, without being destroyed. The phar-ma euti al appli ations ontinuously in rease and vesi les are used more and morein the dermatology for prevention, prote tion and therapy.The �rst en apsulation experiments on atanioni systems were performed by Harg-reaves and Deamer61 on the etyltrimethylammonium bromide / sodium dode ylsul-fate system. They su essfully entrapped glu ose, however, the system was limitedto high temperatures (> 47◦C). Ten years later, Kaler at al.24 pro eeded with sim-ilar experiments using vesi les formed from etyltrimethylammonium tosylate andsodium dode ylbenzenesulfonate (CTAT/SDBS) mixtures. A more omprehensivestudy of the entrapment ability was made by Tondre et al.62 on the CTAT/SDBSand by Kondo et al.50 on the didode yltrimethylammonium bromide / sodium do-de ylsulfate (DDAB / SDS) system. The surfa tant on entration as well as theratio of the two parti ipating surfa tant have proved to have a big e�e t on theentrapment e� ien y. 7

Another problem is the release of the entrapped a tive ingredient. In most ases anon-ioni surfa tant is used (i. e. Triton X-100)50. A more elegant solution for tar-geted drug delivery is designing vesi les that be ome unstable at an easily tuned pHvalue. It is known that, for example, tumors and in�amed tissues exhibit a de reasedextra ellular pH63�67. For this reason a large number of groups have fo used theirattention on the preparation of pH-sensitive liposomes68�78 as possible drug arriersystems. Furthermore, improving the bio ompatibility of produ ts used in osmeti formulation even further is always sought after. For this reason it is important toidentify the skin irritating properties of ommer ially used surfa tants. In additionto their entrapment abilities, vesi les serve also as models for membranes of biolog-i al ells79�81 and as templates82�85 for the synthesis of nanoparti les, extra tion ofrare earth metal ions86, and as gene delivery systems87.1.2 Ion E�e ts1.2.1 Ions in WaterIt has long been known that the dissolution of ions brings about hanges in solventstru ture88. The region of modi�ed solvent surrounding an ion has been denotedas a osphere of the ion89. The degree and manner in whi h ospheres overlap inthe lose-range en ounter of two ions depends spe i� ally on the nature of bothions and the primary for es between them. Ion hydration has been studied exten-sively experimentally88,90�93 as well as theoreti ally94�98. Terms su h as ` onta t'pairs and `solvent-separated' pairs have ome into use to distinguish the results of omplete and partial elimination of solvent mole ules from between two intera tingions (see Se tion 1.2.3.). The basi features responsible for ion-spe i� short-rangeintera tions are, in the ase of monoatomi ions, their harge, their size, their po-larizability, and availability of ele trons and/or orbitals for ovalent ontributions.Additional features of polyatomi ions are the harge density distribution, and, insome ases, the presen e of hydrophobi groups99. The ease with whi h hydratione�e ts a ompanying asso iation-disso iation pro esses an be observed, depends8

on the number of ion pairs existent at any instant. Thus the e�e ts are observablemore readily with weak ele trolytes and with pairs of multivalent ions, than withstrong 1-1 ele trolytes.Ions have long been lassi�ed as being either kosmotropes (stru ture makers) or haotropes (stru ture breakers) a ording to their relative abilities to indu e thestru turing of water. The degree of water stru turing is determined mainly by twoquantities: the in rease or de rease of vis osity in water due to added salt, andentropies of ion solvation. For example, the vis osity η of an aqueous salt solutiontypi ally has the following dependen e on ion on entration c91:η/η◦ = 1 + Ac1/2 + Bc + · · · (1.1)where η◦ is the vis osity of pure water at the same temperature. A is a on-stant independent of c; its orresponding term an be explained by Debye-Hü keltheory as being due to ounterion s reening at low ion on entrations. The onstant

B, whi h is alled the Jones-Dole B oe� ient, is the quantity that des ribes thedegree of water stru turing. B is positive for kosmotropi ions and negative for haotropi ions. (One issue in interpreting experiments is how to separate the on-tributions of the anion from the ation. The standard assumption is that K+ hasthe same B oe� ient as Cl−, be ause K+ and Cl− have approximately the sameioni ondu tan e100 and be ause the value of B for KCl is approximately zero.)Water stru turing is also re�e ted in entropies of ion solvation. Again, the ef-fe ts of the anion from the ation need to be separated (it is assumed that thesolvation entropies are additive90). Furthermore, the ion solvation entropy needsto be divided into ion and hydration water ontributions. Subtra ting the former,∆SII is obtained, whi h des ribes the hange in entropy of hydration water due tothe presen e of an ion90. Ions whi h are kosmotropi in vis osity experiments tendto have a negative hydration omponent to their solvation entropy, implying thatthey order the nearby waters, while haotropi ions have a positive ∆SII . Figure1.3 plots the entropy of water near monovalent ions as al ulated from the entropyof hydration of the ion (from dissolving the ion in water) versus the ioni radius ofthe ion92. A negative ∆SII (upper portion of Figure 1.3) indi ates tightly bound9

0.5 1.0 1.5 2.0 2.5

-15

-10

-5

0

5

10

15

- S

II r (Å)

Li+

Na+

K+

Rb+

Cs+

F-

Cl-

Br-

I-

Figure 1.3: The entropy of pure water minus the entropy of water near an ion in calK−1 mol−1.The rystal radii of the ions in angstroms are plotted along the abs issa. Positive values of ∆SII(lower portion of �gure) indi ate water that is more mobile than bulk water. Negative values of∆SII (upper portion of �gure) indi ate water that is less mobile than bulk water. Kosmotropesare in the upper portion of the �gure; haotropes are in the lower portion of the �gure (Figurereprodu ed from Collins101).water that is less mobile than bulk water, whereas a positive ∆SII (lower portionof Figure 1.3) indi ates loosely held water that is more mobile than bulk water.In reasing ion size (de reasing ion harge density) is asso iated with in reasing mo-bility of nearby water mole ules. If this mobile, loosely held water is immediatelyadja ent to the ion, as suggested by x-ray and neutron di�ra tion data102, then thehorizontal line in Figure 1.3 indi ating ∆SII = 0, separates strongly hydrated ions(above the line) from weakly hydrated ions (below the line). Sin e this transitionfrom weak to strong hydration o urs at a larger size for anions than for ations,the anions must be more strongly hydrated than the ations, sin e anions begin toimmobilize adja ent water mole ules at a lower harge density than do ations. Theexperiments show that water is ordered by small or multivalent ions and disorderedby large monovalent ions. Therefore, water ordering has generally been interpretedin terms of ion harge densities90,103. Charge densities are high on ions that have asmall radius and/or a large harge.The water-ordering e�e t of ions is also theoreti ally extensively studied. Contin-10

uum ele trostati s models su h as that of Debye and Hü kel104 utilize a ma ros opi diele tri onstant and assume that all intera tions involving ions are stri tly ele -trostati , implying the existen e of long range ele tri �elds strong relative to thestrength of water-water intera tions. In these models, ions are often thought of aspoint harges and water as a dipole whi h orients in the long-range ele tri �eld.Su h models are unable to a urately des ribe su h simple ion-spe i� behaviorsas their tenden y to form onta t ion pairs, whi h is a major determinate of thesolubility of spe i� salts and of the role of spe i� ions in biologi al systems. Forexample, models employing a ma ros opi diele tri onstant predi t that all ionsare strongly hydrated and will be repelled from nonpolar surfa es by image for es. Infa t, weakly hydrated ions (e.g., ammonium, hloride, potassium, and the positively harged amino a id side hains) a tually adsorb to nonpolar surfa es105�107 and in-terfa es108,109. The driving for e for this adsorption has been to be the release ofweakly bound water to be ome strongly intera ting bulk110, a pro ess not in ludedin the al ulations utilizing the ma ros opi diele tri onstant. Sophisti ated mi- ros opi al ulations have indi ated a role for the polarizability111,112 of weaklyhydrated ions (as opposed to their dehydration energy) and dispersion for es113 indriving them to neutral interfa es.1.2.2 Hofmeister E�e tA related property is the Hofmeister e�e t114,115. In 1888, Franz Hofmeister om-pleted the �rst systemati study on spe i� -ion e�e ts. He reported that salts a�e tthe solubility of proteins in water. Certain ions pre ipitate proteins in water (`salt-ing out') while others help solubilize them (`salting in'). This behavior has beeninterpreted as a modulation of the hydrophobi e�e t by salts due to the hanges inthe water stru ture brought about by ions. Su h salt e�e ts orrelate with hargedensities of the salts. Small ions tend to ause `salting out', that is, to redu ehydrophobi solubilities in water, whereas large ions tend to ause `salting-in', in- reasing nonpolar solubilities. In Hofmeister's papers an ordering of salts, and lateran ordering of ions, usually denoted as `The Hofmeister series' was developed. `TheHofmeister series' orders ions as a monotoni fun tion of their surfa e harge den-11

sity and thus water a�nity, with the strength of water-water intera tions separatingstrongly hydrated from weakly hydrated spe ies. It is most onvenient to generatea separate series for anions and for ations ( . f. Figure 1.4).SO4

2- > OH- > F- > Cl- > Br- > NO3- > I- > SCN- > ClO4

-

NH4+ > K+ > Na+ > Cs+ > Li+ > Rb+ > Mg2+ > Ca2+ > Ba2+

Cosmotropes

Salting-out

“Hard”

Chaotropes

Salting-in

“Soft”

Organic

Figure 1.4: A typi al Hofmeister series.1.2.3 Ion Pairing in WaterIon pairing des ribes the (partial) asso iation of oppositely harged ions in ele -trolyte solutions to form distin t hemi al spe ies alled ion pairs. If the ion asso- iation is reasonably strong (the value depends on the harges of the ions and therelative permittivity of the solvent), there is usually little di� ulty in separating theproperties of the ion pair from the long-range nonspe i� ion-ion intera tions thatexist in all ele trolyte solutions. However, when the ion asso iation is weak, thereis a strong orrelation between these nonspe i� ion-ion intera tions ( hara terizedin terms of a tivity oe� ients) and ion pair formation ( hara terized in terms ofan asso iation onstant). Spe ies are generally des ribed as ion pairs if two oppo-sitely harged ions in solution stay together at a separation r, whi h is smaller thansome spe i�ed uto� distan e R. Ions further apart than R are onsidered `free'.Various theories have been proposed for hoosing the value of R and for des ribingthe properties of the ion pairs and free ions that together produ e the observed be-havior of ele trolyte solutions. It is generally a epted that ions an not approa hea h other more losely than some `distan e of losest approa h' a due to the strongrepulsive for es of the ele tron shells of the ions, even if polarizable. The distan ea is understood to bear some relation to the sum of the ( rystal ioni ) radii of theoppositely harged ions, generally a ≥ r+ + r−. In summary, two ions of oppositesign are onsidered to form an ion pair if their distan e apart is between a and R for12

a time longer than the time needed to di�use over su h a distan e116. On e ions arepaired, they are thought to have no tenden y to asso iate with other ions in dilutesolutions, although, at higher ele trolyte on entrations, ion triplets, quadruplets,or larger aggregates may form. A major role in the asso iation of ions in solutioninto pairs is thought to be played by long-range ele trostati for es between the ions,usually modeled as a Coulomb's law attra tion, attenuated by the solvent permit-tivity. Very short-range intera tions (hard or nearly-hard sphere repulsions) involvethe mutual ex lusion of ions at r < a. However, at distan es a < r < R, solvation ofthe onstituent ions must be onsidered. On this basis an ion pair may be lassi�edas a (double) solvent-separated ion pair (2SIP), when the primary solvation shellsof both ions remain essentially inta t, as a solvent-shared ion pair (SIP), if a singlesolvent layer exists in the spa e between the ion partners of the pair, or as a onta tion pair (CIP), if no solvent exists between the partners and the ions are in dire t onta t (Figure 1.5).Figure 1.5: S hemati representation of ion-pair types: (a) solvent separated (2SIP), (b) solventshared (SIP), and (c) onta t (CIP). The omplete solvation shell around the ion pair is notdisplayed. (Reprodu ed from Mar us and Hefter117)1.2.4 Collins' Theory of Mat hing Water A�nitiesTaking into a ount experimental �ndings a simple model for the ion-indu ed stru -turing and disordering of water has been proposed103. Collins103 proposed that ione�e ts on water stru ture ould be explained by a ompetition between ion-waterintera tions, whi h are dominated by harge density e�e ts, and water-water inter-a tions, whi h are dominated by hydrogen bonding. For example, lithium is a smallion with high harge density, so it intera ts strongly with the water dipole to strongly13

orient the water mole ules in the ion's �rst solvation shell. Larger ions having lower harge density have a lower tenden y to orient water in the ion's �rst solvation shell( haotropes). A ordingly, ions with high harge density have a high propensity toorder water (kosmotropes). He suggested that anions are stronger than ations atwater ordering be ause of the asymmetry of harge in a water mole ule: the negativeend of water's dipole is nearer to the enter of the water mole ule than the positiveend. Therefore, anions feel a larger ele trostati potential at the surfa e of a watermole ule than ations. The al ulations of Kalyuzhnyi et al.118 indi ate that thesolvation model of Collins yields qualitative agreement with the experimental data.

-60 -40 -20 0 20 40 60

-15

-10

-5

0

5

10

q (k

cal m

ol-1)

WA-- WC+ (kcal mol-1)

CHAOTROPE-KOSMOTROPE

KOSMOTROPE-CHAOTROPE

KOSMOTROPE-KOSMOTROPE

CHAOTROPE-CHAOTROPE

CsF

KF

NaF

CsCl

CsI

NaClNaBr

LiCl

LiBr

LiI

Figure 1.6: Relationship between the standard heat of solution of a rystalline alkali halide (atin�nite dilution) in kcal ·mol−1 and the di�eren e between the absolute heats of hydration of the orresponding gaseous anion and ation, also in kcal ·mol−1. The ions are assi�ed as haotropes(weakly hydrated) or kosmotropes (strongly hydrated). The enthalpy of solution of haotrope- haotrope and kosmotrope-kosmotrope salts is positive (takes up heat), whereas the enthalpy ofsolution of haotrope-kosmotrope and kosmotrope- haotrope salts is either negative (gives o� heat)or positive (takes up heat). Figures reprodu ed from Collins101.Collins' law of mat hing water a�nities states that oppositely harged ions in14

free solution form inner sphere ion pairs spontaneously only when they have equalwater a�nities103. Experimental observations of the systemati dependen e of theheats of solution of simple alkali halides on the water a�nity of the individual ions(absolute free energies of hydration) and the dependen e of the solubilities of thealkali halides on ion size ontributed to the realization that a simple law ontrolledthe tenden y of ions of opposite harge to form inner sphere ion pairs. In Figure 1.6,the enthalpy of solution is plotted on the verti al axis: those salts learly above theline at 0 produ e old solutions upon dissolution; those salts learly below the lineat 0 produ e hot solutions upon dissolution. Plotted on the horizontal axis is thedi�eren e in absolute free energies of hydration (water a�nity) of the onstituentions of the salt. We see that when the onstituent ions of a salt are mat hed in watera�nity (kosmotrope-kosmotrope and haotrope- haotrope salts), old solutions areprodu ed, suggesting that no strong intera tions with water have o urred (whi hwould release heat) and that the oppositely harged ions of the dissolved salt tendto stay together. This is to be expe ted: the point harge at the enter of a (small)kosmotropi ion an get loser to the point harge at the enter of an oppositely harged (small) kosmotropi ion than it an to the point harge at the enter ofthe oppositely harged portion of a medium size zwitterion (water mole ule); and,the point harges at the enters of the two harges on the medium size zwitterions an get loser to the harges on other water mole ules than it an to the point harge at the enter of a (large) haotrope. In ontrast, when the onstituent ionsare mismat hed in water a�nity (kosmotrope- haotrope and haotrope-kosmotropesalts), hot solutions are often produ ed, suggesting that a strong intera tion ofthe small ion with water has o urred and that the oppositely harged ions of thedissolved salt have separated. This is also to be expe ted, sin e the point hargeat the enter of a (small) kosmotropi ion an get loser to the point harge at the enter of the oppositely harged portion of a medium size zwitterion than to the point harge at the enter of the oppositely harged (large) haotrope. The requirementof a haotrope-kosmotrope or kosmotrope- haotrope salt for an exothermi heat ofsolution is a ne essary but not su� ient ondition sin e when su h a salt is dissolvedthe kosmotropi ion will generate heat as it goes from a (large) haotropi partner15

to a (medium size zwitterioni ) water mole ule, and the haotropi ion will take upheat as it goes from a (small) kosmotropi partner to a medium size zwitterioni )water mole ule. The net e�e t an be exothermi or endothermi .+

+

+

+

+

_

_

_

_

_+

+

+

Small-SmallSmall Small

Small

Big Big Big-Big

Big

X

Figure 1.7: Ion size ontrols the tenden y of oppositely harged ions to form inner sphere ionpairs. Small ions of opposite sign spontaneously form inner sphere ion pairs in aqueous solution;large ions of opposite harge spontaneously form inner sphere ion pairs in aqueous solution; andmismat hed ions of opposite harge do not spontaneously form inner sphere ion pairs in aqueoussolution.This is s hemati ally represented in Figure 1.7. Small ions of opposite hargewill tend to ome together be ause the point harges at their enters an get loserto ea h other than with the point harges at the enters of the medium size watermole ules. Large ions of opposite harge will ome together be ause the releasedwater mole ules an form stronger medium - medium intera tions. And (small)kosmotropi ions will not spontaneously dehydrate to form an inner sphere ion pairwith an oppositely harged (large) haotropi ion be ause the point harge at the enter of the kosmotropi ion an get loser to the point harge at the enter of theoppositely harged portion of a medium size zwitterions than to the point hargeat the enter of an oppositely harged (large) haotrope. Thus, it an be on ludedthat oppositely harged ions in free solution spontaneously form inner sphere ionpairs only when they have equal water a�nities.16

1.2.5 Ion-Spe i� E�e ts in Colloidal SystemsIon-water intera tions are important throughout biology and hemistry. Ions af-fe t the onformations and a tivities of proteins and nu lei a ids119�123. Ion om-plexation in ells is ru ial for the a tivities of biomole ules su h as enzymes anddrugs124,125. Ions regulate the ele trostati potentials, ondu tan es, and permeabil-ities of ell membranes126,127, the stru tures of mi elles, and the hydrophobi e�e t,whi h drives partitioning, permeation, and folding as well as binding pro esses107,128.In hemistry, ions a�e t the rates of hemi al rea tions129,130, rates of gelation131(widely used in food appli ations), ion-ex hange me hanisms132 (widely used for hemi al separations), and the expansion and ontra tion of lays, responsible forenvironmental pro esses su h as mudslides133,134.It has been observed that in asso iation olloids, the hanges in the balan e offor es ontrolling the aggregate stru ture are re�e ted in the hanges in interfa ial on entrations of water and other omponents135. In surfa tant systems, this hangeis re�e ted in the di�erent self-aggregation morphologies. This is espe ially true for atanioni mixtures, due to their enhan ed sensitivity to outside parameters. Forthis reason, atanioni surfa tant systems have been hosen to elu idate the ion-spe i� behavior and role of the surfa tant headgroups in olloidal hemistry.

17

18

Part ISalt-Indu ed Morphologi alTransitions in Non-EquimolarCatanioni Systems

19

Chapter 2Blastulae Aggregates: SpontaneousFormation of New Catanioni Superstru tures2.1 Abstra tThe transition of ioni mi elles to vesi les upon the addition of salt was explored.The atanioni surfa tant solution was omprised of sodium dode ylsulfate (SDS)and dode yltrimethylammoniumbromide (DTAB) with an ex ess of SDS. The hangein aggregate size an be a ommodated by the in rease of ounterion binding and onsequent dehydration of the surfa tant headgroups. A new type of intermediatestru ture was found: a symmetri ally shaped spheri al super-stru ture, whi h wenamed blastulae vesi le. In ontrast to known raspberry-like or egg- arton stru -tures, we believe that harge �u tuations within the bilayers are responsible forthis spontaneous super-aggregation to o ur in the presen e of only a small amountof sodium hloride. A possible me hanism for the observed pattern formation isproposed. 21

2.2 Introdu tionIt is known for a long time that double hain surfa tants an spontaneously formvesi les43. A similar phenomenon an be observed with mixtures of ationi andanioni surfa tants (here alled atanioni s). These systems display a wide varietyof phase behavior and stru tures su h as mi elles, vesi les, dis s and folded bilay-ers an be observed24,136,137. Re ently, new self-assembled stru tures, su h as onionphases and i osahedra were found in `true' atanioni solutions with no other ionsthan the surfa tant mole ules138,139. Spontaneously formed vesi les are of appli a-tional interest, espe ially in atanioni systems, sin e they an be tailored at willby varying the anioni / ationi surfa tant ratio, the size of the hain length or thenature of the polar heads26.For dilute systems, lo al aggregate urvature determined by geometri al onstraintsembodied in a surfa tant pa king parameter P = v/(lmaxa), where v and lmax are thevolume and length of the hydrophobi part, respe tively, and a the area per mole uleat the interfa e (headgroup) is a onvenient variable that hara terizes phase dia-grams15. A ne essary ondition for the formation of vesi les from either single ormixed surfa tants an be shown to be that the pa king parameter is 1/2 < P < 1.For a single surfa tant that ondition an be satis�ed by hoosing a double hainedsurfa tant, e.g. didode yldimethylammonium bromide (DDAB); for mixed surfa -tant systems a pseudo-double hain surfa tant is obtained by ion-pair formationbetween an anioni and a ationi surfa tant18,45.However, this ondition is not su� ient. While the type of aggregate that formswith ioni surfa tants an be broadly understood in terms of a balan e betweenfor es due to the pa king properties of surfa tant tails and those due to double-layerele trostati intera tions, onditions for the formation of single or few walled vesi lesare very subtle16,140,141. For e�e ts of global pa king onstraints and inter-aggregateintera tions, see André et al.142. As the surfa tant parameter (intrinsi lo al ur-vature) varies in the range 1/2 < P < 1 by e.g. varying head group area via saltaddition or temperature or mixing hain lengths, then, if the surfa tant hains are�exible, the system forms vesi les that grow as P in reases to form a lamellar phase22

at P = 1.If the hains are not �exible, the system at �rst forms multiwalled vesi les, thenvesi les, then lamellar phases again. At P = 1 this urious phenomenon is due topa king onstraints that emerge be ause the interior and exterior surfa tants of a urved bilayer experien e very di�erent onstraints. Depending on those hain andheadgroup onstraints the system forms ubi phases at P ≈ 1. These are phasesof zero average urvature but varying Gaussian urvature. Again, with in rease insurfa tant on entration equivalent to in reased repulsion between aggregates, thesystem an form equilibrium states of supra aggregation. In these the interior anbe mi elles or ubi phases prote ted by few walled vesi les.The appearan e of so many di�erent stru tures is known for a very long time20,45,143�149.But apart from the system of double hained didode yldimethylammonium saltswith di�erent ounterions studied by Ninham and Evans, they have been little ex-plored.Vesi les from atanioni systems are easily prepared. There is an expe tation thatthey might also be used as vehi les for ontrolled delivery of drugs62,150,151 or astemplates for the synthesis of hollow parti les150�152. One advantage of atanioni vesi les as ompared with more robust genuinely double hained surfa tants is theirgreater sensitivity to parameters su h as temperature153 or the presen e of salts47used to indu e transitions from vesi les to mi elles or to pre ipitation. Of parti ularinterest is the dire t transition from mi elles to vesi les. Su h a phenomenon o�ersin prin iple an easy way of en apsulating a tive agents by dissolving them in themi ellar phase prior to vesi le formation. Mi elle-to-vesi le transition was alreadyobserved when diluting a mi ellar solution with water26,154,155, hanging the anioni / ationi surfa tant ratio156�158, in reasing temperature159, or upon the addition oforgani additives160 and salt161�163.The e�e t of ioni strength on atanioni systems was previously studied experi-mentally by Brasher et al.47. Their results show that the addition of monovalentsalt hanges the phase behavior and aggregate properties of mixed surfa tant solu-tions. Theoreti ally the e�e t of salt on the atanioni s was des ribed by Yuet etal.59, however, due to the onstri tions of the model (smeared surfa e harges and23

point-sized ions), they ould not reveal the spe i� ity of di�erent ions.In the present hapter, we explore salt-mediated transition of mi elles to vesi lesin a well-studied system47,48,164. We are on erned with the in�uen e of salts on a atanioni system omposed of sodium dode ylsulfate (SDS) and dode yltrimethy-lammonium bromide (DTAB) in aqueous solution, with an ex ess of anioni surfa -tant. In reasing amounts of sodium hloride was su essively added to a solution ofmixed SDS / DTAB mi elles. The mi ellar system was �rst hara terized by rheo-metri measurements and ryo-TEM. We report on the parti ular morphologies thatarise during the salt-indu ed mi elle to vesi le transition. The formation of irregu-lar onvex- on ave patterns and a se ondary self-assembly of vesi le-like stru turesupon the addition of sodium hloride is presented. In order to study this transitionin detail, the on entration of salt in the system was in reased in small in rementsand the e�e t was studied using dynami light s attering, ryo- and freeze-fra turetransmission ele tron mi ros opy. Two di�erent mi ros opy te hniques were used inorder to ex lude the artifa ts that might arise during the preparation of samples.FF-TEM provides a dire t visualization of the three-dimensional stru ture of parti- les. The fra ture follows the path of least resistan e, and in olloidal dispersions,the fra ture surfa e propagates along the interfa e of two phases. This makes FF-TEM ideal to study membrane surfa es. To observe whether the vesi les are losedand if the membranes are inta t ryo-TEM was employed. Furthermore, ryo-TEMis very appropriate to study mi ellar solutions.2.3 Experimental Pro eduresMaterials Sodium dode yl sulphate (SDS) (purity: 99%) and sodium hloridewere pur hased from Mer k, Germany. The ationi surfa tant used was 99% do-de yltrimethylammonium bromide (DTAB) pur hased from Aldri h, Germany. All hemi als mentioned above were used as re eived without further puri� ation. Mil-lipore water was used as solvent in all ases.24

Sample Preparation Surfa tant sto k solutions were prepared by dissolving weighedamounts of dried substan es in Millipore water. The solutions were then left for 24hours to equilibrate at 25◦C. The atanioni solutions were prepared by mixing thesurfa tant sto k solutions to obtain a �xed anioni / ationi surfa tant mass ratioof 70/30 (this orresponds to a molar ratio of about 2.5/1). The total surfa tant on entration was kept at 1 wt.% at all times. Salts were added to the mi ellarsolution at in reasing on entrations. The solutions were then stirred and left toequilibrate for a week at 25◦C before making measurements.Dynami Light S attering (DLS) Measurements Parti le size analysis wasperformed using a Zetasizer 3000 PCS (Malvern Instruments Ltd., England), equippedwith a 5 mW helium neon laser with a wavelength output of 633 nm. The s atteringangle was 90◦ and the intensity auto orrelation fun tions were analyzed using theCONTIN software. All measurements were performed at 25◦C.Rheology Rheologi al experiments were performed on a Brook�eld DV - III+ rate ontrolled rheometer. A one-and-plate geometry of 48 mm diameter and with a 0.8deg one angle was used (spindle model CP - 40).Cryo-Transmission Ele tron Mi ros opy ( ryo-TEM) Spe imens for ryo-TEM were prepared by pla ing a small drop ( a. 4µl) of the sample on a holey arbongrid. Immediately after blotting with �lter �lm to obtain a thin liquid �lm over thegrid, the sample is plunged into liquid ethane (at its melting temperature). The vit-ri�ed �lm is then transferred under liquid nitrogen to the ele tron mi ros ope. Thegrid was examined with a Zeiss EM922 EF Transmission Ele tron Mi ros ope (ZeissNTS mbH, Oberko hen, Germany). Examinations were arried out at temperaturesaround 90 K. The TEM was operated at an a eleration voltage of 200 kV. Zero-loss�ltered images (DE = 0 eV) were taken under redu ed dose onditions (100 - 1000e/nm2). Images were registered digitally by a bottom mounted CCD amera system(Ultras an 1000, Gatan, Muni h, Germany) ombined and pro essed with a digitalimaging pro essing system (Digital Mi rograph 3.9 for GMS 1.4, Gatan, Muni h,Germany). 25

Freeze-Fra ture Ele tron Mi ros opy Samples used for ryo-fra ture were ryoprote ted by 30% gly erol and frozen in liquid N2. Freeze-fra ture was per-formed in a Balzers (Balzers, Switzerland) apparatus at −150◦C under a va uumof 10−6 Torrs. Metalli repli as were obtained by Pt and arbon shadowing of fra -ture surfa es. The repli a were examined and photographed with a Philips CM 12transmission ele tron mi ros ope.2.4 Results2.4.1 Chara terization of SDS / DTAB Mi ellar SolutionThe system under study is a well-known mixture of ationi and anioni single- hainsurfa tants. The total surfa tant on entration (1 wt.%; ≈ 33 mM) and the anioni / ationi molar ratio (2.5 / 1) remained onstant throughout all the experiments.The initial sample was olorless and isotropi , orresponding to the mi ellar regionof the phase diagram (Figure 2.1).

Figure 2.1: S hemati ternary phase diagram of the SDS / DTAB system at 25◦C (the bla k ross shows the starting point (referen e sample), to whi h sodium hloride was added)DLS measurements on�rm a mi ellar solution indi ating a hydrodynami radius(RH) of 10 nm and a relatively high polydispersity index (0.27). Figure 2.2 (left)shows a ryo-TEM image of our referen e solution (without added salt), exhibitingvery long rod- or ribbon-like mi elles, in equilibrium with spheri al mi elles (hen e26

explaining the high polydispersity). Long rod-like mi elles have already been ob-served by SANS measurements in systems similar to ours165. Results from rheometryexperiments show that the vis osity de reases with applied strain rate (Figure 2.2(right)) therefore exhibiting properties of non-Newtonian shear-thinning �uids. Thiskind of behavior is ommon for solutions ontaining large non-spheri al mole ules ina solvent with smaller mole ules. It is generally supposed that the large mole ular hains tumble at random and a�e t large volumes of �uid under low shear, but thatthey gradually align themselves in the dire tion of in reasing shear and produ e lessresistan e. This behavior on�rms the presen e of rod-like mi elles. No enthalpy hange ould be dete ted by di�erential s anning alorimetry so that no information ould be dedu ed about possible phase transitions o urring in the system between10 and 80◦C166�168. Probably, the amount of surfa tant was too low for su h adete tion with our equipment.

200 400 600 800 10003

4

5

6

7

8

9

10

11

12

13

/ P

as

/ s-1Figure 2.2: Left: Cryo-TEM image of a SDS/DTAB aqueous solution at the molar ratio of 2.5/1and a total surfa tant on entration of 1 wt.% (reprodu ed from ref.169); right: vis osity of thesame sample as a fun tion of shear rate.2.4.2 Salt-Indu ed Mi elle-to-Vesi le TransitionUpon the addition of sodium hloride, the solutions exhibited a transition from a olorless to a blue solution, the blue olor being typi al of the presen e of largeobje ts. Samples with di�erent salt on entrations were analyzed by dynami light27

0.00 0.01 0.02 0.03 0.040

10

20

30

40

50

60

70

80

90

RH /

nm

cNaCl / molL-1

Figure 2.3: In rease in the mean hydrodynami radius of atanioni aggregates upon the additionof sodium hloride.s attering. The addition of hloride salts auses an in rease in average parti le sizeand a ertain turbidity of the solution (Figure 2.3). DLS indi ated a signi� antin rease in the mean hydrodynami radius (RH) of the mi elles from 10 to a. 70nm. It an be expe ted that the salt s reens the ele trostati intera tions, whi hleads to smaller headgroups and therefore a higher pa king parameter.Freeze-fra ture and ryo-TEM on�rm the formation of lamellar sheets in thesample (Figure 2.4). However, by arefully observing the ryo-TEM images one ansee that at lowest on entration of added salt (10 mM) the long rod-like mi ellespresent in the starting solution start to break-up and luster together, see Figure 2.5(left). Other images from the same solution show that these lusters seem to formsmall pie es of lamellae, whi h eventually lose to form round vesi les. The urvingof membranes is represented by the presen e of darker, sti�-looking edges, due tothe higher ele tron density in these points. Figure 2.5 (right) shows some urvedpie es of lamellar sheets as well as some omplete vesi les. The vesi les representedin ryo-TEM appear to be perforated; su h perforated vesi les have previously been28

Figure 2.4: Cryo-TEM (left) and FF-TEM (right) photographs representing the formation ofmulti-lamellar sheets upon the addition of 10mM NaCl. The arrows show regions where we anobserve the unraveling of ribbon-like mi elles into lamellar sheets. The molarity in the ase ofFF-TEM experiments orresponds to the on entration of the solutions prior to ryo-prote ting.observed in various surfa tant systems170�174. Surprisingly, FF-TEM images do not on�rm su h perforations.The addition of salt produ es dramati e�e ts dete ted by freeze-fra ture. At20 mM of NaCl large spheri al, highly undulated aggregates are observed. As men-tioned previously, FF-TEM exploits the property that surfa es fra ture along thearea of least resistan e. In the ase of vesi les this is within the bilayer. There-fore, only 3-dimensional obje ts an be observed. The size range of the parti lesis from 150 to 500 nm, see Figure 2.6 (left). This apparent polydispersity is mostlikely due to the hara terization te hnique used; the measured size of the obje tdepends on the region where the samples are fra tured. Some of the aggregates arefra tured lose to the middle; Figure 2.6 (right) shows a ring of vesi les. Sin e theaggregates pi tured in Figure 2.6 are observed in high amounts, they most likelyrepresent the same obje t, fra tured in di�erent pla es ( lose to the `poles' of theblastulae vesi les, as opposed to the middle of the vesi le). The images suggestthat the inside of these parti les are hollow and �lled with the same solvent as thesurrounding (water). Due to the similarities in appearan e we named the observedstru ture blastulae, taking the name from biologi al origin; the blastulae are an early29

Figure 2.5: Cryo-TEM photographs representing the breaking-up of ribbon-like mi elles and onsequent lustering of the pie es. Clusters of elongated aggregates are indi ated by bla k (left),whereas individual aggregates are designated by white arrows. The image on the right shows lusters of smaller aggregates (bla k arrows) and vesi les with perforated surfa es (white arrows).

Figure 2.6: FF-TEM photographs representing the formation of blastulae-like lusters uponthe addition of 20mM NaCl ut near the surfa e (left); ut through the middle (right), learlyrepresenting the solvent �lled avity. The arrow shows an individual vesi le with its own membrane.stage of embryoni development onsisting of a spheri al layer of ells surroundinga �uid-�lled avity. On the basis of the present pi tures we annot say, whether theblastula vesi le is an aggregate onsisting of one individual membrane or a lusterof elongated mi elles as observed in ryo-TEM. Both possibilities will be explored30

further in the text. However, it is interesting to note that some unilamellar indi-vidual vesi les are also present. Interestingly, the average diameter of the vesi les( a. 60 nm) is of the same size as the bulges forming the blastulae stru ture. Thismight speak for the possibility that the blastula is a tually a luster of individualsmall vesi les. These, however, are not deformed in a way that it is usually observedin aggregates. A me hanism for this type of lustering will be dis ussed later on,pointing to the similarity with spe i� -site (or ligand-re eptor) binding. It shouldbe noted that similar aggregates have been observed in another ontext175,176 andwill also be dis ussed later.

Figure 2.7: FF-TEM image of individual unilamellar vesi les upon the addition of 30 mM NaCl.The bar represents 100 nm.As more salt is added to the system the lusters begin to disaggregate. At sodium hloride on entrations of 30 mM only individual unilamellar vesi les are observedas an be seen in Figure 2.7.Finally, at NaCl on entrations of 40 - 45 mM, a loose and unstru tured, randomlypa ked aggregation of vesi les is observed, see Figure 2.8. These aggregates arenot spheri ally symmetri al and the individual vesi les forming the aggregates are31

Figure 2.8: FF-TEM photograph representing the aggregation of vesi les upon the addition of45 mM NaCl to the referen e SDS/DTAB mi ellar solution. The bar represents 200 nm.deformed.In summary, two features are new in this system: i) the appearan e of blastulaestru tures, and ii) the series of di�erent stru tures that are indu ed by salt additiononly, without any further additives.2.5 Dis ussion2.5.1 Models of the Mi elle-to-Vesi le TransitionIn the following we propose a possible me hanism for the observed pattern formation.The di�erent steps are represented in Figure 2.9.`Lamellar Model' It is well known that, as an ele trolyte is added to a mixedmi ellar solution, the e�e tive surfa e area of the surfa tant headgroups be omessmaller. This e�e t is mainly due to ounterion on entration and s reening as wellas onsequent dehydration of the neutralized heads. This e�e t favors a lamellar32

pa king (step a → b′). The formation of large multi-lamellar sheets is energeti allyunfavorable, so these start urving and onsequently forming large spheri al obje ts.A theoreti al model has been suggested by Safran et al.13, showing that in ases ofmixed surfa tants vesi les are more stable than lamellar stru tures. This is due tothe formation of surfa tant bilayers, where the two monolayers onsist of di�erentsurfa tant on entrations, whi h results in equal and opposite monolayer sponta-neous urvatures. This model is in a ordan e with our result, where the largespheri al aggregates are shown to form from mi elles via lamellar sheets. The mainfa tor governing the pa king arrangement in these phases is the degree of hydrationof the polar headgroup.As the polar headgroups are dehydrated even more by the addition of salt, the pa k-ing parameter in reases and the formation of individual vesi les is favored. In thepresent atanioni system the outer layer of the vesi les is omposed of positivelyand negatively harged groups with the latter being in ex ess. A redistribution of harges may then take pla e in the highly �exible vesi le membranes leading to lo alpat hes of positive and negative harges. The lo ally positive harges of one vesi le an intera t favorably with the negative harges of another vesi le and vi e versa(Figure 2.9, stage f). In other words, the vesi les an attra t ea h other due to the+,− intera tions among the headgroups of surfa tants belonging to di�erent vesi- les. This is somewhat similar to harge �u tuations in polarizable obje ts leadingto van der Waals intera tions. But the di�eren e is that here the ion harge dis-tribution of the headgroups �u tuates, and not the ele tron distribution. The lo alCoulomb attra tion may be strong enough to over ome the overall repulsive for e ofequally harged vesi les a ting at larger distan es. Sin e vesi les are spontaneouslyformed from lamellar sheets, they are, at this starting point, in lose enough onta tfor this to happen. In this ase only lo alized parts of the membrane (pat hes) areintera ting, similarly to spe i� site binding.As the ion strength in the solution is in reased the blastulae vesi les disaggre-gate. We suppose this is due to the fa t that the intera tions between the oppositely harged headgroups of neighboring vesi les are weaker than those within a vesi le(one must onsider the additional strong van der Waals for es between the hydropho-33

c’

e

+ NaCl + NaCl + NaCl

a b d e

+ NaCl

c

b’

f

Figure 2.9: S hemati al representation of the observed stru tures that form through in reasingsalt on entration: From the starting mi ellar solution, whi h exhibits rod-like mi elles (a) twopossibilities are presented: In the ` lustering model', the rod-like mi elles (a) start breaking-up intosmaller pie es (b), whi h start lustering together to form spheri al aggregates. In the `lamellarmodel', multi-lamellar sheets are formed upon addition of salt (b′). These start losing to formundulated giant vesi les(c′). At higher ioni strength, the formation of blastulae lusters (c) isobserved, where the outer layer is omposed of individual vesi les, whi h en lose a �uid-�lled avity. The vesi les intera t through oppositely harged headgroups of surfa tants belonging todi�erent vesi les (f ; the open and �lled ir les represent headgroups of opposite harges). As salt on entration is further in reased, the attra tion is s reened and the vesi les segregate (d). Stillhigher on entrations of salt in the system produ e a non-spe i� intera tion, where van der Waalsfor es dominate, ausing the vesi les to loosely aggregate while deforming (e).34

bi surfa tant tails). Therefore, as more salt is added and the harges on the vesi lesare s reened, the `bonds' between the vesi les are broken. The luster begins to dis-aggregate due to the net negative harge of the individual vesi les, resulting in theformation of individual unilamellar vesi les.At still higher ioni strength the ele trostati repulsions are s reened and the e�e tof van der Waals for es be omes signi� ant. A loose and unstru tured, randomlypa ked aggregation of vesi les is observed. A similar behavior (�o ulation of vesi- les) was observed upon the addition of NaCl to the mi ellar solution of dode yl-benzene sulfoni a id177.`Clustering Model' In this model, the ribbon-like mi elles start breaking up intosmaller ylindri al aggregates ( orresponding to step a → b). These ould be smallvesi les, whi h grow until retaining a radius at whi h they are most stable (smallvesi les are thermodynami ally unfavorable due to high urvature). Again theseaggregates start lustering together despite an overall net negative harge. Stepsc → e remain the same as in the `lamellar model'.2.5.2 Blastulae Vesi lesMi ros opy images learly show that the blastulae lusters are spheri al and hollow.Figure 2.6 (right) pi tures the �uid-�lled avity and the astounding symmetry of theaggregate. The interesting phenomenon is that in this luster the vesi les are notdeformed in a way that is usually observed in aggregates. Aggregation of vesi leswithout a ompanying deformation of membrane had until now been observed onlyin systems to whi h spe i� ligands and re eptors were added. In su h ase nodeformations o ur when vesi les are brought together by a spe i� site-binding(ligand-re eptor) intera tion. This is be ause not the whole vesi le surfa e is involvedin the intera tion, but only a dis rete number of onta t points on ea h surfa e. Thiskind of self-assembly of vesi les driven by ligand-re eptor oupling was reported byChiruvolu et al. and Walker et al.178�180.It is interesting to translate this on ept to blastulae formation by proposing a�u tuation of the ele tron density in the membrane as was des ribed above. Su h35

a formation of lo alized partitions of opposite harge has been previously reportedby Aranda-Espinoza et al.181. A spontaneous partitioning of positively hargedmixed bilayer vesi les in the presen e of negatively harged parti les resulting in anele trostati repulsion between oppositely harged parti les was observed. In our ase the e�e t is reversed, ele trostati attra tion of vesi les of equal net harge arefound. Re ently, the asymmetry of harge in lipid bilayers indu ed by monovalentsalts was reported. A ording to Gurtovenko182, the di�eren e in the headgrouporientation on both sides of the bilayer, oupled with salt-indu ed orientation ofwater dipoles, leads to an asymmetry in the harge-density pro�les and ele trostati potentials of the bilayer. In this report lipid vesi les with only one type of mole ulesare present. In our ase, where two oppositely harged surfa tants are present, thesalt-indu ed asymmetry may be the reason for the pat hes of the single hargedsurfa tants of opposite sign. Finally, the overall shape of the blastulae formed bythe `atta hed' vesi les an be explained by the preferential lo al aggregate urvaturedetermined by geometri al onstraints, obeying the same rules as the formation ofvesi les from surfa tants. It should be noted that the attra tion of overall equally harged obje ts is dis ussed from time to time in literature183. The present �ndingsand the model proposed here may help to understand su h phenomena.As far as oexisten e of vesi les and mi elles are on erned, intermediate stru tureshave been reported in the literature: open vesi les, mesh phases, or even pat hesand dis s56,170,184. However, to our knowledge su h symmetri ally shaped hollowstru tures as the blastulae vesi les have never been observed before in atanioni systems. One of the possible explanations is that blastulae are the diluted ounter-part of the `oyster phase' whi h has been observed for other harged bilayers in theabsen e of salt185.The rationale in the sequen e of observed shapes (as s hemati ally represented inFigure 2.9), when spontaneous urvature is varied via omposition, is given by ageneral me hanism: a o- rystallization o urs, followed by a segregation of ex essmaterial. The amount of non-stoi hiometri omponent distributes between thelatti e and the edges and thus ontrols the shape of the rystallized olloids186.36

2.5.3 The O urren e of Convex-Con ave Patterns in Biolog-i al SystemsExperimentally, the o urren e of onvex- on ave deformations of high regularity,similar to our blastulae, was �rst observed in a study of lipid extra ts from ba terialmembranes187,188. The egg arton pattern is mostly found in omplex lipid mixturesof biologi al origin su h as Streptomy es hygros opi us and brain sphingomyelin189.Later on, this kind of urved patterns has also been observed in a few examples witha more simple omposition su h as in vesi les of DMPC mixed with a polymerizedamphiphile with butadiene groups190, N-nervonoyl sphingosylphosphoryl holine191both in their gel state and in systems ontaining the lipid omplex soybean le ithinand poly(diallyldimethylammonium hloride)192. However, the repeating onvex- on ave patterns in these so- alled `egg arton' stru tures are found on �at sheetsand the obje ts are multilamellar vesi les with diameters largely ex eeding the diam-eters of the bulges. By ontrast, FF-TEM images of the present atanioni systemshow that the building blo ks forming the blastulae aggregate are of the same av-erage diameter as the individual vesi les present ( a. 60 nm) in solution. Theorigin of onvex- on ave bilayer deformation is believed to result from onstraintsimposed by limiting hydration of the headgroup and a frustration arising from thespontaneous urvature of the bilayers193. This e�e t had already been dis ussed byGebhardt et al.194. Undulation and the formation of so- alled egg arton stru turehad been extensively studied theoreti ally. A `hat and saddle' model was proposedby Helfri h195 to explain the existen e of orrugated membranes of biologi al origin.Fournier has shown that anisotropi in lusions an indu e spontaneous bending in�at lipid membranes, whi h attra t ea h other to form an egg arton stru ture196,197.A model al ulation based on bilayer bending elasti ity yielding disordered egg ar-ton textures was proposed by Goetz and Helfri h198. However, all these al ulationsare not appropriate to explain the pe uliar vesi le patterns experimentally observedin present work. 37

2.5.4 Raspberry Vesi lesFor the same reason the blastulae are di�erent from the so- alled raspberry vesi les.The term `raspberry vesi les' was used previously to des ribe onvex- on ave vesi lesobtained after the indu tion of osmoti sho k on giant unilamellar vesi les in di�erentphospholipid systems175. The `raspberry e�e t' is related to the de�ating of theliposome due to the volume redu tion, the onsequen e being the existen e of anex ess membrane. This ex ess membrane indu es the formation of inverted `daughtervesi les'. Similar observations have been reported by Ménager and Cabuil176 for theosmoti shrinkage of liposomes �lled with a ferro�uid and in the ase of liposomessubje ted to a gradient of glu ose194. However, the evolution of these vesi les isdi�erent from the one observed in our ase. The initial membrane undulation ofgiant vesi les seems to be a ommon step in all ases. In the aforementioned ases,the osmoti shrinkage of vesi les was shown to be reversible whi h is proof of apersistent membrane ne k onne tion between the mother and daughter vesi les. Athigh enough osmoti strength the membrane ruptured. No vesi les with individualmembranes were observed. This is not the ase in our system, where the in reaseof ioni strength promotes the formation of individual membranes, resulting in aspheri ally symmetri al luster of unilamellar vesi les, the blastulae. Higher ioni strength results in a separation of the vesi les.2.5.5 Blastulae Vesi les: A General Trend in Catanioni Sys-tems?Finally, we would like to report that similar images have been found also in samples ontaining esium hloride and in a system omposed of sodium dode yl arboxylateand DTAB in the presen e of sodium hloride ( . f. Figure 2.10). Further experi-ments and theoreti al insights are ne essary in order to larify the exa t me hanismof formation of su h stru tures and the reason for this intermediate stru ture to befound only in atanioni systems ontaining salt.38

Figure 2.10: Blastulae vesi les in an SDS / DTAB system with the addition of 20 mM CsCl.2.6 Con lusionsWe have presented a way to make hollow regularly shaped stru tures by spontaneousse ondary self-assembly of vesi les without additives ex ept salt. These spheri al,symmetri al aggregates of individual vesi les were never observed previously, evenin systems with ligand-re eptor binding. A me hanism of formation for this typeof super-stru tures was proposed, showing the importan e of harge �u tuation.Finally, su h blastulae aggregates ould be onsidered as an intermediate step in theformation of unilamellar vesi les from bilayers in non-equimolar atanioni systems.

39

40

Chapter 3Spe i� Alkali Cation E�e ts in theTransition from Mi elles to Vesi lesThrough Salt Addition3.1 Abstra tA transition from mi elles to vesi les is reported when salts are added to atan-ioni mi ellar solutions omposed of sodium dode anoate (SL) / dode yltrimethy-lammonium bromide (DTAB) and sodium dode ylsulfate (SDS) / DTAB, with anex ess of the anioni omponent. The ounterion binding and in rease in aggregatesize was monitored by mass spe trometry, rheology and dynami light s atteringmeasurements, whereas the vesi les were hara terized by freeze-fra tion and ryo-transmission mi ros opy experiments. The e�e t of ounterions on the formation ofvesi les in both systems was studied and ompared to evaluate the role of the surfa -tant headgroups on the ounterion spe i� ity. The hange in aggregate size an bea ommodated by the in rease of ounterion binding and onsequent dehydration ofthe surfa tant headgroups. A lassi� ation of the ations ould be made a ordingto their ability to in rease the measured hydrodynami radii. It was observed that,if the sulfate headgroup of the anioni surfa tant is repla ed by a arboxyli group,the order of the ions was reversed. 41

3.2 Introdu tionThe investigation of spe i� ion e�e ts has engaged resear hers for de ades. Despitethat, ion properties and their intera tions with other mole ules are still not under-stood in detail, and we are far away from being able to predi t their behavior. Amajor di� ulty in the study of salts presents the fa t that many phenomena involvethe a tion of ations and anions of the ele trolyte.Mole ular self-assembly in surfa tant systems is largely dependent on the numberof water mole ules surrounding the headgroups. When ions are added to a solution,they dehydrate the surfa tant headgroups3,59. This auses a de rease in the valuea (e�e tive area per mole ule at the interfa e) and onsequently an in rease of thestru tural pa king parameter P , whi h may result in the formation of vesi les. Thee�e t of salts on a harged system an di�er mu h depending on their kosmotropi or haotropi hara ter. It is well known that small ions of high harge density (e.g.sulfate, arboxylate, sodium) are strongly hydrated (kosmotropes), whereas largemonovalent ions of low harge density (e.g. iodide, potassium) are weakly hydrated( haotropes)101,199. Ea h salt is therefore expe ted to have an individual in�uen eon vesi le formation, whether it tends to adsorb at the interfa e between mi elle andwater or remains strongly hydrated in the bulk.Collins' on ept of `mat hing water a�nities'101 provides us with a simple modelof spe i� ion-ion intera tions. From various experimental results (for an extensivereview see ref.200) Collins on luded that the dominant spe i� for es on ions ofthe same valen y in water are short-range for es of hemi al nature and that �thelong range ele trostati for es generated by simple ions in water are weak relativeto the strength of water-water intera tions� 200. Therefore, onta t ion-pair forma-tion is a tually dominated by hydration-dehydration. A good agreement was alsofound with re ent al ulations of water pairing, where expli it water mole ules weremodeled201. It is hallenging to see if this simple on ept holds also for spe i� ion-headgroup intera tions.A salt indu ed mi elle-to-vesi le transition was studied in two atanioni systems:sodium dode anoate (SL) / dode yltrimethylammoniumbromide (DTAB) and sodium42

dode ylsulfate (SDS) / DTAB. An ex ess of the anioni surfa tant was present inboth systems. The two systems di�ered only in the headgroup of the anioni sur-fa tant in order to elu idate the fa t that not the types of the ion alone, but ratherthe spe i� ation-surfa tant intera tions shape the surfa e behavior. Large ationspe i� ity was found, however the series experien ed a reversal when the arboxy-late ion was ex hanged for a sulphate one. SDS / DTAB is a well-known throughlystudied system47,48, whereas, by using a fatty a id based surfa tant, the system re-sembled more to a biologi al membrane than a system ontaining SDS; it had beendetermined that ellular membranes onsist of roughly 2 to 5 % of free long- hain arboxyli a ids. The atanioni surfa tant trimethylammonium headgroup, in om-parison, resembles the holine group that is often present in membranes202,203. Thee�e t of di�erent anions and ations on the mi ellar solutions was studied by phasediagrams, rheology, dynami light s attering and mass spe trometry, whereas thevesi les were hara terized by freeze-fra ture and ryo-TEM imaging. The me ha-nism of mi elle-to-vesi le transition was investigated and ompared.3.3 Experimental Pro eduresMaterials The surfa tants, sodium dode yl sulfate (SDS; Mer k, Germany; assay> 99%), sodium dode anoate (SL; Sigma, Germany; grade: 99-100%) and dode- yltrimethylammonium bromide (DTAB; Mer k, Germany; assay > 99%) were usedas re eived. All sodium and hloride salts used in the experiments were supplied byMer k, Germany. Millipore water was used as solvent in all ases.Sample Preparation Surfa tant sto k solutions were prepared by dissolving weighedamounts of dried substan es in Millipore water. The solutions were then left for 24hours to equilibrate at 25◦C. The anioni -ri h region of the phase diagrams wasused in both ases. The atanioni solutions were prepared by mixing the surfa -tant sto k solutions to obtain a �xed anioni / ationi surfa tant mass ratio: 60/40(this orresponds to a molar ratio of about 2/1) and 70/30 (molar ratio 2.5/1) wasused for SL / DTAB and SDS / DTAB mixtures, respe tively. The starting ratio43

was determined from phase diagrams (see Figures 2.1 and 3.1). The total surfa tant on entration was kept at 1 wt.% at all times. Salts were added to the mi ellarsolution at in reasing on entrations up to 50 mM. The solutions were then stirredand left to equilibrate for a day at 25◦C before making measurements.Ele trospray Mass Spe trometry (ES-MS) Cation a�nities for the vesi ularinterfa e/ arboxylate group were determined by ele trospray mass spe trometry.ES-MS was arried out using a Thermoquest Finnigan TSQ 7000 (San Jose, CA,USA) with a triple stage quadrupole mass spe trometer. The solutions were sprayedthrough a stainless steel apillary held at 4 kV, generating multiple harged ions.Data were olle ted using the X alibur software.Other Methods Cryo- and freeze-fra ture transmission ele tron mi ros opy, dy-nami light s attering and rheology were performed as des ribed in Chapter 2.3.4 Results3.4.1 Phase DiagramsThe addition of salts to the atanioni mixtures at a ertain ratio of the surfa tantsindu ed an easily observable aggregation, due to a formation of a bluish olor or tur-bidity. Salts were added to samples from all parts of the phase diagrams, howeverhomogeneous vesi ular systems upon salt addition were observed only at a ertainanioni / ationi surfa tant ratio (see Phase Diagrams, Figures 2.1 and 3.1). Forthis reason, the starting solution was taken from the mi ellar region of the phasediagram, a little way from the equimolarity line, with an ex ess of the anioni spe ies(with a mass ratio of 60/40 and 70/30 for SL / DTAB and SDS / DTAB mixtures,respe tively). 44

0.00 0.25 0.50 0.75 1.00

0.00

0.25

0.50

0.75

1.000.00

0.25

0.50

0.75

1.00

water4% DTAB

4% SL

equimolarity

I+

I_

Figure 3.1: S hemati ternary phase diagram of the SL / DTAB system (right) at 25◦C (the bla karrow shows the starting point (referen e sample), to whi h di�erent salts were added). Figurereprodu ed with permission from the do toral dissertation of A. Renon ourt.3.4.2 Counterion E�e tsThe e�e t of various ations was studied by varying the on entration of di�erentalkali hloride salts (LiCl, NaCl, KCl, CsCl). As the on entration of salt in thesystem was in reased, the solutions exhibited a transition from a olorless to a bluesolution, the blue olor being typi al of the presen e of large obje ts. This transition ould be observed for all salts, however, the on entration at whi h vesi les appearedwas strongly dependent on the nature of the ation. Further in rease of added salt on entration �nally led to a phase separation, with one of the two phases beingblue and isotropi and the other turbid.Samples with di�erent salt on entrations were �rst analyzed by dynami light s at-tering. The addition of hloride salts auses an in rease in average parti le size anda ertain turbidity of the solution (Figure 3.2). DLS indi ated a signi� ant in reasein the mean hydrodynami radius (RH) of the mi elles from 10 up to a.170 nm. Inagreement with our visual observations the rate of in rease of the measured hydro-dynami radius was di�erent for ea h hloride salt. This highlights strong ationspe i� ity. 45

For instan e, in the SL / DTAB system (Figure 3.2, left) a small amount of LiCl (<10 mM) was su� ient to dramati ally in rease the hydrodynami radius (over 100nm), whereas with CsCl, on entrations higher than 60 mM were needed to obtainparti les of that size. A lassi� ation of the ations an be made a ording to theirability to in rease the measured hydrodynami radius of the aggregates forming inthe SL / DTAB surfa tant mixture: Cs+ < K+ < Na+ < Li+.

0 10 20 30 40 500

50

100

150

200

RH (n

m)

cSALT (mmol/L)0 10 20 30 40 50

0

20

40

60

RH(n

m)

CSALT (mmol/L)Figure 3.2: The e�e t of various ations on the growth of the hydrodynami radii RH of the atanioni aggregates in SL / DTAB (left) and SDS / DTAB (right) systems with an ex ess ofanioni surfa tant in both systems: LiCl (△), NaCl (N), KCl (�), CsCl (�).However, if the arboxyli headgroup of the anioni surfa tant was repla ed bya sulfate group, the order in whi h the ions assist the formation of vesi les wasreversed ( ompare with Figure 3.2, right). Salts ontaining big ations having asmaller harge density (for instan e CsCl) more e� iently indu ed vesi le/aggregateformation than small highly harged and highly hydrated ations (su h as LiCl). Areversed Hofmeister series was therefore observed: Cs+ > K+ > Na+ > Li+.Cation a�nities for the vesi ular interfa e / sulphate (or arboxylate) groupwere determined by ele trospray ionization mass spe trometry. Although in thismethod the ionization pro ess takes pla e in the gas phase, the hemi al nature ofsurfa tant monomers and simple ions is the same in the liquid and in the gas phase.Therefore, the preferential ion-surfa tant intera tions an easily be noted204,205. Re-46

sults show that most of the mass signals (peaks) remain un hanged independentlyof the nature of the added salts. A typi al ES-MS spe trum of the omplete m/z re-gion (m and z are the mass and harge of an ion, respe tively) is shown in Figure 3.3.

200 300 400 500 600 700

m/z

0

20

40

60

80

100

A-

(2A + H )- -+

(2A +Na )- -+

(A + K + Cl )- - -+

(2A +K )- -+

Re

lativ

eA

bund

anc

e

Figure 3.3: ES-MS spe trum of 1 wt.% SL / DTAB mi ellar solution with 15mM NaCl; A− :dode yl arboxylate anion (199 Da); C+ : dode yltrimethylammonium ation (228 Da).Figure 3.4 represents the anion fragmentation patterns of the SL / DTAB atan-ioni mixture upon the addition of various salts: for better visibility only a partof the m/z region is presented. A loser look reveals that the position of the peakrepresenting the binding of sodium ions to the arboxylate (2A−+ Na+)− remainedun hanged. Also the size of the peak was omparable in all spe tra with the ex- eption of spe trum (A). When LiCl was added, two same-sized peaks appeared,one representing the binding of sodium, the other of lithium ions to the arboxy-late anion. This suggests that Li+ was able to ome loser to the vesi ular surfa e,repla ing a part of the sodium ions at the interfa e. The exa tly opposite was ob-served in the ase of CsCl (Figure 3.4D). The latter suggests a very small o urren eof esium ions at the vesi le interfa e (very small (2A−+ Cs+)− peak). A generalordering of the ations ould be determined from the ES-MS spe tra, with lithiumshowing the greatest a�nity for the anioni group and the other ations following:Li+ >Na+ >K+ >Cs+. As observed previously, this order exhibited a reversal if the47

surfa tant mixture ontained a sulphate headgroup (SDS).

400 500 6000

20

40

60

80

100

0

20

m/z

B400 500 600 700

0

20

40

60

80

100

0

20

m/z

(2A + H )- -+

(2A + Na )- -+

(A + C + Cl )- - -+

(2A + Li )- -+ (2A + C )

- -+

Re

lative

Ab

un

da

nc

e

ARe

lative

Ab

un

da

nc

e

(2A + H )- -+

(2A + Na )- -+

(A + C + Cl )- - -+

(2A + C )- -+

400 500 600 7000

20

40

60

80

100

0

m/z

C

Re

lative

Ab

un

da

nc

e

(2A + H )- -+

(2A + Na )- -+

(A + C + Cl )- - -+

(2A + C )- -+

(2A + K )- -+

400 500 600

m/z

0

20

40

60

80

100

400 500 6000

Re

lative

Ab

un

da

nc

e

D

(2A + H )- -+

(A + C + Cl )- - -+

(2A + Na )- -+ (2A + C )

- -+

(2A + Cs )- -+Figure 3.4: Ion binding as determined by ES-MS: addition of 15mM of (A) LiCl, (B) NaCl, (C)KCl and (D) CsCl to a SL / DTAB mi ellar solution; A− : dode yl arboxylate anion (199 Da);C+ : dode yltrimethylammonium ation (228 Da).Results from rheology experiments performed on atanioni solutions with vari-ous salts showed that the vis osity de reases with applied strain rate ( . f. Figure3.5). This behavior is ommon for solutions ontaining large non-spheri al mole ules,whi h tumble at random under low shear, but align themselves in the dire tion ofin reasing shear and produ e less resistan e as the shear rate is in reased3. Thisbehavior pointed to the presen e of rod-like mi elles in the solutions, whi h was on�rmed also by mi ros opy (see ahead). No hange in the rheologi al behavior ofthe samples was observed however as the nature of the salt is varied. Probably, the hange in the overall on entration of elongated mi elles present in the solution was48

too low for su h dete tion to be possible with our equipment.

0 50 100 150 200 250 3000

1

2

3

4

5

/ P

as

/ s-1Figure 3.5: Vis osity of 1 wt.% SL / DTAB mi ellar solution upon the addition of 15 mM LiCl(△), NaCl (N), KCl (�), CsCl (�) as a fun tion of shear rate.Homogeneous samples exhibiting a bluish olor typi al for solutions ontainingvesi les were further investigated by ryo- and FF-TEM. The images on�rmed anin rease in the size of the aggregates and high polydispersity ( ommon to non-equimolar atanioni systems). Figure 3.6(A) shows a pi ture similar to what wehave reported in the previous hapter. Not only the SDS / DTAB system, but alsothe SL / DTAB system on�rmed the presen e of long ribbon-like mi elles unravel-ing and forming sheets. It seems that the addition of salts produ ed a similar e�e tin both systems, hinting at a general me hanism of mi elle-to-vesi le transitions in atanioni systems. Figure 3.6(B) shows many half- losed and some already fully- losed vesi les.Cryo-TEM images of the SDS / DTAB system on�rm the high polydispersityin su h mixtures. Vesi les of various sizes were seen to exist in equilibrium withribbon-like mi elles.The in rease in measured RH was therefore most likely due to49

A B

Figure 3.6: Cryo-TEM photographs of a SL/DTAB aqueous solution at the molar ratio ofapproximately 2/1 and a total surfa tant on entration of 1wt.% upon the addition of 20 mM ofNaCl. The arrows in (A) show the presen e of long ribbon-like mi elles unravelling and formingsheets. These sheets then urve and form spheri al aggregates (B).the ombined e�e ts: formation of vesi les as well as the elongation of rods. Thelatter e�e t is well known in literature206,207.FF-TEM, be ause of the advantageous fra ture ourse within hydrophobi zones,provides us with the information regarding the membrane surfa e of the vesi les.A pe uliar di�eren e ould be observed. The addition of sodium ions to the twosystems assisted in the formation of very di�erent vesi les. While the sulfate group ontaining surfa tant mixture formed polyhedral (fa eted) vesi les with very sti�looking membranes, the arbooxylate system exhibited the presen e of individualunilamellar vesi les with ompletely smooth membranes (Figure 3.7). Fa eted vesi- les are present when the tails of the surfa tants are rather sti� due to rystallization.This is ommonly seen in atanioni systems. This sti�ness may also des ribe thehydration of the surfa tant mole ule.Similar behavior ould be found within one system, when the e�e t of various ations on the membrane sti�ness was ompared. The stronger preferential bindingof ions ould also be observed by the distortion of the SDS / DTAB vesi le membraneas it be ame dehydrated when Cs+ and Li+ ions were ompared (see Figure 3.8).50

200nmFigure 3.7: FF-TEM photographs representing the e�e t of 30 mM NaCl on the SL / DTAB (left)and the SDS / DTAB (right) mi ellar solution. The left image presents vesi les with ompletelysmooth membranes. When the arbooxylate group is ex hanged for a sulphate group, the vesi lesbe ome sti�-looking and fa eted.3.4.3 Co-ion E�e tsThe binding of ounterions to a mixed atanioni mi elle/solution interfa e dependson the surfa tant molar ratio1, i.e. the harge of the aggregates. Zeta potentialmeasurements of the atanioni solutions with and without salt on�rmed the pres-en e of negatively harged aggregates. Therefore, no anion spe i� ity was to beexpe ted. Samples were nevertheless he ked also in the presen e of various sodiumsalts (NaCl, NaBr, Na2CO3, NaSCN, NaOA , NaNO3). All salts proved to have asigni� ant in�uen e on the growth of the mi elles, however, no signi� ant spe i� ity ould be found for the di�erent anions (the urves overlap - . f. Figure 3.9).3.4.4 Nonioni E�e tsTwo non ioni additives, glu ose and urea, having respe tively salting-out andsalting-in e�e ts, were also added to the same mixed mi elles solutions. No e�e twhatsoever on the growth of the measured hydrodynami radius ould be noti ed.Even though glu ose and urea were added at a on entration as high as 1 mol/L,no in rease in the hydrodynami radius of the aggregates was found.51

Figure 3.8: FF-TEM images the e�e t of 45 mM CsCl (left) and LiCl (right) on the referen eSDS / DTAB mi ellar solution. The bars represent 100 nm in both ases.3.4.5 E�e ts of `Hydrophobi Ions'Above observations on alkali salts have shown that all alkali ations aused an in- rease in the measured hydrodynami radii. At high ele trolyte on entrations theele trostati repulsions are s reened and the vesi les aggregate. This e�e t has notbeen observed when so alled `hydrophobi ' ions are present. Tetramethylammo-nium hloride (TMAC) and holine hloride (ChCl) were added to mixed mi ellarsolutions of SDS / DTAB and SL / DTAB. Results obtained by DLS show only aslight in rease of the hydrodynami radius upon the addition of TMAC to a mi ellarsolutions (from 20 to 30 nm).A larger di�eren e in the ation a�nity for the aggregate surfa e ould be deter-mined by ES-MS. Figure 3.10 shows anion fragmentation patterns upon the additionof salt, tetramethylammonium hloride (for better visibility only a part of the m/zregion is presented). ES-MS spe tra hint that TMA+ was able to ome loser tothe vesi ular surfa e, repla ing a part of the sodium ions at the interfa e (Fig-ure 3.10(A)). The exa t opposite was observed in the ase of arboxylates (Figure52

0 10 20 30 40 500

25

50

75

100

125

150

175

RH [n

m]

c [mM]

0.00 0.01 0.02 0.03

20

40

60

80

RH[n

m]

c [mM]Figure 3.9: The e�e t of various anions / sodium salts on the growth of the hydrodynami radiiRH of the atanioni aggregates in SL / DTAB (left) and SDS / DTAB (right) systems: NaCl((•), NaSCN (�), NaBr (�), Na2CO3 (△), NaOA (◦) and NaNO3 (N)3.10(B)). The results obtained for holine hloride salts were omparable.FF-TEM images of the SDS / DTAB solution showed the presen e of vesi lesat low TMAC on entrations ( .f. Figure 3.11, left), similarly to the ase of alkalisalts. However, the membranes of the vesi les were more �exible and are thereforemore likely to fuse. Surprisingly, at higher on entration only small dis s and mi-

400 500 600 7000

20

40

60

80

100

Re

lative

Ab

un

da

nc

e

m/z

(2L + K )- + -

(L + C + Cl )- + - -

(2L + Na )- + -

(2L + H )- + -

(2L + TMA )- + -

B

Re

lative

Ab

un

da

nc

e

500 600 700 8000

20

40

60

80

100

m/z

(2DS + TMA )- + -

(2DS + H )- + -

(2DS + Na )- + -

(2DS + K )- + -

AFigure 3.10: Ion binding as determined by ES-MS: addition of 40 mM of TMAC and to a (A)SDS / DTAB and (B) SL / DTAB mi ellar solution. DS- : dode ylsulfate anion (265 Da); L- :dode yl arboxylate anion (199 Da); K+ : dode yltrimethylammonium ation (228 Da).53

100 nm100 nmFigure 3.11: FF-TEM photographs representing the e�e t of 20 mM (left) and 40 mM (right)TMAC on the referen e SDS/DTAB mi ellar solution. elles ould be observed (Figure 3.11, right). In the ase of SL / DTAB, however,no vesi les were observed. Small dis s and mi elles were present over the whole on entration region. The same pi ture ould be found when 50 mM of ChCl wasadded to SDS/DTAB and SL/DTAB.3.5 Dis ussion3.5.1 Aggregation Behavior of Catanioni SystemsThe main thermodynami driving for e for the asso iation of a ationi and an an-ioni surfa tant mixture is the release of ounterions from the aggregate surfa es.This results in a large entropy in rease. For these systems a non-monotoni hangein P is observed, with a pronoun ed maximum as the mixing ratio is varied1,3. Inthe system presented here vesi les ould only be formed when the anioni surfa tantis in molar ex ess ompared to the ationi surfa tant. Therefore, only this ratiowas onsidered here.The geometry of aggregates in olloidal systems is attributed to the pa king of theamphiphili mole ules. The pa king parameter is dependent on the length and vol-54

ume of the hydrophobi tail and the size of the hydrophobi head of the surfa tantmole ule. These fa tors are often expressed in a pa king parameter P = v/(lmaxa),where v and lmax are the volume and length of the hydrophobi part, respe tively,and a the area per mole ule at the interfa e. If the hydro arbon part of the surfa -tant is kept onstant and only the headgroup is varied, then the di�eren e in theaggregation behavior an be attributed solely to the properties of the polar head-group. Be ause a des ribes the e�e tive headgroup size, whi h in the ase of ioni surfa tants is largely determined by repulsive ele trostati for es, and not the ioni radius, the hange of morphology in surfa tant systems is largely dependent on thenumber of water mole ules surrounding the headgroups. It has been observed by hemi al trapping method135 that in asso iation olloids, the hanges in the balan eof for es ontrolling the aggregate stru ture are re�e ted in the hanges in inter-fa ial on entrations of water and other omponents. The ions, when added to ami ellar solution, an dehydrate the surfa tant headgroups. This auses a de reasein the value a and onsequently an in rease of the stru tural pa king parameterP . As a onsequen e, the riti al mi ellar on entration is redu ed60,208 and theaggregate morphology hanged209. The e�e ts are aused by destru tion of the hy-dration layer of the surfa tant, de reased ele trostati repulsions, and an in reased ounterion binding60. Consequently, the surfa tant monomers an be pa ked losertogether leading to a bilayer formation.3.5.2 Counterion PropertiesThe e�e t of salts on a harged system an di�er mu h depending on the salting-inor salting-out properties of their ions107,115,210. Salting-out ions are usually small,with relatively small polarizabilities. They have high ele tri �elds at short distan esand have tightly bound hydration water. NMR eviden e211 suggests that small ionsdo not show spe i� binding to mi elles and retain their mobility and their waterof hydration right up to the mi ellar `surfa e'212 although this result depend on thetype of surfa tants. Salting-in ions are usually large, with signi� ant polarizabilities.They have weak ele tri �elds and a loose hydration sheath, whi h an be easilyremoved. NMR proves that they perturb phospholipid (surfa tant) headgroups in a55

signi� ant way, probably through ion pairing213. Ea h salt is therefore expe ted tohave an individual in�uen e on vesi le formation, whether it tends to adsorb at theinterfa e between mi elle and water or remains strongly hydrated in the solution.3.5.3 Collins' `Law of Mat hing Water A�nities'Re ently, a simple `law of mat hing water a�nities' has been proposed101. It isbased on the tenden y of oppositely harged ions to spontaneously asso iate inaqueous solution. Only oppositely harged ions with mat hing absolute free energiesof hydration should spontaneously form inner sphere ion pairs. This is supposedto be due to the fa t that the strength of intera tion between the ions and thewater mole ules is orrelated to the strength with whi h the ion intera ts with otherions200. Small osmotropi ions an ome lose together forming inner sphere ionpairs without intermediate water mole ules. The same is supposed to be true for big haotropi ions, whereas when a osmotropi ion approa hes a haotropi ounterion,the ions should remain separated by at least one water mole ule. With this modelan impressive number of phenomena and properties an be des ribed.3.5.4 Counterion Sele tivity of Alkyl SulfatesIn the present ase the headgroups in ex ess are alkylsulfates. In ontrast to double- harged sulfate ions, su h headgroups have a more haotropi behavior. With thisassumption and a ording to Collins' on ept, alkylsulfates should ome in lose onta t with haotropi ions like esium, whereas lithium ions remain further away.Therefore it is expe ted that Cs+ s reen more e� iently the negative harge ex esson the aggregates than Li+, and this is pre isely what is observed, .f. Figure 3.2(right). The same ordering of ations was found in systems ontaining pure SDSmi elles214 and in the ase of another anioni surfa tant with a sulfate group, sodiumdode ylbenzenesulfonate177. To make a more quantitative analysis we ompare thesalt on entrations to the ex ess on entration of SDS ( ompared to DTAB), whi his approximately 0.015 mol/L. From Figure 3.2 (right) it an be inferred that this on entration roughly orresponds to the asymptoti limit for mi elle to vesi le56

transition in the ase of K+ and Cs+ hlorides. Therefore it may be on ludedthat these ations have a strong propensity to the sulfate groups and that sulfateheadgroup- ation pairs are formed; in this ase the ex ess negative harge of thevesi les is neutralized. This �nding is in agreement with the asso iation onstantsfound for a similar surfa tant n-o tylbenzenesulfonate154.3.5.5 Counterion Sele tivity of Alkyl CarboxylatesWe have seen that with the sulfate group the binding in reases with the de reasingsize of the hydrated alkali metal ions. The order is reversed when the headgroupis ex hanged for a arboxylate. The arboxylate headgroup exhibits water-orderingproperties and the addition of an alkyl hain should not hange the kosmotropi behavior of the headgroup. A ording to Collins' on ept, alkyl arboxylates shouldthen ome in lose onta t with kosmotropi ions like lithium, making it more possi-ble to form inner-sphere ion pairs, whereas esium ions remain further away. There-fore, it is expe ted that Li+ s reens more e� iently the negative harge ex ess onthe aggregates than Cs+, and this is pre isely what is observed in Figure 3.2 (left).The same order of ounterion-binding to arboxylate headgroups that was found inour atanioni surfa tant system are found in polyele trolyte solutions215�218, whenmeasuring the ation a�nity to ion-ex hange resins219,220, in studies of membranepotentials220, of ele trophoreti mobility of olloids221, and of ion-transport phenom-ena222. Our results are also in agreement with measurements of ounterion bindingto soap mi elles81,212,223,224 and long- hain fatty a ids222,223,225. Re ent MD simu-lations have shown that the arboxylate groups dominate the ounterion behavioreven in omplex systems su h as proteins210.3.5.6 Alkyl Sulfates vs. Alkyl CarboxylatesSimilar ounterion spe i� ity was observed in the ase of polyele trolytes. As inthe ase of surfa tants, the spe i� -ion e�e ts do not depend only on the individualproperties of the parti ipating ioni group and its ounterion ( harge, size, hargedistribution and polarizability), but also on the overall harge of the polyion, as well57

as on the possible ooperating binding sites whi h might in�uen e the binding site.Furthermore, even at high dilution a substantial number of ounterions is for edinto lose proximity to the polyion by the long-range ele trostati for es226,227, sothat there always exist a large number of ion pairs for whi h solvation e�e ts shouldbe observable. Strauss217,218 used the dilatometri method to measure the volume hanges that o urr when polyele trolyte solutions are mixed with solutions ontain-ing di�erent spe i� ally intera ting ounterions. Polyele trolytes ontaining sulfateand arboxylate headgroups exhibit opposite series of ounterion binding, similar tothe series observed in our experiments on surfa tant mi elles. Dilatometri results learly show spe i� ities depending on both the polyion and on the metal ion whi hwould not be expe ted on the basis of long-range ele trostati density of the polyion.The e�e ts of the latter are governed predominantly by the linear harge density ofthe polyion and the linear harge densities of the studied sulfonates and a rylates arethe same. It follows that the intera tions giving rise to the observed volume hangesinvolves spe i� sites on the polyions. Similar results of alkali ations to polya ry-lates and polysulfonates have been dis ussed also by other authors215,216,228�230. Thereversal of the ation series in the presen e of polya rylates was attributed to a ompetition between water of hydration and the anion for a given ation, whi h isrelated to the harge distributions, polarizabilities, and the e�e tive �eld strengthsof the ions221,222. The on ept was however not studied thoroughly.The reversion of series was observed also when measuring the ation a�nity to ion-ex hange resins219,220,231, ion-transport phenomena222, swelling of hydrogels232 and ounter-ion binding to long- hain fatty a ids212,215,223,224,233,234.3.5.7 Mole ular Dynami s SimulationsAlso MD simulations show similar results. Chialvo and Simonson235 modeled ashort- hain polystyrenesulfonate and found the intera tion with Li+ too weak to ause desolvation. Similarly, Lipar et al.230 reported the binding to polyanethole-sulfoni a id to be strongest for Cs+. Re ently, an extended omputational studyof pairing of sodium and potassium with a broad set of biologi ally relevant anionswas des ribed201,236. The monovalent sulfate and the arboxylate groups and their58

pairing with sodium vs. potassium were examined. The stru tures of the four in-vestigated onta t ion pairs, together with the ation - anion oxygen distan es areshown in Figure 3.12.

Figure 3.12: Geometries of the onta t ion pairs of a) sodium-a etate, b) potassium-a etate, )sodium-methylsulfate, and d) potassium-methylsulfate. The ations bind bidentally to two anioni oxygens. Reprodu ed from Vla hy et al.236.A ombination of ab initio al ulations with a polarizable ontinuum solventmodel and mole ular dynami s simulations were used to quantify the relative ation-anion asso iation free energies, i.e., the values of ∆∆G onne ted with repla ingpotassium with sodium in a onta t ion pair with a etate or methylsulfate anion236.These two free energy di�eren es for CH3COO− and CH3SO−

4 are −2.50 and +0.37k al/mol, respe tively. We see that while a etate strongly prefers sodium over potas-sium (by about 2.5 k al/mol), methylsulfate weakly (by roughly 0.4 k al/mol) preferspotassium over sodium. These �ndings are onsistent with our experimental results59

and also with the empiri al Law of mat hing water a�nities101.3.5.8 Generalization of the Con ept: Hofmeister Series ofHeadgroupsIn the pre eding paragraphs a omparison of two headgroups is given. However, moreheadgroups are of interest in olloidal hemistry and biology, espe ially sulfonatesand phosphates. Con erning hydrogen phosphate, both the harge density as wellas several experimental results201,214,237,238 suggest that it an be lassi�ed betweenthe arboxylate and the sulfonate headgroup. Following omputational as well asexperimental results, we propose a way to lassify headgroups in a Hofmeister likeseries, ordering them from osmotropi to haotropi headgroups: - arboxylate, -hydrogen phosphate, -sulfonate, -sulfate. (presented in Figure 3.13). With thisseries many spe i� -ion phenomena in olloidal systems an now be explained.3.5.9 The Anioni E�e tThe addition of sodium salts proved to have a signi� ant e�e t on the growth of themi elles (Figure 3.9). This in rease of parti le size is mainly due to a de rease ofthe head group repulsions be ause of ele trostati s reening. Sin e the mi elles arevery probably negatively harged, the anions will not ome into lose onta t withthe mi ellar surfa es and the onset of vesi le formation o urs at approximately thesame on entration for all sodium salts. The small di�eren es in radius size observedare likely to be due to the di�eren es in the hydration of the anions; in the ase ofstrongly hydrated Cl− larger vesi les were observed than in the presen e of otherless hydrated ions. A similar e�e t was previously observed in studies of negatively harged phospholipid vesi les239.3.5.10 The Non-Ioni E�e tThe fa t that the addition of glu ose and urea had no e�e t on the growth of themi elles (no mi elle to vesi le transition was observed) suggests that ele trostati intera tion is ne essary and is the �rst order e�e t. Hydration numbers show that60

Figure 3.13: Ordering of anioni surfa tant headgroups and the respe tive ounterions regardingtheir apabilities to form lose pairs. The green arrows represent strong intera tions ( lose ionpairs).glu ose dehydrates the headgroups as well as the salts (the hydration numbers forCs+ and glu ose are 4.3116 and 3.5240 respe tively. Even mu h higher on entra-tions of glu ose and urea (up to 1M) had no e�e t on the growth of the surfa tantaggregates. This on�rms that dehydration of the headgroups alone is not su�- ient to trigger the mi elle to vesi le transition. The reason must be sought in theele trostati s. Even if these un harged mole ules modi�ed the surrounding waterstru ture, this e�e t is not strong enough to signi� antly hange the urvature viaadsorption to the headgroups. If the mi elles were onsisted of just one type ofsurfa tant (one harge), then the addition of salt would s reen the repulsion be-tween the headgroups, ausing the surfa tant mole ules to arrange in a bilayer. Theopposite ase would be if the surfa tants were mixed in equimolar amounts. Then61

the salts would s reen the attra tion between them, most likely ausing the distan ebetween the headgroups to in rease. Sin e our mixed mi elles ontain an ex ess ofthe anioni surfa tant, we an assume an asymmetri distribution of harges (as wasalso proposed in the pre eding hapter). Therefore the addition of salt should havea similar e�e t as in the ase of `single- harged' mi elles.3.5.11 E�e ts of `Hydrophobi Ions'Ammonium ations have a salting-in hara ter and enhan e the solubility of other-wise poorly water-soluble ompounds. They are well-known as phase transfer ata-lysts. In water the ioni hara ter of these ions is dominant, whereas in hydrophobi environments the organi hara ter prevails241. However, there is a signi� ant dif-feren e between the properties of ammonium and tetramethylammonium ions. Theammonium ation behaves mu h like a big alkali ation; it is weakly hydrated, itshydration sphere being similar to the hydration sphere of the Cs+ ion.By ontrast, the tetramethylammonium ation is an intermediate ase between `sim-ple' ions and hydrotropes su h as xylene sulfate242. It seems plausible to assumethat this `hydrophobi ' ation an penetrate into the mi ellar surfa e mu h as an-ions like tosylate and sali ylate243�246 in the ase of ationi surfa tants. If this istrue, the tetramethylammonium ion an a t as a o-surfa tant, perhaps even par-tially repla ing the ationi surfa tant and perturbing lamellar stru tures - a featurethat is well known from hydrotropes247. Interestingly, ES-MS suggest that a strongbinding of the hydrophobi ions to the mixed mi elles is present only in the ase ofsulfates, but not also when a arboxylate surfa tant is present. The initial formationof the vesi les, observed in the SDS / DTAB system at low salt on entrations ouldbe solely due to the in reased ioni strength. However, this does not explain theobserved hanges in membrane �exibility. The destru tion of vesi les at higher salt on entrations suggests the in orporation of the hydrophobi ion into the membraneand a hange in the pa king of the surfa tants.Although the information obtained from ele troni mi ros opy is more or less thesame for both systems at high salt on entrations, the reason for the presen e ofonly mi elles and small dis s are most likely di�erent. In the ase of SDS, the62

strong binding of the `hydrophobi ions' should ause a disruption of the membrane,destroying the large vesi ular stru tures. On the other hand, it is likely that theTMA+ ion does not ome into lose onta t to the SL / DTAB mi ellar surfa e.Therefore, it assists aggregate growth only by hanging the properties in the bulk.As ammonium ions have salting-in properties, they do not indu e the dehydrationof the ioni surfa tant headgroups and therefore also not assist in vesi le formation.Further dis ussion on alkyl trimethylammonium groups an be found in Chapter 4.3.6 Con lusionsThrough the addition of salts a transition from rod-like mi elles to vesi les wasobserved in aqueous solutions omposed of DTAB and an ex ess of either SL orSDS. Whereas no anion spe i� ity for the added salts appeared in the formationof vesi les, the nature of the ation was found to in�uen e strongly the riti al salt on entrations around whi h mi elles turn to vesi les. In the present ase of neg-atively harged vesi les, ion-spe i� e�e ts were expe ted for ations, sin e theya umulate in high on entration in the vi inity of the vesi le. The observed ationspe i� ity followed the lassi al Hofmeister series for ation adsorption to sulfateheadgroups. When alkyl arboxylates were present in solution, the ations followeda reversed Hofmeister series. The ion spe i� ity ould be qualitatively explaineda ording to Collins' on ept of mat hing water a�nities. To this purpose, theheadgroup of an alkylsulfate had to be regarded as a haotrope and the alkyl ar-boxylate as a kosmotrope. The morphologies observed during the mi elle-to-vesi letransition are analogous to those presented in the pre eding hapter, hinting at ageneral me hanism ommon to atanioni systems.

63

64

Chapter 4Anion Spe i� ity In�uen ingMorphology in Catanioni Surfa tantMixtures with an Ex ess of Cationi Surfa tant4.1 Abstra tIn the previous hapter we reported on an ion spe i� mi elle-to-vesi le transition,when salts were added to a atanioni mi ellar solution omposed of sodium do-de yl arboxylate (sodium laureate, SL) and dode yltrimethylammonium bromide(DTAB), with an ex ess of SL. In the present hapter, we illustrate the ion spe i-� ity, when DTAB was in ex ess in the same system. In this ase, no transition tovesi les was observed, but an elongation of mi elles upon salt addition. The oun-terion binding and in rease in aggregate size was monitored by mass spe trometry,dynami light s attering measurements, and ryo-transmission mi ros opy. Theme hanism was argued by employing the ability of the ounterion to dehydrate thesurfa tant headgroup. This theory was on�rmed in phase with Collins' on ept ofmat hing water a�nities. 65

4.2 Introdu tionMixtures of surfa tants show enhan ed performan e in te hni al pro esses (e. g.detergen y, tertiary oil re overy, drug arrier systems, �otation), when omparedto pure surfa tant systems. Surfa tant mixtures for spe i� appli ations are of-ten hosen based on empiri al eviden e and experien e. However, to optimize theappli ations it is important to have a general understanding of the interplay of in-tera tions between the surfa tants in a mixed system and of the fa tors in�uen ingthe phase diagram. Therefore, it is of interest to study the self-aggregation andmi ellization of su h mixtures.When oppositely harged surfa tants are mixed, new properties appear. Aqueous atanioni mixtures exhibit a wide range of unique properties that arise from thestrong ele trostati intera tions between the oppositely harged heads. They exhibitlow riti al mi elle on entration (CMC) values and a non-monotoni hange in thesurfa tant pa king parameter (P ) as the mixing ratio is varied2. For this reason alarge number of aggregate stru tures su h as spheri al and rod-like mi elles, vesi les,lamellar phases, and pre ipitate have all been observed depending on the on en-tration, the size of the hain length or nature of the polar heads, and the ratiosof the surfa tants in solutions24,26,48,53�57. One advantage of atanioni systems as ompared with more robust genuinely double hained surfa tants are their greatersensitivity to parameters su h as temperature153 or the presen e of salts47.The e�e t of salt type on various physi o- hemi al properties of a system was �rstobserved over one hundred years ago by Franz Hofmeister who dis overed the de-penden e of protein solubility on the type of added inorgani ele trolyte114. Sin ethen, a wide range of ion-spe i� phenomena in biology, pharma y, and hemistrywas observed. They have been re ently reviewed in a spe ial issue of `Current Opin-ion in Colloids and Interfa e S ien e'115. Extensive studies have shown that the ounterion has a strong e�e t on the thermodynami s and aggregation propertiesof surfa tants248�250. Salt e�e ts on the m , mi ellar size, and degree of disso i-ation for a given surfa tant may follow a lyotropi (or Hofmeister) series251. Anion's position in the lyotropi series an be orrelated with its harge and hydrated66

radius. Depending on the harge density of the anion, it an intera t more or lessstrongly with the ationi headgroups of mi ellar surfa e. Su h a `binding' de reasesthe ele trostati repulsion between the surfa tant headgroups and hen e favors ag-gregation. This lowers for instan e the CMC and the degree of ionization of themi elles. A typi al Hofmeister series for anions is as follows: SO2−

4 < C2H3O−

2 <Cl− < NO−

3 < Br− < ClO−

4 < I− < CNS− (the positions of the NO−

3 and Br− ionsare often swit hed in the lytropi series).Single-tailed surfa tants usually form globular mi elles in aqueous solution abovetheir CMC4. An in rease in surfa tant on entration may indu e the formationof worm-like mi elles252,253. Similarly, addition of organi and inorgani ounteri-ons206,252�254, un harged ompounds like aromati hydro arbons255, or an oppositely harged surfa tant26,158 an transform spheri al mi elles into worm-like mi elles.In the previous hapter, we have reported a way to lassify headgroups in a Hofmeis-ter like series. We also were able to explain how salt-indu ed mi elle-to-vesi letransitions in the anioni -ri h regions of the phase diagrams in atanioni systemsdepend on both ion and headgroup spe i� ities. In the present hapter, this studyis extended to the ationi -ri h region. We fo us on the in�uen e of the ounterionidentity on the aggregation behavior of non-equimolar mixed surfa tant solutions, omposed of sodium dode anoate (SL) and dode yltrimethylammonium bromide(DTAB) with an ex ess of DTAB. The e�e t of di�erent anions and ations on themi ellar solutions will be shown by means of phase diagrams, ryo-TEM, mass spe -trometry and dynami light s attering. Collins' on ept of mat hing water a�ni-ties101 will be shown to be also very valuable for the omprehension of the foundseries of salt sensitivity. And �nally, the aggregation behavior will be ompared tothe one found in the anioni -ri h region of the orresponding phase diagram.4.3 Experimental Pro eduresMaterials The surfa tants, sodium dode anoate (SL; Sigma, Germany; grade: 99-100%) and dode yltrimethylammonium bromide (DTAB; Mer k, Germany; assay >99%) were used as re eived. All sodium and hloride salts used in the experiments67

were supplied by Mer k, Germany. They were also used as re eived without furtherpuri� ation. Millipore water was used as solvent in all ases.Phase Diagrams Surfa tant sto k solutions were prepared by dissolving weighedamounts of dried substan es in Millipore water. The solutions were then left for24 hours to equilibrate at 25◦C. The atanioni solutions were prepared by mixingthe surfa tant sto k solutions to obtain a �xed anioni / ationi surfa tant massratio of 30 / 70. The starting ratio was determined from phase diagrams. The totalsurfa tant on entration was kept at 1 wt.% at all times. Salts were added to themi ellar solution at in reasing on entrations. The solutions were then stirred andleft to equilibrate for a day at 25◦C before making measurements.Other Methods Ele trospray mass spe trometry, ryo-transmission ele tron mi- ros opy and dynami light s attering were performed as des ribed in the previous hapters.4.4 Results and Dis ussion4.4.1 Ion Binding in Catanioni Surfa tant MixturesIn the studied atanioni system, the ationi surfa tant DTAB was in ex ess. There-fore, it was expe ted that the variation of the anions will produ es a larger e�e t asthe variation of the type of ations. Figure 4.1 shows ele trospray ionization massspe tra of the SL / DTAB atanioni solutions (mass ratio 30 / 70) upon the addi-tion of di�erent salts. Although in this method the ionization pro ess takes pla ein the gas phase, the hemi al nature of surfa tant monomers and of simple ions isthe same as in the liquid phase. Therefore, ounterion a�nities for the surfa tant an easily be noted204,205. Be ause our surfa tant mixtures already ontained oun-terions, we added the same on entration of salts. In this way, we ould observe the ompetition of the ounterions for the surfa e of the surfa tant aggregate.Figure 4.1 represents the DTA+ fragmentation patterns upon the addition of68

m/z

100

80

60

40

20

0

100 200 300 400 500 600 700

Re

lativ

eA

bund

anc

e

K+

(2K++ SCN )

- +

(2K++ Br )

- +

(2K++ A )

- +

100

m/z

100

80

60

40

20

0200 300 400 500 600 700

(2K++ Br )

- +

(2K++ A )

- +

(2K++ NO )3

- +

(2K++ A )

- +

K+

Re

lativ

eA

bund

anc

e

100

m/z

0

80

60

40

20

200 300 400 500 600 700

Re

lativ

eA

bund

anc

e

(2K++ Br )

- +

(2K++ A )

- +

K+

A

CD

100

m/z

0

20

40

60

80

100

200 300 400 500 600 700

(2K++ Br )

- +

(2K++ A )

- +

K+

B

Re

lativ

eA

bund

anc

e

Figure 4.1: Ion binding as determined by ES-MS: addition of 15mM of (A) NaSCN, (B) NaBr,(C) NaNO3 and (D) Na2SO4 to a SL / DTAB mi ellar solution; A− : dode yl arboxylate anion(199 Da); K+ : dode yltrimethylammonium ation (228 Da); the peaks between 250 and 300 (m/z)are solvent related.various salts: for better visibility only a part of the m/z region is presented. Itturns out that most of the mass signals (peaks) remain un hanged, independentlyof the nature of the added salts.A loser look revealed that the position of the peak representing formation of ionpairs between the oppositely harged surfa tant ions (2K++ A−)+ remained un- hanged. Also the size of the peak was omparable in all spe tra. Similarly, thepeak orresponding to the binding of bromide ions to the dode yl trimethylammo-nium ation (2K++ Br−)+ was present in all spe tra, regardless of the nature ofthe added salt. The height of this peak, however, was less pronoun ed in spe trum(A). When omparing the binding of anions to the dode yl trimethylammonium ation, we observed that only the SCN− ex hange the bromide anions at the mi- ellar surfa e. NO−

3 and Cl− showed approximately the same a�nity for the alkyltrimethylammonium group, whereas in spe trum (D), when Na2SO4 was added to69

the atanioni mixture, no peak representing the binding of sulfate ions was visible.These results suggest that the strongly hydrated sulfate anion did not ome loseto the mi ellar surfa e, whereas the loosely hydrated thio yanate anion was ableto ome loser to the mi ellar surfa e, repla ing a part of the bromide ions at theinterfa e. A general ordering of the ations ould be determined from the ES-MSspe tra, with thio yanate showing the greatest a�nity for the anioni group and theother ations following: SCN− > Br− > NO−

3 > Cl− > CH3COO− > SO2−

4 . Thevariation of the ation of the salt produ ed no e�e t on the ation fragmentationpatterns, as an be expe ted in a system with an ex ess of ationi surfa tant. Them/z peaks remained un hanged for all added hloride salts.4.4.2 Phase Behavior upon Salt Addition

0.00 0.25 0.50 0.75 1.00

0.00

0.25

0.50

0.75

1.000.00

0.25

0.50

0.75

1.00

water4% DTAB

4% SL

equimolarity

I+

I_

Figure 4.2: S hemati ternary phase diagram of the SL / DTAB system at 25◦C (the arrowrepresents our starting solution (referen e sample; mass ratio 30 / 70).The in�uen e of salts on the SL / DTAB atanioni mixture was further followedby observing the phase behavior; the starting phase point in the phase diagram be-ing a mi ellar solution, . f. diagram shown in Figure 4.2. Di�erent salts a�e tedthe system in various ways. While the addition of some salts produ ed no visiblee�e t, others indu ed an easily observable aggregation (formation of a bluish olor70

or turbidity). As the salt on entration was in reased the aggregation be ame morepronoun ed. The e�e ts are reported in Table 4.1.Salt 25 mM 50 mMNaSCN bluish/turbid turbidNaBr bluish/turbid turbidNaNO3 bluish turbidNaOA lear learNa2SO4 lear learNaCl lear lear/bluishLiCl lear lear/bluishKCl lear lear/bluishCsCl lear lear/bluishCholine Cl lear learTable 4.1: Visual observations of the e�e t of various salts on the SL / DTAB (mass ratio: 30 /70) atanioni solution. The referen e sample was a lear homogeneous solution.The aggregation was he ked by light s attering; however problems were reportedby the apparatus due to the high s attering intensity and the high polydispersity ofthe solutions. Despite that we are able to on�rm an in rease in the hydrodynami radius of the parti les in a ordan e with our visual observations ( . f. Figure 4.3).We emphasize that the absolute values of the hydrodynami radii should be takenwith a grain of salt, be ause the aggregates formed in solutions are rods (as will beshown with ryo-TEM further on), and the CONTIN software used by the Zeta-sizer 3000 al ulates the radius of spheri al aggregates. However, the trend in thein rease of RH is important. The degree of mi ellar growth varied in exa tly theinverse order of the hydrated radius of the ounterion: SO2−

4 ≈ C2H3O−

2 < Cl− <NO−

3 < Br− < CNS−.Salts ontaining big anions having a weakly distributed harge (CNS−) indu edmore e� iently the mi ellar growth than the divalent sulphate (SO2−

4 ). Apparently71

0 20 40 60 80 100 120 140 1600

10

20

30

40

50

60

70

80

90

100

110

120

130

RH [n

m]

csalt [mM]Figure 4.3: The e�e t of various ations / hloride salts on the growth of the hydrodynami radii RH of the atanioni aggregates in SL / DTAB systems: NaOA (N), Na2SO4 (�), NaCl (◦),NaNO3 (�), NaBr (•) and NaSCN (△).the large hydrated radius of the sulfate ion hinders its ability to bind to the mi elle ontaining an ex ess of DTAB.4.4.3 Anion Spe i� ity in Physi o-Chemi al Properties of Alkyl-trimethylammonium SystemsThe ompetition of various anions for the binding sites on the surfa e of CTAB mi- elles was he ked by Larsen and Magid256. They found a strong binding of nitrateto the mi elle, displa ing bromide. The positions of bromide and nitrate ions areoften ex hanged in the lyotropi series. As an example, Cohen and Vassiliades257found that NO−

3 was less e�e tive than Br− at lowering the CMC of CTAB. Thelyotropi series was observed also when mi ellization of alkyltrimethylammoniumhalides, heats of ounterion binding, surfa e tension, and the thermodynami and72

mi roenvironmental properties of the mi elles were ompared256,258�261. The on-trolling fa tor in this ion-spe i� ity seems to be the distan e of losest approa h ofthe ion to the mi elle.4.4.4 In�uen e of Salt on the Aggregation Behavior of Sur-fa tantsThe geometry of aggregates in olloidal systems is attributed to the pa king ofthe amphiphili mole ules. The fa tors governing the shape of the aggregates areexpressed in its simplest form by a pa king parameter P = v/(lmaxa), whi h isdependent on the length (lmax) and volume (v) of the hydrophobi tail and the ef-fe tive size of the hydrophili headgroup (a). The area per mole ule at the interfa edepends on the hydration of the surfa tant headgroup, whi h in turn depends onthe ion harge density and the distan e to the small ounterions. Furthermore, it an be in�uen ed by the ioni strength of the solution. When ions ome lose to theheadgroups, the harges are more or less neutralized the value a (area per mole uleat the interfa e) de reases, and onsequently the stru tural pa king parameter P in- reases60. Therefore, a de rease of the riti al mi ellar on entration and a hangein the aggregate morphology an be observed60,208,209.As was mentioned above, the system under onsideration has an ex ess of DTAB.Previous studies have shown that the self-aggregation behavior of alkyl tetramethy-lammonium surfa tants is relatively independent of the hydro arbon hain length256.The phase behavior of DTAB (C12) and CTAB (C16) is therefore omparable. Theformation of rodlike mi elles in CTAB systems at high on entrations is widelyknown256,262�266. Alkylammonium halides have also been reported to exhibit a tran-sition from spheri al to rod-like mi elle shape with in reasing on entration of addedsalt. At lower ioni strength only spheri al mi elles are formed, while at salt on- entrations higher than a ertain threshold, larger rod-like mi elles are formed inequilibrium with the spheri al mi elles. The length of the rod-like mi elle in reasesstrongly with in reasing ioni strength on entration253,267�278. Ozeki, et al.277 re-ported that the sphere-rod transitions of mi elles took pla e in aqueous NaBr solu-73

tions of DTAB when the salt on entration ex eeded 1.8 M, whereas measurementsof Sudan red B in the same system indi ated that the minimum NaBr on entrationrequired to indu e the sphere-rod transition was 1.0-1.5 M. This may be due to ade rease in intrami ellar repulsion by ele tri al shielding due to ioni atmosphere.At higher salt on entrations, there is a dehydration of the spheri al mi elles due tothe salting-out e�e t of NaBr.By adding an anioni surfa tant, we lower the harge density of the previously

Figure 4.4: Left: spheri al and small rod-like mi elles in the SL / DTAB referen e solution(before the salt was added); right: signi� ant growth and networking of rods upon the addition ofsalt (25 mM NaBr). ationi mi elles. In this way, we also lower the on entration of salts ne essaryto produ e ylindri al mi elles. Similar e�e ts are observed with the addition ofanioni hydrotropes to ationi surfa tants279�281. Hydrotropes bind strongly to op-positely harged surfa tant ions and redu e the headgroup area of the surfa tantby redu ing the headgroup repulsions. Thus, they are e�e tive at promoting theelongated mi elle formation.The elongation of ylindri al mi elles was studied by ryo-transmission ele tronmi ros opy. Figure 4.4 shows that our initial mi ellar solution (referen e sample)exhibits the presen e of spheri al mi elles, as well as some short rods. An elongationof the ylindri al mi elles was observed as salts were added to the solution. These74

Figure 4.5: Cryo-TEM images representing the ylindri al growth as a fun tion of anion type;the e�e t of 25 mM NaNO3 (left) and NaBr (right).mi elles were quite �exible; small loops were observed, and the long mi elles formednetworks. This aused an in rease in the vis osity (observed at the preparation ofthe solutions) and an in rease of the turbidity of the solutions. However, the growthand the on entration in rease of the rods was salt spe i� (Figure 4.5). The addi-tion of NaBr produ ed a higher on entration of longer rods than the addition ofNaNO3. These then started networking together and forming a net-like stru ture.Surprisingly, we ould not see any formation of vesi les. Upon the addition of salts,the system transformed from spheri al mi elles and small rods to long ylindri almi elles, and eventually to pre ipitation as the salt on entration was in reased fur-ther. On the other hand, we have observed the transition from rod-like mi elles tovesi les in the anioni -ri h region of the phase diagram of this sample (when SL wasin ex ess). A similar system to ours was studied by Sierra et al.282. They inves-tigated the phase behavior of atanioni mixtures omposed of DTAB and sodiumunde anoate. They found the presen e of spheri al mi elles on the anioni ri h sideof the phase diagrams and short rods on the ationi ri h side. The rods elongatedand aggregated as the omposition neared equimolarity. No hexagonal mesophasewas dete ted and the phase separation at equimolarity was pre eded by the forma-tion of bundles of rod-like mi elles. The entanglement of long thread-like mi elles75

in CTAB mi elles upon salt addition was observed also by Aswal et al.283.4.4.5 Explaining Counterion Spe i� ity in Surfa tant Sys-temsSin e the ationi surfa tant was in ex ess in the atanioni system, and also inthe mixed mi elles, it is reasonable to suppose that anions were in average loser tothe mi ellar surfa e than ations. Consequently, the e�e t of anion variation wasmore pronoun ed than the e�e t of ation variation. But what makes the di�eren ebetween anions of the same harge? It is their harge density and polarisability, inline with the ordering from hard (high harge density) ions, su h as a etate, to soft(low harge density and often highly polarizable) ions, su h as thio yanate.Again Collins' on ept of so- alled `mat hing water a�nities' will be used to ex-plain ounterion spe i� ity. A ording to Collins' idea, soft ions should ome in lose onta t with soft ions and hard ions in lose onta t with hard ions101. By ontrast, when hard and soft ions ome together, they do not approa h so far thatthey loose their hydration sphere. Therefore, the intera tion between a hard anda soft ion should be weak in water. Numerous phenomena in physi al hemistry an be explained by this on ept. For quaternary ammonium ions, it is not sur-prising that they behave like soft, low-density ions. A ording to Collins' on ept itis therefore lear that DTA+ intera t more with softer ions than with harder ones,and this is pre isely what was found in the present study. For example, SCN− ions ome in lose onta t to the ationi headgroup and they share a ommon hydra-tion shell, whereas a etate anions will stay away from the headgroup and keeps itsown hydration shell. Consequently, the pa king parameter will in rease more, whenNaSCN is added to the mi ellar solution than when NaOA is added, and thereforethe tenden y to form rod-like mi elles is more pronoun ed in the ase of NaSCN.This e�e t was already dis ussed in detail in Chapter 3, where mole ular dynami s(MD) simulation was used to lassify surfa tant headgroups from hard to soft, in thesame spirit as the ions236. A similar ordering as was done in the previous hapterfor anioni headgroups an be done here to represent the ounterion binding to the76

alkyl ammonium headgroup surfa tant ( . f. Figure 4.6).

Figure 4.6: Ordering of anioni ounterions regarding their a�nity for the alkali ammoniumheadgroup.4.4.6 Di�erent self-aggregation behavior of atanioni sys-tems in the atanioni - and anioni -ri h regionsWhy we are able to get vesi les in the anioni -ri h region, but only thread-likemi elles in the atanioni -ri h region is a di� ult questions. This phenomenon wasalready observed by other authors, but never really explained. If the hanges inthe pa king parameter are onsidered, the values of the hydrophobi part remainthe same, only the a value di�ers. However, a (as reported in literature, e. g.referen e Sierra et al.282) does not di�er enough to explain this signi� ant di�eren ein the end pa king. One of the best explanations is the di�erent hydration behaviorof the headgroups ompared to the hydration of the ounterions. With the helpof diele tri relaxation spe tros opy Bu hner et al.284 were able to observe thatthe mi elle surfa e of the SDS surfa tant is strongly hydrophili and the adsorbed ounterions are generally separated by a layer of water mole ules. On the otherhand, is the surfa e of the alkyltrimethylammonium bromides hydrophobi , withthe bound halide ions dire tly atta hed284.77

4.5 Con lusionsA study of the in�uen e of added salt on the aggregation behavior of non-equimolarmixed surfa tant solutions was ondu ted. The system was omposed of sodiumdode anoate (SL) and dode yltrimethylammonium bromide (DTAB) with an ex essof ationi surfa tant. The addition of salts produ ed a sphero ylindri al growthof the mi elles, markedly dependent on the anion identity. The e� ien y of theanions to elongate the mi elles ould be explained by Collins' on ept of mat hingwater a�nities and the lassi� ation of the ationi surfa tant headgroup as a soft,polarizable entity.

78

Part IIIn reasing the Stability of Catanioni Systems

79

Chapter 5In�uen e of Additives and CationChain Length on the Kineti Stability of SupersaturatedCatanioni Systems5.1 Abstra tThe stability of mixed surfa tant solutions of sodium dode ylsulfate (SDS) with etyltrimethylammonium bromide (CTAB) and with dode yltrimethylammoniumbromide (DTAB) was studied as a fun tion of time. These spe i� mixtures wereshown to have a solubility temperature below that of pure surfa tant solutions in theanioni -ri h region. The stability of su h supersaturated solutions was studied withand without di�erent additives. Surfa tant mixtures without additives were shownto destabilize with time varying between 3 and 28 days, depending on the surfa tantratio. Generally, the stability of solutions in reased by in reasing the per entage ofthe anioni surfa tant. The variation of the hain length of the ationi surfa tantprodu ed a large e�e t on the stability of su h mixed surfa tant systems. The pres-en e of simple ele trolytes de reased, while the addition of middle- hain al oholsin reased its stability. Bluish solutions orresponding to a vesi ular region were81

observed at ratios lose to equimolarity in samples without salt, and in the anioni -ri h region upon the addition of middle- hain al ohols. Fluores en e and dynami light s attering measurements showed that the destabilization of the solutions is notdue to the formation of bigger aggregates, but rather due to a shift of the equilib-rium between mi elles and monomers, leading to the liberation of monomers, whi hpre ipitate. The lifetime of vesi les and mi elles ould therefore be ontrolled byvarying the omposition of the surfa tant solutions and by additives. Controlling thepre ipitation phenomena is of importan e for a large number of industrial pro esses,su h as oil / solute re overy pro esses after extra tions or hemi al rea tions.5.2 Introdu tionMixtures of ationi and anioni surfa tants ( atanioni s) are onstituted of �vetypes of spe ies, in luding two ounterions. In these systems a ompetition betweenvarious mole ular intera tions due to van der Waals, hydrophobi , ele trostati and/ or hydration for es may result in a wide variety of phase behaviors and mi rostru -tures24. Among others, atanioni mi elles, vesi les and bilayer stru tures, an bemore parti ularly found. In mixtures of anioni and ationi surfa tants a pre- ipitation zone is observed around equimolarity24,134. The pre ipitation zone anbe ontrolled by hanging the temperature or/and the ioni strength of the solu-tion47,153. In reasing temperature results in a general redu tion in the tenden yof pre ipitation, whereas the in rease of ioni strength generally shows the oppo-site e�e t. The formation of �nite stru tures in equimolar atanioni mixturesis thus limited to high temperatures. It is therefore of obvious interest to studythe parameters whi h ould extend the temperature range in whi h atanioni sur-fa tant self-aggregation an take pla e. Furthermore, ontrolling the pre ipitationand mi ellization phenomena of atanioni mixtures are of pra ti al importan e inmany industrial pro esses, su h as oil / solute re overy pro esses after extra tions or hemi al rea tions47,153,285�287, foaming pro esses and wetting time of textiles288,289.Mi elles and surfa tant monomers are in dynami equilibrium and individual surfa -tant mole ules are onstantly being ex hanged between the bulk and the mi elles.82

Furthermore, the mi elles themselves disintegrate and reassemble ontinuously290.While the pro ess of surfa tant mi ellization is primarily entropi ally driven, mi elleformation depends on the balan e of for es between the fa tors favoring mi elliza-tion (van der Waals and hydrophobi for es) and those opposing it (ele trostati repulsions and kineti energy of the mole ules)60,291,292. The relaxation times of mi- elles are dire tly related to the mi ellar stability and are mu h longer for nonioni surfa tants than for ioni surfa tants, be ause of the absen e of ioni repulsions be-tween the headgroups290. For the same reason mi elle kineti s in atanioni systemswere also found to be very slow293,294. Thus, mixed surfa tant solutions an remainsupersaturated for long periods of time before pre ipitation is omplete295. As theequilibrium states of su h systems are questionable296, the time dependent stabil-ity of ommon atanioni mixtures was studied in this paper. The e�e t of ioni strength and the in�uen e of the variation of ounterions were investigated, and alsothe in�uen e of added al ohols as well as the importan e of a di�erent hain length.5.3 Experimental Pro eduresMaterials The surfa tants, sodium dode yl sulfate (SDS: Mer k, Germany, grade:99%), dode yltrimethylammoniumbromide (DTAB: Mer k, Germany; assay > 99%)and etyltrimethylammonium bromide (CTAB; Sigma, Germany; grade: 99%) wereused as re eived. All sodium and hloride salts used in the experiments were sup-plied by Mer k, Germany. The al ohols: methanol, 1-propanol, 1-butanol, 1-de anol(Mer k, Germany, > 99%), ethanol (J. T. Baker, Holland, 99.9%), 1-pentanol(Aldri h, Germany, 99.5%), 1-hexanol (Fluka, Germany, 99%), 1-heptanol (AlfaAesar, Germany; 99%), 1-o tanol (A ros Organi s, U.S.A., 98%), and itronellol(Mer k, Germany, 98%) were also used as re eived. Millipore water was used assolvent in all ases.Determination of the solubility temperature (This measurement was per-formed by A. Arteaga during his graduate studies. Results are represented with hispermission.)The solubility temperature at di�erent time intervals was determined83

by visual observation. All systems were observed at 1 wt.% total surfa tant on- entration. The ationi / anioni mixtures at di�erent mass ratios were preparedfrom 5 wt.% surfa tant sto k solutions. The samples were �rst heated to 39◦C, then ooled down and left to equilibrate for 5 hours at 0◦C, and then the temperaturewas in reased by approximately 0.5◦C per minute. The measured solubility tem-peratures (between 0◦C and 99◦C) orrespond to the transition from pre ipitate toisotropi phase; i. e. to a mi ellar or vesi ular solution. The pro edure was repeatedafter a ertain time to analyze the kineti s of the system.Phase Diagrams Phase diagrams of the systems were onstru ted a ording tothe pro edure des ribed above. The samples had to be heated after mixing in orderto obtain a homogeneous lear solution. Di�erent surfa tant ratios exhibited di�er-ent initial solubilization temperatures, TI (more throughly explained in the Resultsse tion). The e�e t of additives was studied by adding an appropriate amount ofsalt or al ohol to the SDS / CTAB mixture. Again, in some ases the sample had tobe heated in order to obtain a lear solution. The maximum temperature to whi hthe samples were heated was 39◦C. The samples were then ooled down and keptat 20◦C. Phase diagrams were onstru ted by visual observation and by measuringtransmission using a Hita hi U-1000 spe trophotometer. The samples were also ob-served at 0◦C in order to shorten the destabilization times and to he k whether theorder of destabilization is the same at both temperatures.Counterion Binding Cation a�nities for the surfa tant aggregate interfa e weredetermined by ele trospray mass spe trometry. ES-MS was arried out using a Ther-moquest Finnigan TSQ 7000 (San Jose, CA, USA) with a triple stage quadrupolemass spe trometer. The solutions were sprayed through a stainless steel apillaryheld at 4 kV, generating multiple harged ions. Data were olle ted using the X al-ibur software. The surfa tant on entration was kept 1wt.% and the on entrationof the various hloride salts was 0.1 M in all ases.System Kineti s The evolution of the surfa tant aggregates was further followedusing �uores en e spe tros opy and dynami light s attering (DLS) measurements.84

Pyrene (Aldri h, opti al grade) was used as the external �uores en e probe to mon-itor the hange in surfa tant self-assembly. The preparation of water saturated withpyrene was as reported previously297,298. The �uores en e emission spe tra of pyrenewere re orded on a Cary E lipse �uores en e spe trophotometer (Varian, In .) at20◦C. From the spe tra, the ratio of intensities of the �rst and the third vibra-tional peak of pyrene, I1/I3, was al ulated. Parti le size analysis was performedusing a Zetasizer 3000 PCS (Malvern Instruments Ltd., England), equipped with a5 mW helium neon laser with a wavelength output of 633 nm. The s attering anglewas 90◦ and the intensity auto- orrelation fun tions were analyzed using the Continsoftware.E�e t of Al ohols on the SystemMorphology Spe imens for ryo-transmissionele tron mi ros opy ( ryo-TEM) were prepared as des ribed before169; frozen sam-ples were examined with a Zeiss EM922 EF Transmission Ele tron Mi ros ope (ZeissNTS mbH, Oberko hen, Germany), whereas the repli a were examined and pho-tographed with a Philips CM 12 transmission ele tron mi ros ope.5.4 Results and Dis ussion5.4.1 Shift of Solubility Temperature with TimeThe in�uen e of the ationi surfa tant hain length on the solubility temperaturesof atanioni mixtures was investigated by omparing mixtures of sodium dode- ylsulfate (SDS) with etyltrimethylammonium bromide (CTAB) and with dode- yltrimethylammonium bromide (DTAB). CTAB and DTAB have the same quater-nary ammonium polar headgroups, but CTAB has a longer hain (16 arbon atoms)than DTAB (12 arbon atoms). Owing to stronger hydrophobi repulsions of itshydro arbon hain, CTAB is therefore less soluble in water than DTAB. The Kra�tpoint of pure CTAB in water at 1 wt.% is onsequently higher (≈ 25◦C)299 thanthat of DTAB (< 0◦C)300. The solubility temperatures (TS) of the orresponding atanoni mixtures, i. e. DTAB / SDS and CTAB / SDS with a total surfa tant on entration of 1 wt.% and di�erent anioni / ationi surfa tant (molar) ratios85

were followed as a fun tion of time ( . f. Figures 5.1 and 5.2). Both systems dis-played the same marked di�eren es in their solubility temperatures over the wholeanioni / ationi ratios in the initial days after mixing. The TS in the ationi -ri hregions were generally higher than the TK of the pure ationi surfa tant. The sametrend was observed for only a part of the phase diagram in the anioni -ri h region.Around equimolarity, a maximum of solubility temperatures (TS) was rea hed. Thismaximum lay within the zone of pre ipitation usually observed at room tempera-ture. A bluish zone, orresponding to a vesi ular region, was lo ated in both sides ofthe entral area of insolubility301,302. It was mu h more extended in the anioni -ri hregion (approximately between 0.6 - 0.7) than in the ationi -ri h region (around0.35). The vesi ular regions ould be extended by raising the temperature. Harg-reaves and Deamer61 noted the formation of liposomes in mixtures of CTAB andSDS above ≈ 47◦C. The sequen e of di�erent phases resembled those observed insimilar atanioni systems24,49,303. Obviously, the behaviour of two single- hain sur-fa tants is mu h di�erent in a atanioni system. The pa king between the ationi and anioni surfa tants is quite di�erent from the single ioni one and ontributes toin rease the stability of the atanioni rystal, leading thus to higher solubility (andKra�t) temperatures. Interestingly, however, at a ertain ratio of anioni / ationi surfa tants a `solubility temperature depression' exists. This depression was lo atedin a range of a molar fra tion of the anioni surfa tant between 0.65 − 0.85 and ould be observed immediately after mixing. With time, the mixtures undergo a ontinuous evolution of TS toward higher temperatures. Surprisingly, the solubilitytemperature was almost onstant over time in the ationi ri h region, whereas alarge time dependen y ould be observed in the anioni ri h region, . f. Figures 5.1and 5.2. Furthermore, the shift in the solubility temperature was more pronoun edin systems ontaining a longer hain ationi surfa tant (Figure 5.2). For this reasonwe fo used our attention on the anioni part of the phase diagram of the CTAB /SDS system.

86

0

10

20

30

40

50

60

70

80

90

100

0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0

Molar fraction SDS

Te

mp

era

ture

(°C

)

Figure 5.1: Shift of solubility temperature in DTAB / SDS mixtures at di�erent surfa tant ratiosas a fun tion of time (ctot = 1wt.%): observations after (N) 1 day, (�) 4 weeks and (•) 6 weeks.

0

10

20

30

40

50

60

70

80

90

100

0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0

Molar fraction SDS

Te

mp

ert

aru

re(°

C)

Figure 5.2: Shift of solubility temperature in CTAB / SDS mixtures at di�erent surfa tant ratiosas a fun tion of time (ctot = 1wt.%): observations after (N) 1 day, (�) 4 weeks and (•) 6 weeks.Both Figures reprodu ed with permission from the do toral dissertation of A. Arteaga.87

5.4.2 Behavior of the Anioni -Ri h Region of the Phase Di-agrams Without AdditivesThe samples (with an ex ess of SDS) were prepared as des ribed in the Experi-mental Se tion and heated in order to obtain a lear homogeneous solution. Thetemperature needed an be des ribed by the equation TI(◦C) = −0.43 · X + 64.9,where TI(

◦C) is the initial solubility temperature expressed in Celsius and X is theper entage of SDS in the solution ( . f. Figure 5.3 (left)). Then the samples were leftto ool down and observed at 20◦C. The results from spe tros opy measurementsof the SDS / CTAB system at 20◦C are presented in Figure 5.3 (right).

60 65 70 75 80

30

32

34

36

38

40

T(o

C)

XSDS

0 5 10 15 20 250

20

40

60

80

100

T

rans

mis

sion

(%

)

time (day)Figure 5.3: Left: The temperature needed to obtain lear homogeneous solutions in the anioni -ri h region an be des ribed by a straight line with a slope of k = −0.43(±0.03) and n = 64.9(±2.6);right: Transmission of the SDS / CTAB atanioni system as a fun tion of time for the followingratios: 65/35 (�), 70/30 (•), 75/25 (N), 80/20 (�), and 85/15 (◦)We see that the higher the ratio of the anioni surfa tant, the more stable thesystem is. Whereas the solution with 65% and 70% of SDS were pre ipitated after10 days, solutions with 80% and 85% were still lear after 30 days. The ele tro-stati intera tions and onsequent harge neutralization between the SDS mole ulesand the oppositely harged CTAB promote long relaxation times of atanioni mi- elles. These, in turn, are a ountable for the supersaturation of our system andthe temporary solubility temperature depression observed. The anioni and ationi 88

surfa tant mole ules an be present in solution either as monomers, in orporatedinto surfa tant aggregates (mi elles, vesi les, et .) or as pre ipitate (DSCTA(s) - asalt formed from DS−CTA+ ions). Thus, there are two equilibria present in su hmixtures: monomer-mi elle and monomer-pre ipitate. Two di�erent types of phaseseparation in su h systems an be observed. Phase separation in atanioni surfa -tant mixtures may o ur due to the entanglement of rod-like mi elles, the formationof lamellar phase, or the formation of densely pa ked vesi les. In some ases, how-ever, there seems to be a ompetition between the kineti s of vesi le formation andthe kineti s of pre ipitation of ationi -anioni pairs134,304. Earlier measurements on atanioni systems have shown that pre ipitation is primarily enthalpi ally driven,while mi ellar formation is primarily entropi ally driven. This is due to the orderingof the surfa tant headgroups in the rystalline pre ipitate as a onsequen e of hargeneutralization153.In order to understand whether phase separation is a onsequen e of pre ipita-tion of solid atanioni salt or the onsequen e of mi ellar (vesi ular) aggregation,dynami light s attering and �uores en e te hniques were employed. Parti le sizemeasurements showed a de rease in the size of the parti les with time, as an beobserved in Figure 5.4. After 3 days the solution was pre ipitated and the laserbeam passing through the upper ( lear) part of the solution dete ted only verysmall parti les (RH ≈ 2.5nm).This behavior is quite di�erent from that observed in a previously studied SDS /DTAB system169. There, an in rease in turbidity was observed in a ordan e withthe in rease of the hydrodynami radius of the orresponding parti les. Althoughthe solutions were turbid, no pre ipitation took pla e, even after months of observa-tion. The behavior of the SDS / CTAB was di�erent. A pre ipitation was observedwithout any dete tion of an intermediate turbidity. The reasons for the di�erent be-haviors of the two systems are most likely due to the di�erent Kra�t temperatures ofthe two ationi surfa tants. Pre ipitation is a ommon phenomenon in atanioni surfa tant mixtures, whi h is regulated by two ompeting for es: the free energy ofthe solid rystalline state and that of the mi ellar solution305. The strong intera -tions make pa king into a rystalline latti e energeti ally favorable for surfa tant89

0 2 4 6 8 10 12 14 16 18 200

2

4

6

8

10

12

%ofpartic

les

RH

[nm]

Figure 5.4: Parti le size followed by DLS: SDS / CTAB atanioni system with a (molar) ratio65/35 after 1 day (�) and after 3 days (�).monomers, parti ularly when the hydrophobi mat h is high. It has been shownthat by ex hanging the ounterion, we are able to in�uen e the energeti state ofthe mi ellar solution. A mu h bigger variation between systems is observed in thefree energies of the rystalline state. As we an see from the phase diagrams, thesolubility of the atanioni salt is dire tly orrelated to the solubility of the surfa -tant monomers. The solubility (and Kra�t point) de reases with in reasing hainlength, and is therefore lower for the CTAB mixture in omparison to the DTABsurfa tant mixture. We an not ex lude at this point the possibility that the pre ip-itate was formed from an entanglement of rod-like mi elles or similar big parti lesas was already reported by other authors282,283. However, su h transitions take alonger time, suggesting the formation of larger parti les would be dete ted beforethe a tual pre ipitation takes pla e. For this reason, we believe our system yields90

a similarity to the SDS / DPLC (dode ylpyridinium hloride) mixture studied byStellner et al.134. In this ase, the destabilization of solutions was not due to parti legrowth (and formation of big aggregates), rather was the observed pre ipitation dueto the liberation of monomers.Pyrene �uores en e emission spe tros opy was used to further investigate the hanges in phase behavior. The low �uores en e ratio, I1/I3, of pyrene on the �rstday after preparation of the solutions showed the presen e of highly hydrophobi parti les in the solution (I1/I3 = 1.1; . f. Figure 5.5). The I1/I3 ratio in 1 week oldsolutions (ma ros opi ally the solutions were two-phased: the lower part ontainingthe pre ipitant, whereas the upper part onsisted of a lear supernatant) showedan in rease from a value of 1.1 to a value around 1.3. This value orresponds tothe polarity found for SDS mi elles306. This suggests that as the atanioni saltpre ipitated, the remaining pyrene was solubilized inside the mi elles formed by theex ess anioni surfa tant. In this way, a large amount of the surfa tant was removedfrom the solution (as pre ipitate); only the ex ess amount remained in solution. Hadall of the SDS surfa tant pre ipitated as well, a mu h higher pyrene ratio would havebeen observed (I1/I3(water) ≈ 1.8307,308), whereas, if bigger aggregates had formed,one would expe t the pyrene intensity ratio to remain un hanged. In the latter ase, the pyrene probe would be solubilized either inside rod-like mi elles or thehydrophobi interior of vesi ular membranes, whose environment mi ro-polarity isthe same as that in spheri al mi elles.A higher stability of systems when the fra tion of one of the omponents wasin reased in omparison to the other omponent was already observed by Stellneret al.134. When the SDS on entration is in reased (and the CTAB on entrationheld onstant), pre ipitate an not form due to a hange in the mi ellar (and onse-quently monomer) omposition. The monomer on entration then does not ex eedthe monomer-pre ipitate boundary (for a more detailed explanation the reader isreferred to Stellner et al.134). 91

Figure 5.5: Pyrene �uores en e emission spe trum of a SDS / CTAB atanioni system with a(molar) ratio 70/30 after (left) 1 day, (right) 7 days.5.4.3 E�e t of Additives on the Stability and the `SolubilityTemperature Depression'E�e t of Ioni Strength Addition of ele trolytes to ioni surfa tant solutionsmay modify both the inter- and intrami ellar intera tions. Therefore, the phasebehavior of surfa tant solutions is expe ted to be a�e ted signi� antly upon theaddition of salts. Sodium bromide was added to the atanioni mixture at di�erentratios (in the anioni ri h region) to study the e�e t of ioni strength on the phasestability. The observed e�e t was similar for all ratios. Generally, the in rease ofioni strength de reased the destabilization time of the sample ( . f. Figure 5.6).The higher the on entration of the added NaBr was, the shorter were the desta-bilization times. Whereas the solution with 0.5 M of NaBr was pre ipitated afteronly 2 days, the solution with 0.1 M NaBr started showing �rst signs of turbidityafter 10 days. The destabilization times are shortened down to two hours when thesolutions are kept in an i e bath. A general in rease in the tenden y of pre ipita-tion of atanioni mixtures upon the in rease of ioni strength was already observedpreviously47,153. When surfa tant monomers are salted out by the presen e of an92

ele trolyte, mi ellization is favored2. Catanioni systems exhibit di�erent behav-iors upon salt addition. This is due to the ability of mixed surfa tant systems toadjust the aggregate omposition to attain the lowest possible free energy state forgiven onditions. In atanioni systems two salt-indu ed phenomena are observed,depending on whether the equimolar47 (�true�) or non-equimolar169 atanioni sys-tems are studied. To qui kly review, if the surfa tant mi elles onsist of just onetype of surfa tant (one harge) or one of the omponents is in ex ess, then the ad-dition of salt s reens the repulsion between the headgroups, ausing the surfa tantmole ules to arrange in a bilayer. The opposite ase is true, if the surfa tants arepresent in equimolar amounts. The salts then s reen the attra tion between them,most likely ausing the distan e between the headgroups to in rease. In these ases avesi le-to-mi elle transition was observed. Sin e we are dealing with non-equimolar atanioni mixtures, where SDS is in ex ess, the addition of salt to our surfa tantmixture is expe ted to have the same e�e t as on pure SDS mi elles. Addition ofNaCl to SDS, for instan e, is known to de rease the CMC and in rease the mi ellarsize6. Furthermore, by raising the ioni strength of the solution, the harges on thesurfa tant headgroups are s reened due to a stronger ounterion binding. This, ine�e t, de reases the solubility of the surfa tant monomers and therefore assist inpre ipitation.E�e t of Cation / Anion Variations A good orrelation has been observedbetween the temperature to whi h it is required to heat the solution for ompletesolubilization to o ur (TI) and the stability of the then isotropi solution. For thisreason, the heating temperatures of solutions ontaining various salts and al oholsare reported in Table 5.1. The salt on entration was kept onstant in all ases at 0.1M. The solubility temperatures for all sodium salts are 29(±0.5)◦C. The variation ofthe anion apparently has no e�e t on the solubility of the atanioni mixture. Thiswas expe ted as the overall net harge was negative (we had an ex ess of the anioni surfa tant SDS). A big variation of the temperature was observed when di�erent hloride salts were added to the mi ellar solution. Mass spe tros opy results (Figure5.7) showed that at this relatively high on entration of added salt (0.05 M `new'93

0 5 10 15 20 250

20

40

60

80

100

Tran

smis

sion

(%)

time (day)Figure 5.6: Transmission of the SDS / CTAB atanioni system: the e�e t of ioni strength onthe stability of a SDS / CTAB solution (molar ratio 80/20) without salt (�), with 0.1 (•) and 0.5M (N) NaBr. ounterions vs. ≈ 0.016 M of `original' sodium ounterions) all ounterions were inthe lose proximity of the mixed mi elle surfa e, regardless of the fa t whether theynormally exhibit salting-in or salting-out behavior.Despite this, we saw a large di�eren e in the stability of the solutions when thesolutions ontaining LiCl or NaCl were ompared with solutions ontaining KCl orCsCl. After one week at 20◦C, the �rst two solutions were still lear and isotropi ,while the later two were already pre ipitated (see Figure 5.8). This di�eren e in TI an be attributed to the stronger binding of less hydrated potassium and esium ionsto the ex ess dode yl sulfate ions309. The use of KCl pre ipitation to remove SDSis ommonly used in biology310,311. The e�e t of `hydrophobi ' ounterions ( holineand TMA+) was found to be in the middle of both.A stronger binding of organi ounterions in omparison to sodium to dode ylsulfate mi elles was already observed using �uores en e te hniques312 showing that94

Salt TI (0.1 M) Al ohol TI (0.1 M) TI (0.2 M) TI (0.05 M)No salt 35.5 Methanol 35.5LiCl 29 Ethanol 36 36NaCl 29 n-Propanol 36 36KCl 37 n-Butanol 35 34.5CsCl 31.5 n-Pentanol 30.5 27ChCl 37 n-Hexanol 21∗ 21∗TMACl 37.5 n-Heptanol 21∗ 21∗NaBr 29.5 n-O tanol 21∗ 21∗NaSCN 29 n-De anol > 39∗∗ > 39∗∗Na2SO4 28.5 Citronellol > 39∗∗ > 39∗∗Table 5.1: E�e t of additives on the initial solubility temperature (TI); the initial sample is a70/30 mixture (with an ex ess of SDS); ∗ the samples were isotropi and slightly bluish already atroom temperature (21◦C); ∗∗the samples were not lear even at 39◦C.the binding of alkylammonium ions is dominated by their hydrophobi ity. As waspointed out previously, we are able to in�uen e the energeti state of the mi ellarsolution by ex hanging the ounterion. This an explain the small di�eren es inthe stability of the systems with di�erent ounterions. Mukerjee et al.313 haveinvestigated the binding of ounterions to dode yl sulphate anions by ondu tivity.They have reported that the Li+ ion is less �rmly atta hed to the mi elle than theNa+ ion and this one in turn less than Cs+. The lower binding therefore promotedthe formation of ionized spe ies. The gain in free energy was however not largeenough to over ome the free energy of the rystallization, and thus all of the systemseventually pre ipitated.The a�nity of spe i� ion- ounterion pairing has been previously extensivelystudied on the vesi ular169 and two-phase regions31 of a similar system. The orderingof ations was found to follow the Hofmeister series in all ases. The e�e ts of theions were found to be strongly dependent on the ability of the ounterions to form lose ion pairs with the surfa tant headgroups and the behavior was explained usingCollins' Law of Mat hing Water A�nities101. As was shown in Chapters 3 and 4,95

500 600 700 800 900

m/z

0

20

40

60

80

100(2A +Li

- + +)

500 600 700 800 9000

20

40

60

80

100

500 600 700 800 9000

20

40

60

80

100

600 700 8000

20

40

60

80

100

(2A +Na- + +

)(2A +M

- + +)

(2A +M- + +

)

(2A +Na- + +

)

(2A +H- + +

)

(2A +M- + +

)

(2A +M- + +

)

(2A +Na- + +

)

(2A +H- + +

)

(2A +H- + +

)

(2A +Na- + +

)

(2A + Cs- + +

)(2A +K

- + +)

(A +M- + - +

+Cl )

(A +M- + - +

+Cl )

(A +M- + - +

+Cl )

(A +M- + - +

+Cl )

m/z

m/zm/z

A B

C DFigure 5.7: Ion binding as determined by ES-MS: addition of 0.1 M of (A) LiCl, (B) NaCl, (C)KCl and (D) CsCl to an SDS / CTAB mi ellar solution. A− : dode ylsulphate anion (265 Da);M+ : etyltrimethylammonium ation (284 Da).the alkyl sulfate headgroup behaves like a soft, haotropi ion.The anions showed a destabilization time omparable to sodium hloride. Thevariations of the anion produ ed no hange in the stabilization times and thereforeno spe i� ity ( . f. Figure 5.9). This is most likely due to the fa t that the anionsdid not a umulate near the same- harged mi ellar surfa e, but assisted purely byin reasing the ioni strength of the medium.E�e t of al ohols It is known that al ohol penetrates in the palisade layerof mi elles314 and depresses the surfa tant CMC315�317. Nakayama et al.318 andKaneshina et al.319 have observed a depression of the CMC and the Kra�t point ofioni surfa tant in the presen e of longer hain al ohols. The Kra�t point depressionwas explained quantitatively by a model in whi h it is proposed that the melting96

Figure 5.8: E�e t of additives on the stability of an SDS / CTAB atanioni system with a(molar) ratio 70/30: (from left to right) LiCl, NaCl, KCl and CsCl (all salts 0.1 M) after 1 weekat 20◦C

Figure 5.9: E�e t of additives on the stability of an SDS / CTAB atanioni system with a(molar) ratio 70/30: (from left to right): without salt, with LiCl, NaCl, ChCl (Choline Chloride),NaSCN, Na2SO4 (all 0.1 M) after 3 hours on i e.97

point of the hydrated solid agent is depressed owing to the formation of a mixedmi elle of surfa tant and al ohol318.

Figure 5.10: E�e t of additives on the stability of an SDS / CTAB atanioni system witha (molar) ratio 70/30: (from left to right) methanol, ethanol, n-propanol, n-butanol n-pentanol,n-hexanol, n-heptanol, n-o tanol, n-de anol, itronellol (all 0.1 M) after 1 week. Solutions frommethanol to butanol are slightly turbid / pre ipitated, pentanol is lear; hexanol to o tanol arebluish and homogeneous, de anol and itronellol are turbid (due to the low al ohol solubility).The heating temperatures of solutions ontaining al ohols of various hain lengthand on entrations are reported Table 5.1. A ording to the results reported, the al- ohols an be lassi�ed into four groups: (i) short- hain al ohols (< C4), whi h haveno e�e t on the TI or on the stability of the system; (ii) pentanol, whi h de reasesTI and the solubility somewhat; (iii) middle- hain al ohols (between C6 and C8)whi h lower TI and in rease the stability substantially; and (iv) long- hain al ohols(> C10), whi h are inappropriate for use due to their low solubility ( . f. Figure5.10). Variation of on entration generally produ ed little e�e t. Middle- hain al- ohols from hexanol to o tanol have provided best results. The initial solubilitytemperature was redu ed to below room temperature (< 21◦C) and the stability ofthe solutions was in reased drasti ally. The samples were homogeneous and bluish,hinting at the presen e of larger parti les. Best e�e ts were observed when the al- ohol on entration range was kept between 0.02 and 0.2 M. Lower on entrationsprodu ed no e�e t and larger amounts of al ohol resulted in an al ohol-surfa tantgel. 98

Figure 5.11: Pyrene �uores en e emission spe trum of a SDS / CTAB atanioni system with a(molar) ratio 70/30 upon the addition of 0.05 M hexanol after (left) 1 day, (right) 1 week.Fluores en e measurements reported a presen e of hydrophobi domains in thesolution ( . f. Figure 5.11). In fa t, the polarity of the aggregates sensed bythe pyrene probe was even lower when hexanol was added to the atanioni mix-ture (I1/I3 = 1.04); lose to that found in toluene (I1/I3 = 1.03)307,308 or in 2-propanol (I1/I3 = 1.07)307. This hints at the fa t that hexanol forms mixed mi- elles/aggregates with the two surfa tants, thus redu ing the on entration of sur-fa tant monomers. Solutions with di�erent amounts of hexanol were observed atroom temperature for up to 6 months and still showed no sign of destabilization.Also the I1/I3 ratio in su h samples remained stable over time.Furthermore, solutions ontaining small amounts of hexanol (0.05 M) were keptin an i e bath for a period of up to 2 weeks and again no visible hanges appeared.After 2 weeks, the sample was warmed to 20◦C and he ked by ryo-transmissionmi ros opy. Solutions exhibited a high polydispersity. Spheri al and ribbon-likemi elles, as well as vesi les were observed ( .f. Figure 5.12). This proves that it99

Figure 5.12: Cryo-TEM photographs of a SDS / DTAB aqueous solution at a molar ratio of70/30 and a total surfa tant on entration of 1 wt.% upon the addition of 0.05 M hexanol: spheri al(bla k dots) and ribbon-like mi elles (top), as well as vesi les (bottom) are seen.

100

is possible to obtain vesi les stable for a long time in mixed surfa tant systems attemperatures mu h lower than the Kra�t temperatures of the two orrespondingsurfa tants.5.5 Con lusionsThe o urren e of supersaturation has been investigated in mixed surfa tant solu-tions omposed of SDS / CTAB and SDS / DTAB. A temporary `solubility temper-ature depression' has been observed in the anioni -ri h part of the phase diagram.We have shown that the stability of su h supersaturated solutions an be tuned byin reasing the ioni strength, by variation of ounterions, and by the addition ofal ohols. The presen e of simple ele trolytes generally de reased, while the additionof middle- hained al ohols in reased its stability.Destabilization and on urrent pre ipitation of the systems was shown not to o urdue to the formation of bigger aggregates, but rather due to a shift of the equilibriumbetween mi elles and monomers. The addition of low on entrations of middle- hainal ohols resulted in the formation of vesi les in the atanioni systems at tempera-tures signi� antly lower than the Kra�t temperatures of the ioni surfa tants.The lifetime of vesi les and mi elles ould therefore be ontrolled by varying the omposition of the surfa tant solutions and by additives. Controlling the pre ipita-tion phenomena is of importan e for a large number of industrial pro esses, whereformulations need to be tuned.

101

102

Part IIIToward Appli ation

103

Chapter 6Use of Surfa tants in Cosmeti Appli ation: Determining theCytotoxi ity of Catanioni Surfa tant Mixtures on HeLa Cells6.1 Abstra tThe ytotoxi ity of ommonly used syntheti surfa tants and atanioni mixturesof those was evaluated using MTT on HeLa ells. The 50% inhibition on entration(IC50) for MTT redu tion was al ulated. The e�e t on hain length in rease andin lusion of polyoxyethylene groups on the toxi ity was tested on single surfa tantsystems. A general trend of in reasing toxi ity with in reasing hain length andthe presen e of polyoxyethylene groups was observed. The measured IC50 valuesof atanioni systems lie between those of parti ipating surfa tants. The in reasein toxi ity as the ationi surfa tant was added to the anioni one was howevernot linear. A steep de rease of the IC50 values (and therefore in rease in the toxi properties) was observed immediately already at low on entrations of the ationi surfa tants. This behavior is analogous to the enzyme a tivity in atanoni mi- roemulsions. 105

6.2 Introdu tionVesi les are ommonly used in osmeti s and pharma y as vehi les for a tive agents.A tive mole ules an thus be en apsulated in the bilayer membrane, if they arelipophili , or in the ore of the vesi le, if they are hydrophili . En apsulation isuseful to prote t a tives in preventing any undesired rea tion. Vesi les an thus beused as ve tors to deliver drugs to a spe i� pla e, without being destroyed.Improving the bio ompatibility of produ ts used in osmeti formulation further issought after. For this reason it is important to identify the irritating properties of ommer ially used surfa tants. Lately, atanioni systems have been investigatedas potential delivery systems due to their ability to spontaneously form vesi ularphases. Furthermore, new surfa tants are being synthesized ontinuously. For thisreason, it is important to establish an easy way to estimate the potential toxi ity ofsu h substan es.Performing ytotoxi ity tests on established ell lines is a good alternative to ex-pensive and morally questionable animal tests. In the present study, ytotoxi ityof ommer ially used surfa tants and ommon atanioni surfa tant mixtures weremeasured on HeLa ells in order to identify their toxi ity for use in pharma euti- al and osmeti appli ations. HeLa ells are a well-known ell line that is usedto asses the ytotoxi ity of hemi al ompounds320�323 and reportedly show goodreprodu ibility and a signi� ant orrelation with in vivo results324. Cell viabilitywas evaluated by the tetrazolium MTT redu tion assay, based on the uptake andthe redu tion of the soluble yellow MTT tetrazolium salt by mito hondrial dehydro-genase to a blue insoluble MTT formazan produ t. The IC50 value ( on entrationof test substan e that lowers MTT redu tion by 50% ompared with the untreated ontrol) was al ulated from absorban e data. Previous tests have shown that themito hondrion-based MTT test is a very sensitive indi ator of surfa tant ytotoxi- ity325. Intra ellular organelles (su h as mito hondria) are a�e ted already by lowsurfa tant on entrations due to their high sensitivity to surfa tant membrane a -tions. Furthermore, high orrelation between MTT redu tion assay tests and y-tosoli la tate dehydrogenase leakage (LDH) in vitro and the Draize eye irritan y106

data in vivo were reported325. In fa t, ytotoxi ity in vitro proved to be more sen-sitive and quantitative than the Draize test326.A few reports on erning the toxi ity of surfa tants using di�erent assays should benoted325,327�332. The toxi ity of SDS is ommonly investigated as an example of anioni surfa tant. It has been established that ioni surfa tants show higher toxi ityvalues than the non-ioni ones, and that ationi surfa tants are more potent thantheir anioni ounterparts328. However, reports on the toxi properties of atanioni mixtures are very s ar e. The ytotoxi ity of a SDS/CTAB/ holesterol mixture wastested on murine ma rophage-like ells by Kuo et al.333. A dose-dependent apoptosiswas observed. The study did however not in lude the e�e t aused by the variationof the ratio of the surfa tants and their hain lengths.6.3 Experimental Pro edures6.3.1 MaterialsHeLa Cell Line HeLa ells were distributed by the Ameri an Type Culture Col-le tion (ATCC). Cells were ultured in Earle's minimum essential medium (MEM) ontaining 0.85 g/L NaHCO3 supplemented with FCS (fetal alf serum)(10%), L-Glutamin (2 mM), NEA (non-essential amino a ids)(1%), Amphoter in B (0.4µg/ml)and Peni ilin G / Streptomy in sulfate (100 u/mL).Surfa tant Solutions The surfa tants, sodium o tanoate (SO; Sigma, Germany;grade: 99%), sodium dode anoate (SL; Sigma, Germany; grade: 99-100%), sodiumdode yl sulfate (SDS; Mer k, Germany; assay > 99%), sodium laurelth sulfate(Texapon 70; Na+C12(EO)2.2OSO−

3 , a tive assay: 70%, gift from Cognis, Germany)and sodium dode ylether arboxylate (Akypo Soft 45NV; Na+C11(EO)4.5OCH2COO−,a tive assay: 21%, gift from Kao Chemi als, Germany), o tyltrimethylammoniumbromide (OTAB; Mer k, Germany; assay > 99%), de yltrimethylammonium bro-mide (DeTAB; Mer k, Germany; assay > 99%), dode yltrimethylammonium bro-mide (DTAB; Mer k, Germany; assay > 99%), etyltrimethylammonium bromide107

(CTAB; Sigma, Germany; grade: 99%) were used as re eived. Surfa tant sto k solu-tions were prepared by dissolving weighed amounts of dried substan es in Milliporewater. The solutions were then left for 24 hours to equilibrate at 25◦C. The atan-ioni solutions were prepared by mixing the surfa tant sto k solutions to obtaindi�erent anioni / ationi surfa tant mass ratios.6.3.2 Growing HeLa Cell CulturesSubstan es: MEM - EARLE Medium, PBS - DUBECCO Bu�er, Trypsin / EDTA,Flask ontaining HeLa ells1. The old medium was removed from the �ask ontaining HeLa ells that we wishedto transplant.2. The ells were rinsed with approximately 4 ml of PBS bu�er solution. With thiswe have removed all the �free-swimming� ells; only the ells stu k to the bottom ofthe �ask remained.3. 3 ml of Trypsin / EDTA was added and the ells were in ubated at 37◦C for5 minutes. Trypsin disables the ell's ability to bind to surfa es ( onsequently, the ells are swimming freely in the suspension).4. The ell suspension was transferred into a new �ask, 3 ml of new MEM mediumwas added to stop the rea tion of Trypsin and the ells were entrifuged at 800/minfor 5 minutes (20◦C).5. On e entrifuged, the supernatant was arefully removed.6. 4 ml of new MEM medium was added and the solution was homogenized using apipette until ell- lusters disappeared.7. A desired amount of ell suspension (depending on when we needed the ellsready for the next experiments) was transferred into a new �ask and MEM mediumwas added to a olle tive volume of 20 ml.8. The ells were left to in ubate at 37◦C for minimum 3 days.108

6.3.3 HeLa Toxi ity Test1. The viability of the ells was he ked with Trypanblue (solution: 0.04 g Tryptan-blue in 10 ml PBS bu�er); 10 µl ell suspension +10 µl Trypanblue => we ountedthe living ells under the mi ros ope (the defe ted ells were olored blue, the inta tone's white).2. Number of ells in the 16 squares: f.i. 330

330 · 2 = 660 dilution with Trypanblue660 · 10 = 6.600 ells in 1µl6.600 · 1.000 = 6.600.000 ells in 1 ml3. We need 150.000 ells per experiment:6.600.000 : 150.000 = 44That means we need to dilute 1 ml of the ell suspension with 44 ml of medium.For 1 plate (60 wells with 75 µl) 4.5 ml of ell suspension is needed (2 plates = 9ml); we prepare 0.5 ml ell suspension + 22 ml of medium.4. Con entration of substan es of whi h we wish to measure toxi ity: 1 mg/ml (1wt.%); 0.1 mg/ml if the substan es are very potent ( ationi surfa tants).5. Filling the plates:o The wells at the borders were �lled with 150µl of pure medium (36 · 150 µl = 5.4ml)o The middle 60 wells were �lled with 75 µl mediumo 60 µl medium was added to wells 2 / B-G (the olumns are des ribed using num-bers, and rows using letters)o Wells 2B/2C were additionally �lled with 15 µl of sample 1; 2D/2E with 15 µl ofsample 2; 2F/2G with 15 µl of sample 3.6. With a 6- anal pipette the solutions were thoroughly mixed in wells 2B - 2Gand then 75 µl of the solutions was transfered from olumn 2 to olumn 3 (wells3B - 3G). This pro edure was repeated until row 10 (10B - 10G). The last 75 µl ofsolutions from row 10 was dis arded. Row 11 was reserved for our �blind� probe.7. Finally, 75 µl of ell suspension was added to all 60 middle wells with (withoutmixing)(in the end the whole plate should ontain 150 µl / well).8. The ells were left to in ubate for 68 hours (±1 hour).109

Figure 6.1: 96-well plates used for toxi ity studies immediately after preparation (top) and afterin ubation with MTT (bottom). 110

6.3.4 Dete tion1. 15 µl of MTT solution (4 mg/ml of MTT in PBS bu�er) was added to all wells ontaining ells (0.9 ml MTT solution per plate).2. The plates were left to in ubate for 4 hours.3. The ex ess MTT ontaining medium was arefully remove. Well 1A was emptiedas well.4. The empty wells were �lled with 150 µl 10 wt.% SDS solution (in water; 18.3 mlfor two plates).5. The mi ro-plates were pla ed in laminar �ow overnight, then the opti al densityof ea h plate was measured with a mi roplate reader at 560 nm.A sample plate immediately after the preparation pro ess and after the MTT in u-bation is presented in Figure 6.1. The wells at the rims are �lled only with medium(pink olor), the blue olor is a sign of inta t ells and the transparent solution of ell apoptosis.6.3.5 Evaluation of spe tros opi al dataThe IC50 value (in µg /ml), whi h represents the on entration of test substan e thatlowers MTT redu tion by 50% ompared with the untreated ontrol, was al ulatedfor ea h substan e from the on entration-response urve. A sample al ulationsheet is presented in Figure 6.2. Experiments were repeated four times (n = 5),and the average IC50 is reported. Maximal observed (absolute) standard deviationwas about 15%. Positive ontrol measurements were performed with xanthohumol(HeLa ells: IC50 ≈ 7 µg/mL).6.4 Results and Dis ussion6.4.1 Single-Chain Surfa tantsCell viability was determined by the tetrazolium MTT dye redu tion in the ell ulture system. Evaluation of the results revealed a dose-dependent ytotoxi ityafter 68 hours of exposure. The IC50 values are reported in Table 6.1.111

HeLa -Cells

Testsubstance:SDS/DTAB 90/10

Konz[µg/ml] OD Average corr. ave. Restakt.[%]

StAbw[%]

Empty value 0.04280.043

Control 0.8590 0.8325 0.87390.8141 0.9065 0.8372 0.854 0.811

1000 0.0519 0.0532 0.053 0.0098 1.2 0.1500 0.0511 0.0472 0.049 0.006 0.8 0.3250 0.0457 0.0433 0.045 0.002 0.2 0.2122.5 0.0661 0.0821 0.074 0.031 3.9 1.462.5 0.3885 0.3824 0.385 0.343 42.2 0.531.25 0.4840 0.5725 0.528 0.485 59.9 7.715.63 0.7098 0.7054 0.708 0.665 82.0 0.47.8 0.8245 0.8394 0.832 0.789 97.3 1.33.9 0.9018 0.9091 0.905 0.863 106.4 0.6

Conc. >50%: 31.25Value: 59.9µg/mlConc.<50%: 62.5 Value: 42.2µg/ml

IC 50 =48.73µg/ml

Figure 6.2: Cal ulation of the IC50 value from the spe tros opi al data.Comparison of the IC50 values of the surfa tants revealed the following ytotoxi properties:1) Anioni surfa tants generally exhibited a lower ytotoxi ity than atanioni sur-fa tants. 112

Anioni surfa tants Cationi surfa tantsname IC50 [µg/mL℄ name IC50 [µg/mL℄SO 530.5 ± 41.7 OTAB 23.3 ± 1.8SL 115.8 ± 8.2 DeTAB 4.8 ± 0.6SDS 177.1 ± 7.8 DTAB 5.2 ± 0.3Texapon 70 69.2 ± 1.7 CTAB 1.6 ± 0.2Akypo Soft 45NV 76.3 ± 4.8Table 6.1: The IC50 values ( on entration of test substan e that lowers MTT redu tion by 50% ompared with the untreated ontrol) of singe- hain ioni surfa tants.2) Surfa tants ontaining sulphate headgroups showed higher IC50 values than thosebearing a arboxylate (having the same hain-length).3) An in rease of the toxi ity of the surfa tant was observed as the length of thehydro arbon hain was in reased ( . f. Figure 6.3). An exponential de rease in theIC50 values was observed, whi h leveled o� at a ertain hain length.4) The presen e of the oxyethylene groups in reases the toxi properties of the sur-fa tants. Two surfa tants ontaining 12 arbon atoms, and polyoxyethylene groups,one with a sulphate (Texapon 70) and one with a arboxylate headgroup (AkypoSoft 45 NV) were ompared to those without the EO groups. An in rease in thetoxi properties is observed for both surfa tants, however Akypo Soft 45 NV exhibitsa higher IC50 value, than its sulfate ontaining ounterpart, despite the fa t thatSDS shows higher IC50 values than SL.6.4.2 Catanioni Surfa tant SystemsThe toxi ity of atanioni surfa tant mixtures omposed of SDS / DTAB and SL/ DTAB were analyzed for di�erent surfa tant ratios. The results are presented inTable 6.2 and Figure 6.4.We an see that the presen e of only a small amount of a ationi surfa tantin reased the toxi ity of the mixture signi� antly. This was somewhat surprising,as a linear in rease of the toxi ity was expe ted.113

8 10 12 14 160

10

20

30

100

200

300

400

500

600

IC50

[g/

mL]

n (C)Figure 6.3: E�e t of hydro arbon hain length on the ytotoxi ity of ioni surfa tants: n(C)represents the number of arbons in ationi CnTAB (�) and anioni SCn (�) surfa tants.SDS / DTAB SL / DTABxSDS IC50 [µg/mL℄ xSL IC50 [µg/mL℄0 4.2 ± 0.2 0 4.2 ± 0.2

0.1 3.5 ± 0.1 0.1 5.1 ± 0.4

0.2 3.8 ± 0.3 0.2 4.6 ± 0.6

0.3 3.7 ± 0.2 0.3 6.3 ± 0.5

0.4 6.7 ± 0.5 0.4 7.1 ± 0.2

0.5 8.2 ± 0.8 0.5 7.2 ± 0.4

0.6 7.9 ± 0.6 0.6 13.0 ± 1.2

0.7 11.6 ± 1.0 0.7 16.4 ± 0.9

0.8 20.1 ± 1.6 0.8 28.6 ± 2.1

0.9 43.1 ± 3.2 0.9 49.7 ± 3.4

1.0 177.1 ± 7.8 1.0 115.8 ± 8.2Table 6.2: The IC50 values of atanioni surfa tant mixtures.114

0.0 0.2 0.4 0.6 0.8 1.0

0

25

50

75

100

125

150

175

200

IC50

[g/

mL]

XSDS/SLFigure 6.4: The mito hondrial redu tion of MTT after a 68-hour in ubation of SDS / DTAB(�) and SL / DTAB (�) mixture at di�erent surfa tant ratios with HeLa ells.

This behavior is similar to the one found when the enzymati a tivity in atan-ioni emulsions was studied334. It was observed that the presen e of DTAB exertsan inhibiting e�e t on the enzyme. At weight fra tions higher than 0.22, the enzymewas ompletely inhibited. This is approximately the weight fra tion at whi h thetoxi ity of the mixtures equals that of a pure DTAB solution. It is highly proba-ble that at weight fra tion lower than 0.22 DTAB is ompletely in orporated in the atanioni mi elles and its toxi ity is redu ed by the strong ele trostati intera tionswith the oppositely harged anioni surfa tant.Results suggest that in order to keep the toxi ity of potential drug delivery systemslow, a high fra tion of the anioni omponent is ne essary. So far, vesi les have beenshown to exist in the anioni -ri h region of the atanioni phase diagrams, howeverat ratios lose to equimolarity. The vesi ular region was shifted to higher anioni fra tions when salt was added to the system169,335. A system of lower toxi ity om-posed of a mixture of anioni surfa tants will be reported in the following hapter115

(Chapter 7).6.5 Con lusionsThe ytotoxi ity of single- hain ioni surfa tants and atanioni mixtures was eval-uated using MTT on HeLa ells in order to examine their potential use in osmeti and pharma euti al formulation. It was on�rmed that anioni surfa tants generallyexhibit higher IC50 values than atanioni ones. The toxi ity an further be in�u-en ed by the hydro arbon hain length and the presen e of polyoxyethylene groups.A general trend of in reasing toxi ity with in reasing hain length and the pres-en e of polyoxyethylene groups was observed. A non-linear in rease was observedas ationi surfa tants were added to anioni ones. A steep de rease of the IC50values was observed already at low fra tions of the ationi surfa tants, suggestingthat in potential drug delivery systems a high fra tion of the anioni omponent isne essary.

116

Chapter 7Spontaneous Formation of Bilayersand Vesi les in Mixtures ofSingle-Chain Alkyl Carboxylates:E�e t of pH and Aging7.1 Abstra tWe report the observation of bilayer fragments, some of whi h lose to form vesi les,over a large range of pH at room temperature from mixtures of single- hain bio om-patible ommer ially available non-toxi alkyl arboxyli surfa tants after neutral-ization with HCl. The pH at whi h the morphologi al transitions o ur, was variedonly by hanging the ratio between two surfa tants: the alkyl-oligoethyleneoxide- arboxylate and sodium laureate. The e�e t of aging of the mixed surfa tant sys-tems in the pH region desired for dermatologi appli ation (4.5 < pH < 7) was alsostudied.7.2 Introdu tionThere is a growing interest in the �eld of syntheti surfa tants used for dermatologi purposes. These surfa tants exhibit a wide range of stru tures; parti ularly useful is117

the formation of vesi les, whi h an be used as delivery systems. For osmeti rea-sons, the surfa tants forming vesi les must be skin ompatible (non-irritant), easy tomanufa ture, and the vesi le region must be stable within the range of physiologi alpH at room temperature.A simple geometri al hara terization of hain pa king an be used to analyze trendsin surfa tant phase behavior15,60. The geometri properties of surfa tants dependon the ratio between the ross-se tional area of the hydro arbon part and that ofthe headgroup. Low pa king parameters (around 1/3) are found for single hainsurfa tants with a strongly polar headgroup. These systems tend to form spheri- al mi elles, whereas a pa king parameter value around one favors the formationof lamellar stru tures. An in rease in the pa king parameter an be obtained byadding a se ond hain, whi h doubles the hydro arbon volume. Double- hain sur-fa tants18,21, two surfa tants of opposite harge23�26, or the asso iation of a surfa -tant and a o-surfa tant27�33, an be used. In the latter two ases a pseudo-double hain surfa tant is obtained by either an ion-pair formation between the anioni and ationi surfa tant, or due to asso iation of the two di�erent mole ules via hydrogenbonds.As ationi surfa tants and short al ohols are undesirable in osmeti formulationsdue to their toxi ity, and be ause long- hain al ohols (> C12) exhibit high meltingpoints, monoalkyl arboxylates were hosen for our study. Fatty a ids form a rangeof aggregates depending on the a id on entration and the ionization degree of theterminal arboxyli group336,337. The formation of vesi les from mono arboxyli a ids has long been known. Gebi ki and Hi ks338 �rst observed the formation ofvesi les from unsaturated, long- hain fatty a ids. Later, Hargreaves and Deamershowed that also saturated fatty a ids an form vesi les61. A vesi le phase is spon-taneously formed, when short or middle hain (< C12) fatty a ids are neutralizedwith HCl61,81; two types of amphiphiles are then present in the solution, the proto-nated and the ionized form. The ratio between the two determines the aggregationmorphology.Despite the simpli ity of the me hanism of vesi le formation, possible appli ationsof fatty a id vesi les in osmeti s remain largely unexplored339,340. This may be a118

onsequen e of two obsta les: (i) the high solubility temperature of the long- hain arboxylates and, (ii) the generally too basi pH ne essary for the solubilization ofthe arboxylate. The problem of the solubility temperature of the alkyl arboxylateshas been dis ussed by Hargreaves et al.61 Below 25◦C only alkyl arboxylates withshort alkyl hains are water-soluble. However, these are inappropriate due to theirskin irritating properties. The solubility temperature of sodium laureate is reportedto be equal to or above room temperature, depending on the on entration336. Withlonger alkyl hains (n > 12) the solubility temperature be omes even higher.It has previously been reported that the formation of vesi les o urs at pH valuesat or near the pKa, where approximately half of the arboxyli groups are ion-ized61,336,339. For this reason, fatty a id vesi les are present only over a narrow pHrange. Designing vesi les that be ome unstable at an easily tuned pH value is ofgreat interest for targeted drug delivery. It is known that, for example, tumors andin�amed tissues exhibit a de reased extra ellular pH63�67. For this reason a largenumber of groups have fo used their attention on the preparation of pH-sensitiveliposomes68�78 as possible drug arrier systems.A range of sugar-based gemini surfa tants has been re ently studied, be ause theyexhibit pH-dependent aggregation behavior. However, these surfa tants are ationi and therefore vesi les are observed only at neutral and high pH values341,342. Thesame problem o urs also in solutions of bola-amphiphiles343. Systems omposed ofmixed single- and double-short-tailed PEO ether phosphate esters might be promis-ing for the formation of pH-sensitive vesi les. In that ase a vesi le phase wasobserved at higher on entrations. The pH e�e t, however, was only studied onvesi ular solutions, so one an not on�rm a mi elle-to-vesi le transition as a onse-quen e of protonation344.For arboxylates with longer alkyl hains, desirable for osmeti formulations (i.e.lauri a id; pKa ≈ 8 − 8.524,61, depending on the on entration), the pH of vesi leformation is also generally too basi . The addition of medium- and long- hain al o-hols has been proven to expand the pH region of vesi le formation61,339, but towardeven higher pH values, whi h is undesirable. One way to lower the pH range ofvesi le formation is to introdu e another arboxyli group to the fatty a id hain,119

thus lowering the pKa of the a id. This has been previously done by de Groot et al.by the formation of 2-(4-butylo tyl) maloni a id345. However, to the best of ourknowledge, no one has tried to use a mixture of single- hain fatty a id soaps yet.Re ently, spontaneous formation of vesi les below room temperature, at a idi pH(between 2 and 4), by neutralization of a parti ular industrial single- hain alkyl ar-boxylate surfa tant was reported346. Our intent was to obtain spontaneously-formedvesi les over a wide range of physiologi al pH at room temperature, by hanging theratio of two surfa tants: Akypo Soft 45NV (an alkyl-oligoethyleneoxide- arboxylatealready used in osmeti formulations; AS) and sodium laureate (SL). The e�e tof pH and aging on vesi le formation was studied by visual observation, dynami light s attering and transmission ele tron mi ros opy. The bio ompatibility of thesurfa tant mixture was he ked by measuring the ell viability. Based on our results,we propose a general method to obtain vesi les on a large range of physiologi al pHusing only non-toxi alkyl arboxylates.7.3 Experimental Pro eduresChemi als Sodium laureate (SL) was pur hased from Sigma-Aldri h at a purity of99%. Sodium dode ylether arboxylate (Akypo Soft 45NV; Na+C12(EO)4.5OCH2COO−)was a gift from Kao Chemi als (Germany). A ording to the information given byKao Chemi als, AKYPO Soft 45NV ontains mole ules with roughly a Gaussiandistribution of the number of EO groups with an average value of 4.5. AKYPO Soft45NV is supplied as an approximately 21 wt.% aqueous solution. Sto k solutions of1wt.% were prepared by dissolving weighed amounts of sodium laurate in deionizedwater and by dilution of the original Akypo Soft 45NV solution. Di�erent solutionswere then prepared by mixing di�erent ratios of the two sto k solutions.Phase Diagrams The phase behavior as a fun tion of temperature was deter-mined by visual observations. The samples were ooled down and left to equilibrateat 0◦C, then the temperature was raised by approximately 0.5◦C per minute. Themeasured solubility temperatures orrespond to the transition from pre ipitate to120

isotropi phase, i.e., to a mi ellar or vesi ular solution. The samples were titratedwith 0.1 M HCl at 25◦C. A Con ort, type C831 pH-meter, with a Bioblo k S ienti� glass ele trode (Consort ref. nr. SP02N), was used.Dynami Light S attering Parti le size analysis was performed by a Zetasizer3000 PCS (Malvern Instruments Ltd., England), equipped with a 5 mW helium-neonlaser with a wavelength output of 633 nm. The s attering angle was 90◦ and theintensity auto orrelation fun tions were analyzed using the CONTIN software. Allmeasurements were performed at 25◦C.Cytotoxi ity Tests HeLa ells were provided by the ATCC (Ameri an TypeCulture Colle tion). The passage numbers of the HeLa ells used in the proje tvaried from 25 to 30. Cells were ultured in Earle's minimum essential medium(MEM) ontaining 0.85 g/L NaHCO3 supplemented with FCS (10%), L-Glutamine(2 mM), NEA (1%), Amphoter in B (0.4 µg/ml) and Peni illin G / Streptomy insulfate (100 u/mL). Keratyno ytes (SK-Mel-28) were provided by the ATCC. Thepassage numbers varied from 15 to 20. Cells were ultured in Dulbe o's MEM ontaining 3.7 g/L NaHCO3 supplemented with FCS (10%), L-Glutamine (2 mM),NEA (1%), Amphoter in B (0.4 µg/ml) and Peni illin G / Streptomy in sulfate (100u/mL). MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) assaypro edure was prepared a ording to Mosmann347 in 96-well mi ro-plates using thedoubling dilution method (as des ribed in the previous hapter). First, the wells ofthe �rst olumn were �lled with 135µl of MEM medium. All subsequent olumnswere �lled with 75 µl of medium. Then, 15µl of 1wt.% test solutions was addedto ea h well in the �rst olumn and the solutions were thoroughly mixed with a6- hannel pipette. 75µl of the test substan e / medium mixture was transferredfrom the wells in the �rst olumn to those in the se ond. Again the solutions weremixed and the pro edure was repeated for the whole plate. Then, 75µl of ellswere seeded (2.5 · 103 ells / well). After a 68-hour in ubation at 37◦C, 15µl ofMTT (5 mg /mL) were added to ea h well. After a 4-hour in ubation, the medium ontaining MTT was removed and 150µl of 10% SDS solution was added. The121

mi ro-plates were pla ed in laminar �ow overnight, then the opti al density of ea hplate was measured with a mi roplate reader at 560 nm. The IC50 value (µg /ml),whi h represents the on entration of test substan e that lowers MTT redu tion by50% ompared with the untreated ontrol, was al ulated for ea h substan e fromthe on entration-response urve. Experiments were repeated four times (n = 5),and the average IC50 is reported. Maximal observed (absolute) standard deviationwas about 15%. Positive ontrol measurements were performed with xanthohumol(HeLa ells: IC50 ≈ 7 µg/mL).Cryo-TEM We prepared vitri�ed ryo-TEM spe imens in a ontrolled environ-ment vitri� ation system (CEVS), at a ontrolled temperature of 25◦C and �xed100% relative humidity, followed by quen hing into liquid ethane at its freezingpoint348. We examined the spe imens, kept below −178◦C, by an FEI T12 G2transmission ele tron mi ros ope, operated at 120 kV, using a Gatan 626 ryo-holder system. Images were re orded digitally in the minimal ele tron-dose modeby a Gatan US1000 high-resolution ooled-CCD amera with the DigitalMi rographsoftware pa kage.7.4 Results and Dis ussion7.4.1 Lowering of the Solubility Temperature of Fatty A idsTo form vesi les at room temperature, the solubility temperature TSol of the sur-fa tant mixture has to be equal to or below this temperature. The use of sodiumlaureate is therefore limited, be ause it is insoluble in water at temperatures below24◦C (for solutions of 1 wt.%). We were able to lower the solubility temperature bymixing sodium laurate with a surfa tant that is soluble at lower temperatures. ASwas hosen be ause its TSol = 5◦C. Figure 7.1 shows that the solubility temperaturede reased linearly with in reasing amounts of AS for a �xed total surfa tant on en-tration of 1wt.%, and ould be des ribed by the equation: TSol (◦C) = −19.9·A+23.9,where TSol represents the solubility temperature of the mixture, and A the amountof AS, expressed in wt.%. 122

0.0 0.2 0.4 0.6 0.8 1.00

5

10

15

20

25

T solu

bilit

y [o C]

amount of Akypo Soft 45NV (wt.%)Figure 7.1: Solubility temperature of mixed surfa tant solutions (AS and sodium laureate; ctot

S=

1 wt.%) as a fun tion of the amount of AS in the solution. A linear �t des ribes the dependen ewith the equation: TSol (◦C) = (−19.9±0.1) ·A+(23.9±0.1), where TSol represents the solubilitytemperature of the mixture and A the amount of Akypo expressed in wt.% (the total surfa tant on entration was kept at 1 wt.%)7.4.2 The E�e t of pH on Vesi le FormationThe titration of alkaline soaps with HCl has been des ribed by Rosano et al.225,when investigating the e�e ts of surfa e harge on lipid-water interfa es. In theseexperiments, the appearan e of a plateau (bu�ering apa ity) at a pre ise pH duringtitration was oin ided with the formation of lipid liquid- rystals.We observed thisplateau in solutions of dominating quantities of sodium laureate. As more SL wasrepla ed by AS, the plateau region be ame smaller, disappearing below 0.5 wt.%SL (mass ratio 1 : 1). By redu ing the pH of the samples we observed a su es-sion of two (or three) phases, depending on the mass ratio of the two surfa tants(Figure 7.2). At high pH, all solutions were isotropi and olorless, orrespond-ing to the mi ellar region of the phase diagram. By de reasing the pH, a bluish olor appeared, attributed to a formation of vesi ular stru tures. In the presen e123

of pure AS, only two phases appeared onse utively, while in SL / AS mixtures, ase ond phase transition to pre ipitation was observed at low pH values. In 1 wt.%SL solutions pre ipitation began at pH values below 7, whereas the addition of ASlowered the pre ipitation boundary to pH = 3.5 (in solutions with SL / AS ratio25/75). Pre ipitation when the surfa tant ratio nears equimolarity, had been ob-served in previous studies and was attributed to the in reased on entration of the onjugated a id169. The titrations were performed one week after the preparationof the samples. Ma ros opi ally, the pH values at whi h the phase transitions tookpla e remained the same with respe t to time. This is not true mi ros opi ally, asdes ribed below.The formation of fatty a id vesi les is restri ted to a narrow pH region lose to thepKa of the a idi omponents61,336,339. The apparent shift in the pKa of fatty a idanions (from the pKa = 4.67 of arboxyli group349 to the pKa = 8 of lauri a id61)has been attributed to the lo al de rease in pH at highly harged surfa es61. Sin epure fatty a id/soap vesi les (without additional amphiphiles) ontain an amphiphilethat is not harged (the neutral form of fatty a id), the two bilayer-forming om-ponents are asso iated by hydrogen bonds instead of ele trostati intera tions350.The formation of vesi les near the pKa of the a id an then be explained by theformation of stable hydrogen bond networks between the ionized and neutral a idforms.It has been observed that the addition of al ohols to fatty a ids auses an in reasein the pH of the region of vesi le formation339. In that ase, vesi les are formed dueto stable hydrogen bonds between the al ohol headgroup and the ionized a id head-group. The addition of the al ohol means larger presen e of the non-ionized spe iesin the solution (due to the high pKa of the al ohols); a larger amount of arboxyli a id will be in the ionized form to a hieve the same protonated/ionized headgroupratio needed for vesi le formation (1 : 1). An opposite e�e t an then be expe tedin the ase when the added omponent has a mu h lower pKa than the fatty a id,as is the ase of AS. This will be ompletely ionized at a neutral pH, therefore moreprotonated a id groups will be required for the formation of vesi les, resulting in ade rease of the pH of vesi le formation. This was on�rmed by our titrations, where124

the pKa (and orresponding bluish region) de reased with in reasing amount of ASin the mixture.

0.00 0.25 0.50 0.75 1.00 1.251

2

3

4

5

6

7

8

9

pH

VHCl [mL]Figure 7.2: Di�erent phase transitions observed by titration of mixed surfa tant solutions(sodium dode anoate and AS; ctot

S= 1 wt.%; Vo = 15 mL) ontaining (from bottom to top)

1; 0.75; 0.5; 0.25; and 0 wt.% of AS. The dashed lines represent a bluish solution (possibly vesi les);the solid lines represent the lear (mi ellar) solution (to the left) and the turbid phase (to the rightof the vesi ular phase).The hydrodynami radius of parti les in solutions at di�erent surfa tant ratioswas measured by dynami light s attering. The results on�rmed the presen e oflarge aggregates (possibly vesi les) over a large pH region (Figure 7.3) in mixedsurfa tant solutions. Be ause solutions of industrial surfa tants are highly polydis-perse (the polydispersity of AS has been previously reported169), only the averageradius ould be measured. The regions where RH was found to be 80 − 120 nm(the average radius of vesi les) is plotted in Figure 7.3 for di�erent surfa tant ratios.In reasing the amount of AS lead to a de rease of the pH, at whi h vesi les wereformed. The width of the pH range depended strongly on the surfa tant ratio. Thisrange is the largest for the ratio AS / SL : 0.75 / 0.25, where it extends from pH 3.5125

to 7.5. At this ratio the omplete range of skin pH or physiologi al pH is overed.A redu tion of this range was observed when the amount of Akypo Soft 45NV waseither in reased or de reased.

0.0 0.2 0.4 0.6 0.8 1.00

2

4

6

8

10

pH

AS [wt.%]

Micelles

Precipitation

Figure 7.3: pH range where (after 3 weeks) the average hydrodynami radius RH was approx-imately 100 nm (possible region of vesi le formation) in mixed surfa tant solutions (SL and AS)as a fun tion of AS on entration (ctot

S= 1 wt.%); the open squares (�) represent the solutionsexamined by ryo-TEM.The (average) hydrodynami radius (RH) of the obje ts present in the vesi le ontaining solutions is shown in Figure 7.4 for di�erent surfa tant ratios as a fun tionof the pH. The obtained values suggest the presen e of vesi les, but large variationsexist. In presen e of 0.75 wt.% of AS only small vesi les are measured with a smallpolydispersity index (around 0.2). At smaller ratios of AS, or in presen e of puresodium laurate, bigger obje ts of higher polydispersity are observed. Be ause lights attering is insu� ient to hara terize systems of high polydispersity, ryo-TEMwas used to analyze the system further.126

2 3 4 5 6 7 8 925

50

75

100

125

150

175

200

RH [n

m]

pHFigure 7.4: Average hydrodynami radius, RH , of the aggregates present in bluish solutions asa fun tion of pH for di�erent surfa tant ratios: (△) 0; (◦) 0.25; (•) 0.5; (N) 0.75; (�) 0.85; (�) 1(all expressed in wt.% of AS; ctot

S= 1 wt.%).7.4.3 Cryo-TEM Study of Time-Dependent Vesi le Forma-tionThe AS / SL ratio (75/25), at whi h the largest pH interval of vesi le formation wasmeasured, was further investigated by ryo-transmission ele tron mi ros opy. Theevolution of self-assembly aggregates was followed over a period of two months atthree di�erent pH values. The sequen e of observed morphologies is summarized inTable 7.1.We see from Table 7.1 that in the �rst weeks after preparation mi elles anddis s are the observed stru tures; these are presented in Figure 7.5. With time,membranes started forming and eventually vesi les appeared (Figure 7.6). Thehigher the pH of the solution, the faster this transition took pla e; in samples withpH = 6.7 vesi les were observed already after 2 weeks, whereas in solution withpH = 5.5, these were visible just after 4 weeks. After 2 months vesi les were still127

pH No. of days after preparation5 10 17 28 604.5 D, MP M D MP5.5 M, D M, D M, V M, MP6.7 MP, V M, D, V (C) M, MP, V (I)Table 7.1: Development of various surfa tant self-assembled aggregates with time; M - mi elles,D - dis s, MP - membrane pie es, V - vesi les, V (C/I) - vesi les with either usps, or in ompletevesi les.not observed in solution with the lowest pH ( . f. Figure 7.7 a). However, thetwo samples at higher pH values already showed in omplete vesi les and membranepie es ( . f. Figure 7.7 b). It seems that after a ertain time, the vesi les inthese systems begin to destabilize. We noted that despite the ri h morphologi aldevelopment that ould be observed by mi ros opy, ma ros opi ally the solutionsappeared about the same during the period of observation. Furthermore, we ansee that not all regions where big obje ts were dete ted by DLS ontain vesi les.These were only found at higher pH values (above 5.5), whereas at low pH, dis sand membrane pie es were the dominating form.The results suggest it is possible to obtain vesi les in the region de�ned by thetwo pKa; at ertain surfa tant ratios an even broader pH range was observed. Thisis in a ordan e with results obtained by de Groot et al.345 for a bran hed monoalkylsurfa tant with a malonate headgroup. A bluish zone was observed between the twopKa values (the pKas of the two arboxylate groups on maloni a id are 2.85 and5.70, respe tively)345. In that region o-existen e of small unilamellar (SUV) andmultilamellar (MLV) vesi les was found. Above pH = 5.8 the solutions were learand only SUVs were observed by TEM345.Su h a high polydispersity of morphologies, where �at dis s oexist in equilibriumwith spheri al open and omplete unilamellar vesi les, was previously observed in atanioni systems. A possible explanation for the oexisten e of these stru tureswas proposed by Jung et al. by evaluating the parameters ontributing to thespontaneous urvature351. In some systems however (namely mixtures of anioni 128

Figure 7.5: Cryo-TEM images of a SL / AS mixed surfa tant solution at (left) pH = 4.5, 10days after mixing, and (right) pH = 5.5, 5 days after mixing. The bla k arrows point to dis s,the double arrows to a partially folded dis , whereas the white arrowheads to globular stru tures(possibly mi elles). The white arrow points to an elongated stru ture.

Figure 7.6: The formation of small vesi les after 2 weeks in samples with pH = 6.7 and largemultilamellar vesi les after 4 weeks (pH = 5.5); arrowheads point to very small perfe t vesi les,while the arrow points to an in omplete two-lamellar vesi le (`F' indi ates a frost parti le).129

Figure 7.7: Phase behavior after 2 months: (a) small membrane fragments at low pH (= 4.5);(b) destabilization of vesi les at high pH (= 6.7).and zwitterioni surfa tants), dis oidal stru tures are only short-lived intermediatesin the mi elle-to-vesi le transitions352. The vesi ulation depends on the balan ebetween the unfavorable edge energy of the disks and the bending energy requiredto form spheri al stru tures. In the ase of atanioni surfa tant mixtures, the edgesof the dis s are stabilized by the ex ess ationi surfa tant. It is possible that therelatively bulky EO groups stabilize the dis s observed in our system.7.4.4 Cytotoxi ity Potential of Surfa tant MixturesAkypo Soft 45 NV is ommonly used in osmeti formulations, su h as hair on-ditioners and hair dyes353�357. Improving the bio ompatibility of produ ts used in osmeti formulation even further is always sought after. Cytotoxi ity of our mix-tures on HeLa ells and Keratyno ytes was measured in order to identify the toxi properties of the two surfa tants for use in pharma euti al and osmeti al appli a-tions. Cell viability was evaluated by the tetrazolium MTT redu tion assay, basedon the uptake and the redu tion of the soluble yellow MTT tetrazolium salt bymito hondrial dehydrogenase to a blue insoluble MTT formazan produ t. The IC50value was al ulated from absorban e data. HeLa ells were used, be ause theyreportedly show good reprodu ibility and a signi� ant orrelation with in vivo re-130

sults324, whereas Keratyno ytes were hosen to he k the skin ompatibility of su hsurfa tant mixtures.

0.0 0.2 0.4 0.6 0.8 1.040

60

80

100

120

IC

50 [

g/m

L]

Akypo / SL

Figure 7.8: The mito hondrial redu tion of MTT after a 3-day in ubation of SL / AS mixtureat di�erent surfa tant ratios with HeLa ells (�) and Keratyno ytes (�).IC50 values reported in Figure 7.8 show a dose-dependent de rease in toxi itywhen AS is ex hanged for SL. Therefore any surfa tant mixture formed with SL isbene� ial. The order of the IC50 values was the same for both ell lines; however theabsolute values were slightly lower in the ase of the Keratyno ytes. The loweringof the pH of the surfa tant test solutions showed no pronoun ed e�e t. The volumeof the added surfa tant solution was probably too small to hange the pH value ofthe ell medium signi� antly and the bu�ering apa ity of the medium too large forany e�e ts to be observed.7.5 Con lusionsTwo obsta les mentioned in the introdu tion, namely, the high solubility temper-ature of alkyl arboxylates and limited pH region of vesi le formation have been131

over ome by using a mixture of two alkyl arboxylates with two very di�erent pKavalues. The presented results show that it is then possible to spontaneously formbilayer stru tures, su h as vesi les, in a pH range between the two onsidered pKa.Sodium laureate (with a pKa around 8.5) and Akypo Soft 45 NV (with a pKa = 4.67) ould be thus used at di�erent mixing ratios to obtain vesi ular solutions over theentire range of pH omparable to skin or physiologi al pH. Furthermore, we haveused ommer ially available and bio ompatible surfa tants that are already used for osmeti purposes and are inexpensive, whi h is of signi� ant importan e for appli- ation purposes358. However, more resear h is required to improve their appli ationpotential by in reasing the olloidal stability of fatty a id vesi les and assure thevesi les are ompletely losed. Both are ommon problems in su h systems81,359.

132

Con lusionIn this thesis the formation and stabilization of atanioni vesi les was studied.Catanioni surfa tant mixtures were hosen, be ause they were shown to form vesi- le spontaneously under ertain onditions and exhibit an enhan ed sensitivity tooutside parameters, su h as temperature or the presen e of salts. Due to this sen-sitivity, atanioni mixtures are a good system to study spe i� -ion e�e ts. Theinterest in vesi les was further motivated by their potential as models for biologi almembranes and as en apsulation systems. To optimize the appli ations it is impor-tant to have a general understanding of the interplay of intera tions between thesurfa tants and of the fa tors in�uen ing the phase diagram of a mixed system.First, the morphologi al transitions o urring in mixed surfa tant systems upon thein rease of the ioni strength were explored. The transition of mixed ioni mi ellesto vesi les in a atanioni surfa tant solution, omprised of sodium dode ylsulfateand dode yltrimethylammonium bromide, with an ex ess of the anioni omponent,upon the addition of salt is des ribed in Chapter 2. A new type of intermediatestru ture was found: a symmetri ally shaped spheri al super-stru ture, whi h wenamed blastulae vesi le. A me hanism of formation for this type of super-stru tureswas proposed, suggesting the importan e of harge �u tuation in the vesi le mem-brane.The spe i� -ion e�e ts on this system was further studied and ompared to onewhere the sulphate surfa tant was repla ed by a arboxylate surfa tant (sodium do-de anoate). No anion spe i� ity was found in the anioni -ri h region of the phasediagram for the added salts, whereas the nature of the ation was found to stronglyin�uen e the riti al salt on entrations around whi h mi elles turn to vesi les. Thiswas explained by taking into a ount that in the present ase of negatively harged133

vesi les, the ations a umulate in high on entration in the vi inity of the vesi le.The observed ation spe i� ity followed the lassi al Hofmeister series for ationadsorption to sulfate headgroups. When alkyl arboxylates were present in solution,the ations followed a reversed Hofmeister series. The ion spe i� ity was qualita-tively explained a ording to Collins' on ept of mat hing water a�nities. To thispurpose, the headgroup of an alkylsulfate had to be regarded as a haotrope andthe alkyl arboxylate as a kosmotrope. Based on MD simulations performed on theaforementioned headgroups, a general `Hofmeister series of surfa tant headgroups'is established in Chapter 3.In Chapter 4, the study of ion-spe i� ity was extended to the ationi -ri h regionof the phase diagram of the atanioni system. No vesi les were found, however,the addition of salts produ ed a sphero ylindri al growth of the mi elles, markedlydependent on the anion identity. The e� ien y of the anions to elongate the mi- elles ould again be explained by the Collins' on ept and the lassi� ation of thetrimethylammonium surfa tant headgroup as a soft, polarizable entity.The e�e t of various additives on the stability of mixed surfa tant solutions ofsodium dode ylsulfate with etyltrimethylammoniumbromide and with dode yltrimethy-lammonium bromide was studied as a fun tion of time and is reported in Chapter 5.The lifetime of vesi les and mi elles in these systems ould be ontrolled by varyingthe omposition of the surfa tant solutions and by additives. Controlling the pre- ipitation phenomena is of importan e for a large number of industrial pro esses,where formulations need to be tuned. These spe i� mixtures were shown to havea solubility temperature below that of pure surfa tant solutions in the anioni -ri hregion. The stability of su h supersaturated solutions was in reased by in reasingthe per entage of the anioni surfa tant and shortening the hain of the ationi surfa tant. The presen e of simple ele trolytes de reased, while the addition ofmiddle- hain al ohols in reased their stability. Experimental results suggested thatthe destabilization and on urrent pre ipitation of the systems was not due to theformation of bigger aggregates, but rather to a shift of the equilibrium between mi- elles and monomers. The addition of low on entrations of middle- hain al oholsresulted in the formation of vesi les in the atanioni systems at temperatures sig-134

ni� antly lower than the Kra�t temperatures of the parti ipating ioni surfa tants.In Chapter 6 we evaluated the ytotoxi ity of single- hain ioni surfa tants and atanioni mixtures using MTT on HeLa ells. It is important to onsider the tox-i ity of parti ipating surfa tants when formulating new en apsulating systems foreither osmeti or medi al use. It was on�rmed that anioni surfa tants gener-ally exhibit higher IC50 values (lower toxi ity) than atanioni ones. The toxi ity ould further be in�uen ed by the hydro arbon hain length and the presen e ofpolyoxyethylene groups. A general trend of in reasing toxi ity with in reasing hainlength and in lusion of polyoxyethylene groups was observed. A non-linear in reasewas observed as ationi surfa tants are added to anioni ones. A steep de reaseof the IC50 values is observed already at low fra tions of the ationi surfa tants,suggesting that in potential drug delivery systems a high fra tion of the anioni omponent is ne essary.Seeing that the presen e of only a small amount of a ationi surfa tant in themixture resulted in a large in rease in its toxi ity, we fo used on a mixture of twoanioni and in osmeti formulation already ommonly used surfa tants. In hapter7 we present a way to form vesi les in a mixture of two alkyl arboxylates with twovery di�erent pKa values. We have shown that it is then possible to spontaneouslyform vesi les in a pH range between the two onsidered pKa. Sodium laureate andAkypo Soft 45 NV ould be thus used at di�erent mixing ratios to obtain vesi ularsolutions over the entire range of pH omparable to skin or physiologi al pH.

135

136

Bibliography[1℄ K. Holmberg, B. Joensson, B. Kronberg, and B. Lindman. Surfa tants andPolymers in aqueous solutions, 2nd ed. John Wiley & Sons, 2003.[2℄ M. J. Rosen. Surfa tants and Interfa ial Phenomena (2nd ed.). John Wiley& Sons, 1989.[3℄ D. F. Evans and H. Wennerström. The Colloidal Domain. Wiley-VCH, 1999.[4℄ B. Lindman and H. Wennerström. Topi s in Current Chemistry, Vol. 87.Springer-Verlag, Germany, 1980.[5℄ K. Meguro, M. Ueno, and K. Esumi. Nonioni Surfa tants, Physi al Chem-istry. Mar el Dekker, New York, 1987.[6℄ K. Shinoda, T. Nakagawa, B. I. Tamamush, and T. Isemura. Colloidal surfa -tants, Some physi o- hemi al properties. A ademi Press, London, 1963.[7℄ G. Gunnarsson and H. Jönsson, B.and Wennerström. J. Phys. Chem., 84:3114,1980.[8℄ J. Goodwin. Colloids and Interfa es with Surfa tants and Polymers. JohnWiley & Sons, 2004.[9℄ J. H. Fendler. Membrane Mimeti Chemistry. Wiley, New York, 1982.[10℄ W. Helfri h. Z. Naturfors h., 28:693, 1973.[11℄ M. Kléman. Pro . R. So . London, Ser. A, 347:387, 1976.[12℄ W. Helfri h. J. Phys., 47:321, 1986.137

[13℄ S. A. Safran, P. Pin us, and D. Andelman. S ien e, 248:354, 1990.[14℄ J. N. Israela hvili, D. J. Mit hell, and B. W. Ninham. Bio hem. Biophys.A ta, 470:185, 1977.[15℄ J. N. Israela hvili, D. J. Mit hell, and B. W. Ninham. J. Chem. So ., FaradayTrans. 2, 72:1525, 1976.[16℄ D. J. Mit hell and B. W. Ninham. J. Chem. So ., Faraday Trans. 2, 77:601,1981.[17℄ M. Dubois and T. Zemb. Langmuir, 7:1352, 1991.[18℄ E. F. Marques, O. Regev, A. Khan, and B. Lindman. Adv. Coll. Interf. S i.,100:83, and refs. therein, 2003.[19℄ D.D. Miller, J.R. Bellare, T. Kaniko, and D.F. Evans. Langmuir, 4:1363, 1988.[20℄ B.W. Ninham, D.F. Evans, and G.J. Wei. J. Phys. Chem., 87:5020, 1983.[21℄ Y. Talmon, D. F. Evans, and B. W. Ninham. S ien e, 221:1047, 1983.[22℄ M.I. Viseu, M.M. Velazquez, C.S. Campos, I. Gar ia-Mateos, and S.M.B.Costa. Langmuir, 16:4882, 2000.[23℄ B. A. Coldren, H. Warriner, R. van Zanten, and J. A. Zasadzinski. Langmuir,22:2465, 2006.[24℄ E. W. Kaler, A. K. Murthy, B. E. Rodriguez, and J. A. Zasadzinski. S ien e,245:1371, 1989.[25℄ A. Kihisa and T. A. Hatton. Langmuir, 18:7341, 2002.[26℄ M. T. Yat illa, K. L. Herrington, L. L. Brasher, E. W. Kaler, S. Chiruvolu,and J. A. Zasadzinski. J. Phys. Chem., 100:5874, 1996.[27℄ M. Bergmeier, H. Ho�mann, and C. Thunig. J. Phys. Chem. B, 101:5767,1997. 138

[28℄ M. Bergmeier, H. Ho�mann, F. Witte, and S. Zouraub. J. Colloid Interfa eS i., 203:1, 1998.[29℄ M. Gradzielski, M. Bergmeier, M. Mueller, and H. Ho�mann. J. Phys. Chem.B, 101:1719, 1997.[30℄ M. Gradzielski, M. Mueller, M. Bergmeier, H. Ho�mann, and E. Hoinkis. J.Phys. Chem. B, 103:1416, 1999.[31℄ J. Hao, H. Ho�mann, and K. Horbas hek. Langmuir, 17:290, 2001.[32℄ J. Huang, Y. Zhu, B. Zhu, R. Li, and H. Fu. J. Colloid Interfa e S i., 236:201,2001.[33℄ R. Oda and L. Bourdieu. J. Phys. Chem. B, 101:5913, 1997.[34℄ M. Andersson, L. Hammarström, and K. Edwards. J. Phys. Chem., 99:14531,1995.[35℄ A. M. Carmona-Ribeiro. Chem. So . Rev., 21:207, 1992.[36℄ J.B.F.N. Engberts and D. Hoeskstra. Bio him. Biopphys. A ta, 1241:323,1995.[37℄ E. Feitosa and G. Barreleiro, P.C.A and; Olofsson. Chem. Phys. Lipids,105:201, 2000.[38℄ E. Feitosa and W. Brown. Langmuir, 13:4810, 1997.[39℄ P. Saveyn, J. Co quyt, P. Bomans, P. Frederik, M. De Cuyper, and P. Van derMeeren. Langmuir, 23:4775, 2007.[40℄ H. Hauser. Chem. Phys. Lipid, 43:283, 1987.[41℄ N. E. Gabriel and M. F. Roberts. Bio hemistry, 23:4011, 1984.[42℄ A. K. Murthy, E. W. Kaler, and J. A. Zasadzinski. J. Colloid Interf. S i.,145:598, 1991. 139

[43℄ J. E. Brady, D. F. Evans, B. Ka har, and B. W. Ninham. J. Am. Chem. So .,106:4279, 1984.[44℄ S. Hashimoto, J. K. Thomas, D. F. Evans, S. Mukherjee, and B. W. Ninham.J. Colloid Interfa e S i., 95:594, 1983.[45℄ B. W. Ninham and D. F. Evans. Faraday Dis uss. Chem. So ., 81:1, 1986.[46℄ Z. Lin, M. He, L. E. S riven, and H. T. Davis. J. Phys. Chem., 97:3571, 1993.[47℄ L. L. Brasher, K. L. Herrington, and E. Kaler. Langmuir, 11:4267, 1995.[48℄ K. L. Herrington, E. W. Kaler, D. D. Miller, J. A. Zasadzinski, and S. Chiru-volu. J. Phys. Chem., 97:13792, 1993.[49℄ E. W. Kaler, K. L. Herrington, A. K. Murthy, and J. A. N. Zasadzinski. J.Phys. Chem., 96:6698, 1992.[50℄ Y. Kondo, H. U hiyama, N. Yoshino, K. Nishiyama, and M. Abe. Langmuir,11:2380, 1995.[51℄ G. Gregoriadis. Liposome Te hnology. Volume II: Entrapment of Drugs andOther Materials, 2nd ed. CRC Press: Bo a Raton, FL, 1988.[52℄ D. D. Lasi . Liposomes: From Physi s to Appli ations. Elsevier PublishingCo., Amsterdam, 1993.[53℄ M. Dubois and T. Zemb. Curr. Opin. Colloid Interfa e S i., 5:27, 2000.[54℄ R.D. Koehler, S.R. Raghavan, and E.W. Kaler. J. Phys. Chem. B, 104:11035,2000.[55℄ X. Li and H. Kunieda. Curr. Opin. Colloid Interfa e S i., 8:327, 2003.[56℄ E. Marques, A. Khan, M. Miguel, and B. Lindman. J. Phys. Chem., 97:4729,1993.[57℄ S.R. Raghavan, G. Fritz, and E.W. Kaler. Langmuir, 18:3797, 2002.140

[58℄ A. Khan. Curr. Opin. Colloid Interfa e S i., 1:614, 1996.[59℄ P. K. Yuet and D. Blanks htein. Langmuir, 12:3802, 1996.[60℄ C. Tanford. The Hydrophobi E�e t: Formation of Mi elles and Biologi alMembranes (2nd ed.). John Wiley & Sons, New York, 1980.[61℄ W. R. Hargreaves and D. W. Deamer. Bio hemistry, 17:3759, 1978.[62℄ C. Tondre and C. Caillet. Adv. Colloid Interfa e S i., 93:115, 2001.[63℄ P. M. Gullino, F. H. F. H. Grantham, S. H. Smith, and M. Haggerty. J. Natl.Can er Inst., 34:857, 1965.[64℄ H. Kahler and W. V. B. Robertson. J. Natl. Can er Inst., 3:495, 1943.[65℄ K. A. Meyer, E. M. Kummerling, L. Altman, M. Koller, and S. J. Ho�man.Can er Res., 8:513, 1948.[66℄ J. Naeslund and K. E. Swenson. A ta Obstet. Gyne ol. S and., 32:359, 1953.[67℄ M. Stubbs, P. M. J. M Sheehy, J. R. Gri�ths, and C. L. Bashford. Mol. Med.Today, 6:15, 2000.[68℄ M. C. De Oliveira, E. Fattal, C. Ropert, C. Malvy, and P. Couvreur. Journalof Controlled Release, 48:179, 1997.[69℄ I. M. Hafez, S. Absell, and P. R. Cullis. Biophys. J., 79:1438, 2000.[70℄ L. Huang and C. Y. Wang. Bio hemistry, 28:9508, 1989.[71℄ X. Li and M. S hi k. Biophys. J., 80:1703, 2001.[72℄ C. Ropert, M. Lavignon, C. Dubernet, P. Couvreur, and C. Malvy. Bio hem.Biophys. Res. Commun., 183:879, 1992.[73℄ S. Simões, J. Nuno Moreira, C.; Fonse a, N. Duzgunes, and M. C. Pedroso deLima. Adv. Drug Deliv. Rev., 56:947, 2004.141

[74℄ S. Simões, V. Slepushkin, N. Duzgunes, and M. C. Pedroso de Lima. Bio him.Biophys. A ta, 1515:23, 2004.[75℄ J.J. Sudima k, W. Guo, W. Tjarks, and R.J. Lee. Bio him. Biophys. A ta,1564:31, 2002.[76℄ V.P. Tor hilin, F. Zhou, and L. Huang. J. Liposome Res., 3:201, 1993.[77℄ P. Venugopalan, S. Jain, S. Sankar, P. Singh, A. Rawat, and S. P. Vyas.Pharmazie, 57:659, 2002.[78℄ M. B. Yatvin, W. Kreutz, B. A. Horwitz, and M. Shinitzky. S ien e, 210:1253,1980.[79℄ I. A. Chen and J. W. Szostak. Pro . Natl. A ad. S i., 101:7965, 2004.[80℄ I. Chen and J. W. Szostak. Biophys. J., 82:988, 2004.[81℄ K. Morigaki, S. Dallavalle, P. Walde, S. Colonna, and P. L. Luisi. J. Am.Chem. So ., 119:292, 1997.[82℄ C.G. Goltner and M. Antonietti. Adv. Mat., 9:431, 1997.[83℄ R. Hoss and F. Voegtle. Angew. Chem., 106:389, 1994.[84℄ S. Mann, S.L. Burkett, S. A. Davis, C.E. Fowler, N.H. Mendelson, S.D. Sims,D. Walsh, and N.T. Whilton. Chemistry of Materials, 9:2300, 1997.[85℄ C.A. M Kelvey, E.W. Kaler, J.A. Zasadzinski, B. Coldren, and H.-T. Jung.Langmuir, 16:8285, 2000.[86℄ Z. Yuan, Z. Yin, S. Sun, and J. Hao. J. Phys. Chem. B, 112:1414 � 1419, 2008.[87℄ A. Bonin ontro, M. Falivene, C. La Mesa, G. Risuleo, and M. R. Pena. Lang-muir, 24:1973, 2008.[88℄ J. D. Bernal and R. H. Fowler. J. Chem. Phys., 1:515, 1933.[89℄ R. W. Gurney. Ioni Pro esses in Solution. M Graw-Hill Book Co., 1953.142

[90℄ G. A. Krestov. Thermodynami s of Solvation. Ellis Horwood: New York,1990.[91℄ R. A. Robinson and R. H. Stokes. Ele trolyte Solutions. Butterworth S ienti� Publi ations: London, 1959.[92℄ O. Y. Samoilov. Water and Aqueous Solution: Stru ture, Thermodynami s,and Transport Pro esses, pages 597�612. Wiley-Inters ien e: New York, 1972.[93℄ O. Y. Samoilov. Dis uss. Faraday So , 24:141, 1957.[94℄ G. Galli and M. Parrinello. Computer Simulations in Materials S ien e.Kluwer: Dordre ht, 1991.[95℄ K. Heinzinger and P. C. Vogel. Z. Naturfors h., 29:1164, 1974.[96℄ G. Hummer, L. R. Pratt, and A. E. Gar ia. J. Phys. Chem. A, 102:7885, 1998.[97℄ D. Marx, M. Sprik, M. Sprik, and M. Parinello. Chem. Phys. Lett., 273:360and refs therein, 1997.[98℄ M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias, and J. D. Joannopoulos.Rev. Mod. Phys., 64:1045, 1992.[99℄ W. Kauzmann. Adv. Protein Chem., 14:1, 1959.[100℄ M. Kaminsky. Dis uss. Faraday So ., 24:171, 1957.[101℄ K. D. Collins. Methods, 34:300, 2004.[102℄ N. T. Skipper and G. W. Neilson. J. Phys.: Condens. Matter, 1:4141, 1989.[103℄ K. D. Collins. Biophys. J., 72:65, 1997.[104℄ P. Debye and B. Hü kel. Phys. Z., 24:185, 1923.[105℄ K.D. Collins. Pro .Natl.A ad. S i., 92:5553, 1995.[106℄ K. D. Collins and M. W. Washabaugh. J. Biol. Chem., 261:12477, 1986.[107℄ K. D. Collins and M. W. Washabaugh. Q. Rev. Biophys., 18:323, 1985.143

[108℄ S. Gopalakrishnan, D. Liu, H.C. Allen, M. Kuo, and M. J. Shultz. Chem.Rev., 106:1155, 2006.[109℄ P.B. Petersen and R.J. Saykally. Annu. Rev. Phys. Chem., 57:333, 2006.[110℄ J. Cheng, C.D. Ve itis, M.R. Ho�mann, and A. J. Colussi. J. Phys. Chem. B,110:25598, 2006.[111℄ P. Jungwirth and D. J. Tobias. Chem. Rev., 106:1259, 2006.[112℄ L. Vrbka, M. Mu ha, B. Minofar, P. Jungwirth, E. C. Brown, and D. J. Tobias.Curr. Opin. Colloid Interfa e S i., 9:67, 2004.[113℄ M. Bostrom, F.W. Tavares, B.W. Ninham, and J.M. Prausnitz. J. Phys.Chem. B, 110:24757, 2006.[114℄ F. Hofmeister. Ar h. Exp. Pathol. Pharmakol., 24:247, 1888.[115℄ W. Kunz, P. Lo Nostro, and B. W. Ninham. Curr. Opin. Colloid Interfa eS i., 9:1, 2004.[116℄ Y. Mar us. Ion Solvation. Wiley: Chi hester, U.K., 1985.[117℄ Y. Mar us and G. Hefter. Chem. Rev., 106:4585, 2006.[118℄ Y. V. Kalyuzhnyi, V. Vla hy, and K. Dill. A ta Chim. Slov., 48:309, 2001.[119℄ T. V. Chalikian, J. Volker, E. Plum, and K. Breslauer. Pro . Natl. A ad. S i.,96:7853, 1999.[120℄ J. Darnell, H. Lodish, and D. Baltimore. Mole ular Cell Biology. S ienti� Ameri an Books, New York, 1990.[121℄ K. A. Dill. Bio hemistry, 29:7133, 1990.[122℄ N. Korolev, A. P. Lyubartsev, A. Ruppre ht, and L. Nordenskiold. Biophys.J., 77:837, 1999.[123℄ J. Rupley and G. Careri. Adv. Protein Chem., 41:37, 1991.144

[124℄ T. P. Lybrand, A. M Cammon, and G. Wi�. Pro . Natl. A ad. S i., 83:833,1986.[125℄ F. Sussman and H. Weinstein. Pro . Natl. A ad. S i., 86:7880, 1989.[126℄ P. C. Jordan. Biophys. J., 58:1133, 1990.[127℄ B. Katz. Nerve, Mus le, and Synapse. M Graw-Hill: London, 1966.[128℄ M. G. Ca a e, E. M. Landau, and J. J. Ramsden. Q. ReV. Biophys., 30:241,1997.[129℄ M. F. Kropman and H. J. Bakker. S ien e, 291:2118, 2001.[130℄ M. Maron elli, J. Ma Innins, and G. R. Fleming. S ien e, 243:1674, 1989.[131℄ V. L. Larwood, B. J. Howlin, and G. A. Webb. J. Mol. Model., 2:175, 1996.[132℄ S. Habu hi, H. B. Kim, and N. Kitamura. Anal. Chem., 73:366, 2001.[133℄ M. Chavez-Paez, K. Van Workum, L. de Pablo, and L. L. de Pablo. J. Chem.Phys., 114:1405, 2001.[134℄ K. L. Stellner, J. C. Amante, J. F. S amehorn, and J. H. Harwell. J. ColloidInterfa e S i., 123:186, 1988.[135℄ L. S. Romsted. Langmuir, 23:414, 2007.[136℄ H. Fukuda, K. Kawata, and H. Okuda. J. Am. Chem. So ., 112:1635, 1990.[137℄ Th. Zemb, M. Dubois, B. Demé, and T. Gulik-Krzywi ki. S ien e, 283:816,1999.[138℄ M. Dubois, B. Demé, T. Gulik-Krzywi ki, J.-C. Dedieu, C. Vautrin, S. Désert,E. Perez, and Th. Zemb. Nature, 411:672, 2001.[139℄ A. Song, S. Dong, X. Jia, J. Hao, W Liu, and T. Liu. Angew. Chem. Int. Ed.,44:4018, 2005. 145

[140℄ P. André, A. Filankembo, I. Lisie ki, C. Petit, T. Gulik-Krzywi ki, B.W. Nin-ham, and M. P. Pileni. Adv. Mat., 12:119, 2000.[141℄ S. T. Hyde, S. Andersson, K. Larsson, Z. Blum, T. Landh, S. Lidin, and B. W.Ninham. The Language of Shape. Elsevier, Amsterdam, 1997.[142℄ P. André, B. W. Ninham, and M. P. Pileni. New J. Chem., 25:563, 2001.[143℄ P. André, B. W. Ninham, and M. P. Pileni. Adv. Coll. Interf. S i., 89:155,2001.[144℄ D. Miller, J. Bellare, D. F. Evans, Y. Talmon, and B. W. Ninham. J. Phys.Chem., 91:674, 1987.[145℄ B. W. Ninham, S. Hashimoto, and J. K. Thomas. J. Colloid Interfa e S i.,95:594, 1983.[146℄ B. W. Ninham, Y. Talmon, and D.F. Evans. S ien e, 221:1047, 1983.[147℄ J. L. Parker, H. K. Christenson, and B. W. Ninham. J. Phys. Chem., 92:4155,1988.[148℄ E. Z. Radlinska, B. W. Ninham, J. P. Dalbiez, and T. N. Zemb. J. Coll. Surf.,46:213, 1990.[149℄ E. Z. Radlinska, T. N. Zemb, J. P. Dalbiez, and B. W. Ninham. Lamgmuir,9:2844, 1993.[150℄ T. Bramer, M. Paulsson, K. Edwards, and K. Edsman. Pharm. Res., 20:1661,2003.[151℄ H. P. Hentze, S. R. Raghavan, C. A. M Kelvey, and E. W. Kaler. Langmuir,19:1069, 2003.[152℄ M. Kep zynski, F. Gana haud, and P. Hemery. Adv. Mat., 16:1861, 2004.[153℄ J. C. Amante, J. F. S amehorn, and J. H. Harwell. J. Colloid Interfa e S i.,144:243, 1991. 146

[154℄ N. Kamenka, M. El Amrani, J. Appell, and M. Lindheimer. J. Colloid Interfa eS i., 143:463, 1991.[155℄ A. Renon ourt, P. Bauduin, D. Touraud, W. Kunz, N. Azemar, and C. Solans.Comuni a iones presentadas a las Jornadas del Comite Espanol de la Deter-gen ia, 34:273, 2004.[156℄ N. Filipovi -Vin ekovi , M. Bujan, I. Smit, Lj. Tusek-Bozi , and I. Stefani .J. Colloid Interfa e S i., 201:59, 1998.[157℄ A. J. O'Connor, T. A. Hatton, and A. Bose. Langmuir, 13:6931, 1997.[158℄ O. Soederman, K. L. Herrington, E. W. Kaler, and D. D. Miller. Langmuir,13:5531, 1997.[159℄ H. Yin, Z. Zhou, J. Huang, R. Zheng, and Y. Zhang. Angew. Chem., 42:2188,2003.[160℄ H. Yin, S. Lei, S. Zhu, J. Huang, and J. Ye. Chem. Eur. J., 12:2825, 2006.[161℄ L. Hao, Y. Nan, H. Liu, and Y. Hu. J. Dispersion S i. and Te hnology, 27:271,2006.[162℄ L. Zhai, G. Li, and Z. Sun. Colloids Surf., A, 190:275, 2001.[163℄ L. Zhai, M. Zhao, D. Sun, J. Hao, and L. Zhang. J. Phys. Chem. B, 109:5627,2005.[164℄ A. Renon ourt. PhD thesis, University of Regensburg, 2005.[165℄ M. Bergström and J. S. Pedersen. Langmuir, 15:2250, 1999.[166℄ E. F. Marques, A. Khan, and B. Lindman. Thermo himi a A ta, 394:31, 2002.[167℄ K. Tsu hiya, H. Nakanishi, H. Sakai, and M. Abe. Langmuir, 20:2117, 2004.[168℄ C. Vautrin, M Dubois, Th. Zemb, St. S hmölzer, H. Ho�man, M. ColloidsGradzielski, and 165. Surfa es A: Physi o hem. Eng. Aspe ts 2003, 217. Col-loids Surf., A, 217:165, 2003. 147

[169℄ A. Renon ourt, N. Vla hy, P. Bauduin, M. Dre hsler, D. Touraud, J.-M. Ver-bavatz, M. Dubois, W. Kunz, and B. W. Ninham Langmuir 23 (2007) 2376.Langmuir, 23:2376, 2007.[170℄ K. Edwards, J. Gustafsson, M. Almgren, and G. Karlsson. J. Colloid Interfa eS i., 161:161, 1993.[171℄ J. Gustafsson, G. Orädd, G. Lindblom, U. Olsson, and M. Almgren. Langmuir,13:852, 1997.[172℄ J. Gustafsson, G. Orädd, M. Nydén, P. Hansson, and M. Almgren. Langmuir,14:4987, 1998.[173℄ H. Ho�mann, C. H. Thunig, U. Munkert, H. W. Meyer, and W. Ri hter.Langmuir, 8:2629, 1992.[174℄ M. Silvander, G. Karlsson, and K. Edwards. J. Colloid Interfa e S i., 179:104,1996.[175℄ A.-L. Bernard, M.-A. Guedeau-Boudeville, L. Jullien, and J.-M. di Meglio.Bio him. Biophys. A ta, 1567:1, 2002.[176℄ C. Ménager and V. Cabuil. J. Phys. Chem. B, 106:7913, 2002.[177℄ J. B. F. N. Sein, A.and Engberts. Langmuir, 11:455, 1995.[178℄ S. Chiruvolu, S. A. Walker, J. Israela hvili, F. J. S hmitt, D. Le kband, andJ. A. Zasadzinski. S ien e, 264:1753, 1994.[179℄ S. A. Walker, M. T. Kennedy, and J. A. Zasadzinski. Nature, 387:61, 1997.[180℄ S. A. Walker and J. A. Zasadzinski. Langmuir, 13:5076, 1997.[181℄ H. Aranda-Espinoza, Y. Chem, N. Dan, T. C. Lubensky, P. Nelson, L. Ramos,and D. A. Weitz. S ien e, 1999:285, 394.[182℄ A. A. Gurtovenko. J. Chem. Phys., 122:244902, 2005.[183℄ B. Hribar and V. Vla hy. J. Phys. Chem. B, 101:3457, 1997.148

[184℄ A. Walter, P. K. Vinson, A. Kaplun, and Y. Talmon. Biophys. J., 60:1315,1991.[185℄ B. Demé, M. Dubois, T. Gulik-Krzywi ki, and Th. Zemb. Langmuir, 18:997,2002.[186℄ M. Dubois, V. Lizunov, T A. Meister, Gulik-Krzywi ki, J.-M. Verbavatz,E. Perez, J. Zimmerberg, and Th. Zemb. Pro . Natl. A ad. S i., 101:15082,2004.[187℄ H. W. Meyer, W. Ri hter, and J. Gumpert. Bio him. Biophys. A ta, 1026:171,1990.[188℄ B. Sternberg, J. Gumpert, G. Reinhardt, and K. Gawris h. Bio him. Biophys.A ta, 898:223, 1987.[189℄ H. W. Meyer, H. Bunjes, and A. S. Ulri h. Chem. Phys. Lipids, 99:111, 1999.[190℄ E.; Sa kmann and T. Feder. Mol. Membr. Biol., 12:21, 1995.[191℄ V. S. Kulkarni, W. H. Anderson, and R. E. Brown. Biophys. J., 69:1976, 1995.[192℄ M. Antonietti, A. Wenzel, and A. Thünemann. Langmuir, 12:2111, 1996.[193℄ H. W. Meyer, K. Semmler, W. Rettig, W. Pohle, A. S. Ulri h, S. Grage,C. Selle, and P. Quinn. J. Chem. Phys. Lipids, 105:149, 2000.[194℄ C. Gebhardt, H. Gruler, and E. Sa kmann. Zeits hrift für Naturfors hung,32:581, 1977.[195℄ W. Helfri h. Liquid Crystals, 5:1647, 1989.[196℄ P. G. Dommersnes and J. B. Fournier. Biophys. J., 83:2898, 2002.[197℄ J. B. Fournier. Phys. Rev. Lett., 76:4436, 1996.[198℄ R. Goetz and W. Helfri h. J. Phys. II, 6:215, 1996.[199℄ K. D. Collins. Biophysi al Chemistry, 119:271, 2006.149

[200℄ K. D. Collins, G. W. Neilson, and J. E. Enderby. Biophysi al Chemistry,128:95, 2007.[201℄ B. Jagoda-Cwiklik, R. Vá ha, M. Lund, M. Srebro, and P. Jungwirth. J. Phys.Chem. B (2007), doi: 10.1021/jp709634t, doi: 10.1021/jp709634t, 2007.[202℄ I. Boime, E. E. Smith, and F. E. Hunter. Ar h. Bio hem. Biophys., 139:425,1970.[203℄ G. J. van de Vusse, R. S. Reneman, and M. van Bilsen. Prostaglandins,Leukotrienes Essent. Fatty A ids, 57:85, 1997.[204℄ A. Jakubowska. ChemPhysChem, 6:1600, 2005.[205℄ M. Yamashita and J. B. Fenn. J. Phys. Chem., 88:4671, 1984.[206℄ C. Gamboa and L. Sepulveda. J. Colloid Interfa e S i., 113:566, 1986.[207℄ S. Ikeda, S. Hayashi, and T. Imae. J. Phys. Chem., 85:106, 1981.[208℄ M. L. Corrin and W. D. Harkins. J. Am. Chem. So ., 69:683, 1947.[209℄ N. A. Mazer, G. B. Benedek, and M. C. Carey. J. Phys. Chem., 80:107, 1976.[210℄ L. Vrbka, P. Jungwirth, P. Bauduin, D. Touraud, and W. Kunz. J. Phys.Chem. B, 110:7036, 2006.[211℄ E. Leontidis. Curr. Opin. Colloid Interfa e S i., 7:81 and refs. therein, 2002.[212℄ H. Gustavsson and B. Lindman. J. Am. Chem. So ., 100:4647, 1978.[213℄ N. Hedin, I. Furo, and P. O. Eriksson. J. Phys. Chem. B, 140:8544, 2000.[214℄ V. E. Haverd and G. G. Warr. Langmuir, 16:157 and refs. therein, 2000.[215℄ H. P. Gregor and M. J. Frederi k. Polymer S i., 23:451, 1957.[216℄ K. Huts hneker and H. Deuel. Helv. Chim. A ta, 39:1038, 1956.[217℄ U. P. Strauss. Polyele trolytes, page 79. D. Reidel Publishing Company,Dordre ht-Holland, 1974. 150

[218℄ U. P. Strauss and Y. P. Leung. J. Am. Chem. So ., 87:1476, 1965.[219℄ H. P. Gregor, M. J. Hamilton, R. J. Oza, and F. Bernstein. J. Phys. Chem.,60:263, 1956.[220℄ F. Hel�eri h. Ion Ex hange. M Graw-Hill Book Co., In ., New York, N.Y.,1962.[221℄ H. G. Bungenberg de Jong. Colloid S ien e, Vol. II, hapter 9. ElsevierPublishing Co., In ., New York, N. Y., 1949.[222℄ G. Eisenman. Symposium on Membrane Transport and Metabolism, page 163.A ademi Press, New York, N. Y., 1961.[223℄ M. E. Feinstein and H. L. Rosano. J. Colloid Interfa e S i., 24:73, 1967.[224℄ I. Weil. J. Phys. Chem., 70:133, 1966.[225℄ H. L. Rosano, A. P. Christodoulou, and M. E. Feinstein. J. Colloid Interfa eS i., 29:335, 1969.[226℄ R. M. Fuoss, A. Kat halsky, and S. Lifson. Pro . Natl. A ad. S i. U. S., 37:579,1951.[227℄ G. S. Manning. J. Chem. Phys., 51:924, 1969.[228℄ H. Daoust and M.-A. Chabot. Ma romole ules, 13:616, 1980.[229℄ D. Dolar. Polyele trolytes, page 97. Reidel Publishing Company, Dordre ht-Holland, 1974.[230℄ I. Lipar, P. Zalar, C. Pohar, and V. Vla hy. J. Phys. Chem. B, 111:10130,2007.[231℄ G. E. Boyd and B. A. Soldano. Z. Elektro hem., 57:162, 1953.[232℄ L. Xu, X. Li, M. Zhai, L. Huang, J. Peng, J. Li, and G. Wei. J. Phys. Chem.B, 111:3391, 2007. 151

[233℄ H. P. Gregor, L. B. Luttinger, and E. M. Loebl. J. Am. Chem. So ., 76:5879,1954.[234℄ H. Gustavsson and B. Lindman. J. Am. Chem. So ., 97:3923, 1975.[235℄ J. M. Chialvo, A. A.; Simonson. J. Phys. Chem. B, 109:23031, 2005.[236℄ N. Vla hy, B. Jagoda-Cwiklik, R. Vá ha, D. Tourauda, P. Jungwirth, andW. Kunz. 2008, submitted.[237℄ M. M. A. E. Claessens. Size regulation and stability of harged lipid vesi les,Do toral dissertation. PhD thesis, 2003.[238℄ H. Ti Tien. J. Phys. Chem., 68:1021, 1964.[239℄ M. M. A. E. Claessens, B. F. van Oort, F. A. M. Leermakers, F. A. Hoekstra,and M. A. Cohen Stuart. Biophys. J., 87:3882, 2004.[240℄ J. A. W. M. Beena kers, B. F. M. Kuster, and H. S. . van der Baan. Carbohydr.Res., 140:169, 1985.[241℄ W. Kunz, P. Calmettes, T. Cartailler, and P. Turq. J. Chem. Phys., 99:2074,1993.[242℄ P. Bauduin, A. Renon ourt, A. Kopf, D. Touraud, and W. Kunz. Langmuir,21:6769, 2005.[243℄ J Narayanan, C. Manohar, D. Langevin, and W. Urba h. Langmuir, 13:398,1997.[244℄ S. R. Raghavan, H. Edlund, and E. W. Kaler. Langmuir, 18:1056, 2002.[245℄ J. F. A. Soltero, J. E. Puig, O. Manero, and P. C. S hulz. Langmuir, 11:3337,1995.[246℄ H. Wu, S. Kawagu hi, and I. Koi hi. Colloid Polym. S i., 283:636, 2005.[247℄ S. Friberg and I. Blute. Hydrotropy, Surfa tant S ien e Series, Liquid Deter-gents 2nd ed. Dekker, New York, 2006.152

[248℄ S. Berr, R. R. M. Jones, and J. S. Johnson Jr. J. Phys. Chem., 96:5611, 1992.[249℄ G. Biresaw, D. C. M Kenzie, C. A. Bunton, and D. F. Ni oli. J. Phys. Chem.,89:5144, 1985.[250℄ M. S. Vethamuthu, M. Almgren, E. Mukhtar, and P. Bahadur. Langmuir,8:2396, 1992.[251℄ P. H. von Hippel and T. S hlei h. Biologi al Ma romole ules, Vol. 3. Mar elDekker, New York, N. Y., 1969.[252℄ S. Gravsholt. J. Colloid Interfa e S i., 57:575, 1976.[253℄ T. Imae, R. Kamiya, and S. Ikeda. J. Colloid Interfa e S i., 108:215, 1985.[254℄ C. Gamboa, H. Ríos, and S. J. Sepúlveda. J. Chem. Phys., 93:5540, 1989.[255℄ A. J. Hyde and D. W. M. Johnstone. J. Colloid Interfa e S i., 53:349, 1975.[256℄ J. W. Larsen and L. J. Magid. J. Am. Chem. So ., 96:5774, 1974.[257℄ I. Cohen and T. Vassiliades. J. Phys. Chem., 65:1774, 1961.[258℄ H. Fabre, N. Kamenka, A. Khan, G. Lindblom, B. Lindman, and G. J. T.Tiddy. J.Phys.Chem., 84:3428, 1980.[259℄ N. Jiang, P. Li, Y. Wang, J. Wang, H. Yan, R. K. Thomas, I. Chakraborty,and S. P. Moulik. J. Phys. Chem. B, 111:3658, 2007.[260℄ G. Para, E. Jarek, and P. Warszynski. Adv. Colloid Interfa e S i., 122:39,2006.[261℄ G. Para, E. Jarek, and P. Warszynski. Colloids Surf., A, 261:65, 2005.[262℄ U. Henriksson, L. Ödberg, J. C. Eriksson, and L. Westman. J. Phys. Chem.,81:76, 1977.[263℄ G. Lindblom, B. Lindman, and L. Mandell. J. Colloid Interfa e S i., 42:400,1973. 153

[264℄ F. Reiss-Husson and V. Luzzati. J. Phys. Chem., 68:3504, 1964.[265℄ J. Ulmius, B. Lindman, G. Lindblom, and T. Drakenberg. J. Colloid Interfa eS i., 65:88, 1978.[266℄ J. Ulmius and H. Wennerström. J. Magn. Reson., 28:309, 1977.[267℄ A. Abe, T. Imae, and S. Ikeda. Colloid Polym. S i., 265:637, 1987.[268℄ J. K. Ba kus and H. A. S heraga. J. Colloid Interfa e S i., 6:508, 1951.[269℄ S. J. Candau, E. Hirs h, and R. Zana. J. Colloid Interf. S i., 105:521, 1985.[270℄ S. J. Candau, E. Hirs h, and R. Zana. J. Physique, 45:1263, 1984.[271℄ P. Debye and E. W. Ana ker. J. Phys. Colloid Chem., 55:644, 1951.[272℄ R. Dorshow, J. Briggs, C. A. Bunton, and D. F. Ni oli. J. Phys. Chem.,86:2388, 1982.[273℄ R. B. Dorshow, C. A. Bunton, and D. F. Ni oli. J. Phys. Chem., 87:1409,1983.[274℄ S. Ikeda. Surfa tants in Solution, Vol. 2, page 825. Plenum, New York, 1984.[275℄ T. Imae, A. Abe, Y. Tagu hi, and S. Ikeda. J. Colloid Interfa e S i., 109:567,1986.[276℄ T. Imae and S. Ikeda. J. Phys. Chem., 90:5216, 1986.[277℄ S. Ozeki and S. Ikeda. J. Colloid Interfa e S i., 87:424, 1982.[278℄ H. A. S heraga and J. K. Ba kus. J. Am. Chem. So ., 73:5108, 1951.[279℄ N. A. Riegelman, M. K. Allawa, M. K. Hreno�, and L. A. Strait. J. ColloidS i., 12:208, 1958.[280℄ L. Sepulveda. J. Colloid Interfa e S i., 46:372, 1974.[281℄ B. C. Smith, L. C. Chou, and J. L. Zakin. J. Rheol., 38:73, 1994.154

[282℄ M. B. Sierra, P. V. Messina, M. A. Morini, J. M. Russo, G. Prieto, P. C.S hulz, and F. Sarmiento. Colloids Surf., A, 277:75, 2006.[283℄ V. K. Aswal, P. S. Goyal, and P. Thiyagarajan. J. Phys. Chem., 109:2469,1998.[284℄ R. Bu hner, C. Baar, P. Fernandez, S. S hrödle, and W. Kunz. J. Mol. Liq.,118:179, 2005.[285℄ M. C. Brum and J. F. Oliveira. Minerals Engineering, 20:945, 2007.[286℄ M. Nyman, J. M. Bieker, S. G. Thoma, and D. E. Trudell. J. Colloid Interfa eS i., 316:968, 2007.[287℄ B. Wu, S. D. Christian, and J. F. S amehorn. Prog. Colloid Polym. S i.,109:60, 1998.[288℄ A. Patist, S. G. Oh, R. Leung, and D. O. Shah. Colloids Surf., A, 176:3, 2000.[289℄ D. O. Shah. Mi elles, Mi roemulsions and Monolayers. Dekker, New York,1998.[290℄ A. Patist, J. R. Kani ky, P. K. Shukla, and D. O. Shah. J. Colloid Interfa eS i., 245:1 and refs. therein, 2002.[291℄ P. C. Hiemenz and R. Rajagopalan. Prin iples of Colloid and Surfa e Chem-istry, 3rd ed. Dekker, New York, 1997.[292℄ R. J. Hunter. Foundations of Colloid S ien e. Oxford Univ. Press, New York,1987.[293℄ H. Chakraborty and M. Sakar. Langmuir, 20:3551, 2004.[294℄ E. F. Marques. Langmuir, 16:4798, 2000.[295℄ J. M. Pea o k and J. E. Matijevi . J. Colloid Interfa e S i., 77:548, 1980.[296℄ R. G. Laughlin. Colloids Surf., A, 128:27, 1997.155

[297℄ D. Chu and J. K. Thomas. J. Am. Chem. So ., 108:6270, 1986.[298℄ K. Kogej and J. Skerjan . Langmuir, 15:4251, 1999.[299℄ C. Vautier-Giongo and B. L. Bales. J. Phys. Chem. B, 107 (2003) 5398:5398,2003.[300℄ A. R. Tehrani-Bagha, H. Bahrami, B. Movassagh, M. Arami, S. H. Amirshahi,and F. M. Menger. Colloids Surf., A, 307:121, 2007.[301℄ M.-P. Nieh, T. A. Harroun, V. A. Raghunathan, C. J. Glinka, and J. Katsaras.Biophys. J., 86:2615, 2004.[302℄ B. Sohrabi, H. Gharibi, S. Javadian, and M. Hashemianzadeh. J. Phys. Chem.B, 111:10069, 2007.[303℄ M. Bergström, J. S. Pedersen, P. S hurtenberger, and S. U. Egelhaaf. J. Phys.Chem. B, 103:9888, 1999.[304℄ I. Yaa ob and A. Bose. J. Colloid Interfa e S i., 178:638, 1996.[305℄ B. Lindman. Handbook of Applied Surfa e and Colloid Chemistry. John Wiley& Sons Ltd, England, 2002.[306℄ N. Vla hy, K. Kogej, D. Touraud, and W. Kunz. J. Colloid Interfa e S i.,315:445, 2007.[307℄ E. B. Abuin and J. C. S aiano. J. Am. Chem. So ., 106:6274, 1984.[308℄ D. S. Karpovi h and G. J. Blan hard. J. Phys. Chem., 99:3951, 1995.[309℄ D. H. Kim, S. G. Oh, and C. G. Cho. Colloid Polym. S i., 279:39, 2001.[310℄ R. J. Anderson, C. Delgado, D. Fisher, J. M. Cunningham, and G. E. Fran is.Analyti al Bio hemistry, 193:101, 1991.[311℄ S. Askolin, T. Nakari-Setälä, and M. Tenkanen. Appl. Mi robiol. Biote hnol.,57:124, 2001. 156

[312℄ J. B. S. Bonilha, R. M. Z. Georgetto, E. Abuin, E. Lissi, and F. Quina. J.Colloid Interfa e S i., 135:238, 1990.[313℄ P. Mukerjee, K. Mysels, and P. Kapauan. J. Phys. Chem., 71:4166, 1967.[314℄ W. D. Harkins, R. W. Mattoon, and R. Mittelman. J. Chem. Phys., 15:763,1947.[315℄ A. E. Alexander. Trans. Faraday So ., 38:54, 1942.[316℄ P. F. Grieger and C. A. Krauss. J. Am. Chem. So ., 70:3803, 1948.[317℄ K. Shinoda. J. Phys. Chem., 58:1136, 1954.[318℄ H. Nakayama, K. Shinoda, and E. Hut hinson. J. Phys. Chem., 70:3502, 1966.[319℄ S. Kaneshina, H. Kamaya, and I. Ueda. J. Colloid Interfa e S i., 83:589, 1981.[320℄ T. Okubo, K. Hiraiwa, S. Kinoshita, and M. Watanabe. Alternatives to AnimalTesting and Experimentation, 1:2, 1990.[321℄ M. C. S aife. International Journal of Cosmeti S ien e, 4:179, 1982.[322℄ J. Selling and B. Ekwall. Xenobioti a, 15:713, 1985.[323℄ A. Trivedi, N. Kitabake, and E. Doi. Agri ultural and Biologi al Chemistry,54:2961, 1990.[324℄ K. Chiba, I. Makino, J. Ohu hi, Y. Kasai, H. Kakishima, K. Tsukumo,T. U hiyama, E. Miyai, J. Akiyama, Y. Okamoto, H. Kojima, H. Okumura,Y. Tsurumi, M. Usami, K. Katoh, S. Sugiura, A. Kurishita, M. Sunou hi,A. Miyajima, M. Hayashi, and Y. Ohno. Toxi ology in Vitro, 13:189, 1999.[325℄ W. Yang and D. A osta. Toxi ol. Lett., 70:309, 1994.[326℄ O. P. Flint. Altern. Test. Lab. Anim., 18:11, 1990.[327℄ M. Cornelis, Ch. Dupont, and J. Wepierre. Toxi . in Vitro, 6:119, 1992.[328℄ R. L. Grant, C. Yao, D. Gabaldon, and D. A osta. Toxi ology, 76:153, 1992.157

[329℄ H. Kojima, J. Ohu hi, Y. Kasai, J. Okada, K. Tsukumo, H. Kakishima,E. Miyai, J. Akiyama, Y. Okamoto, M. Kotani, K. Inoue, M. Hibata,H. Okumura, M. Arashima, T. Atsumi, I. Makino, K. Chiba, Y. Ohno, andA. Takanaka. Alternatives to Animal Testing and Experimentation, 3:191,1995.[330℄ T. Ohno, Y.and Kaneko, T. Kobayashi, T. Inoue, Y. Kuroiwa, T. Yoshida,J. Momma, M. Hayashi, J. Akiyama, T. Atsumi, K. Chiba, T. Endo, A. Fu-jii, H. Kakishima, H. Kojima, Y. Masamoto, M. Masuda, K. Matsukawa,K. Ohkoshi, J. Okada, K. Sakamoto, K. Takano, T. Suzuki, and A. Takanaka.Alternatives to Animal Testing and Experimentation, 3:123, 1995.[331℄ E. Owen, J. Cli�ord, and A. Marson. J. Cell S i., 32:363, 1978.[332℄ J. F. Sina, G. J. Ward, M. A. Laszek, and P. D. Gautheron. Fundam. Appl.Toxi ol., 18:515, 1992.[333℄ J.-H. S. Kuo, M.-S. Jan, C.-H. Chang, H.-W. Chiu, and C.-T. Li. ColloidsSurf., B, 41:189, 2005.[334℄ S. Mahiuddin, A. Renon ourt, P. Bauduin, D. Touraud, and W. Kunz. Lang-muir, 21:5259, 2005.[335℄ N. Vla hy, M. Dre hsler, J.-M. Verbavatz, D. Touraud, and W. Kunz. J.Colloid Interfa e S i., 319:542, 2008.[336℄ D. P. Cistola, J. A. Hamilton, D. Ja kson, and D. M. Small. Bio hemistry,27:1881, 1988.[337℄ K. Fontell and L. Mandell. Colloid Polym. S i., 271:974, 1993.[338℄ M. Gebi ki, J. M.and Hi ks. Nature, 243:232, 1973.[339℄ C. L. Apel, D. W. Deamer, and M. N. Mautner. Bio himi a et Biophysi aA ta, 1559:1, 2002.[340℄ K. Morigaki and P. Walde. Curr. Opin. Colloid Interfa e S i., 12:75, 2007.158

[341℄ M. Bergsma, M. L. Fielden, and J. B. F. N. Engberts. J. Colloid Interfa eS i., 243:491, 2001.[342℄ M. Johnsson, A. Wagenaar, M. C. A. Stuart, and J. B. F. N. Engberts. Lang-muir, 19:4609, 2003.[343℄ S. Fran es hi, V. Andreu, N. de Viguerie, M. Riviere, A. Lattes, andA. Moisand. New J. Chem., 22:225, 1998.[344℄ Z. Yuan, J. Hao, and H. Ho�mann. J. Colloid Interfa e S i., 302:673, 2006.[345℄ A. de Groot, R. W. and; Wagenaar, A. Sein, and J. B. F. N. Engberts. Re l.Trav. Chim. Pays-Bas, 114:371, 1995.[346℄ A. Renon ourt, P. Bauduin, E. Ni holl, D. Touraud, J.M. Verbavatz,M. Dubois, M. Dre hsler, and W. Kunz. ChemPhysChem, 7:1892, 2006.[347℄ T. Mosmann. J. Immunol. Methods, 65:55, 1983.[348℄ Y. Talmon. Modern Chara terization Methods of Surfa tant Systems, page147. Mar el Dekker: New York, 1999.[349℄ R. Smith and C. Tanford. Pro . Natl. A ad. S i., 70:289, 1973.[350℄ T. H. Haines. Pro . Natl. A ad. S i., 80:160, 1983.[351℄ H.-T. Jung, S. Y. Lee, E. W. Kaler, and J. A. Zasadzinski. Pro . Natl. A ad.S i., 99:15318., 2002.[352℄ T. M. Weiss, T. Narayanan, C. Wolf, M. Gradzielski, P. Panine, S. Finet, andW. I. Helsby. Phys. Rev. Lett., 94:38303, 2005.[353℄ H. Hoe�kes, K. Nelles, and B. Bergmann, 2000. DE 19835327 A1 200002210.[354℄ H. Hoe�kes, M. Seiler, W. Gross, G. Knuebel, D. Oberkobus h, S. Kainz, andA. Rei hert, 2007. DE 102005055340 A1 20070524.[355℄ S. Kainz, H. Hoe�kes, M. Krippahl, and C. Kolonko, 2006. WO 2006097167A1 20060921. 159

[356℄ A. Kleen, M. Akram, S. Hoepfner, and H. Manne k, 2005. WO 2005058260A1 20050630.[357℄ A. Kleen and B. Frauendorf, 2007. EP 1762220 A2 20070314.[358℄ Y. Barenholz. Curr. Opin. Colloid Interfa e S i., 6:66, 2001.[359℄ J. Dong, D. R. Whit omb, A. V. M Cormi k, and H. T. Davis. Langmuir,23:7963, 2007.

160

A knowledgementsFrom experien e I an tell you that these last pages of a PhD thesis are the mostwidely read pages of the entire publi ation. It is here where you supposedly �nd outto what extent you have in�uen ed the life and work of the author. While this maybe true to some extent, you have to weigh my verdi t with the disturbingly low levelof sanity left in this PhD andidate after several years of studying toxi mole ules,rede�ning the olor blue and sear hing for `berries' in aqueous surfa tant solutions.The work des ribed in this thesis has been arried out under the guidan e of Prof.Dr. Werner Kunz at the Institute for Physi al and Theoreti al Chemistry, Universityof Regensburg. I would like to thank him for the trust he showed by letting mework independently and the opportunity to ollaborate with a number of othergroups around the world. Furthermore, thanks to all of his various personalities(the arabi one I am espe ially fond of) for keeping us amused during lun hes andlong onferen e rides.Part of the experimental work was performed at the Department of Pharma eu-ti al Biology, University of Regensburg. Thank you to Prof. Dr. Jörg Heilmann,Dr. Birgit Kraus, and Gabi Brunner for their patien e and trust they showed byletting this near-theoreti ian work in their sterile environment.Spe ial thanks to Dr. Didier Touraud for de orating our lab in pink olor andmaking ammonia the ommonly used air freshener. Furthermore, Didier, thank youfor your en ouragement through the years; for knowing when to push me further andwhen to ba k o� and, espe ially, for always taking the time to dis uss new (mostly razy) ideas.The good basi s were developed already during my undergraduate studies. Forthat, I am thankful to Prof. Dr. Ksenija Kogej, for tea hing me to pay attention161

to detail and work in an organized fashion. An indispensable skill that was provenne essary to survive in the sometimes he ti half-Fren h institute.No s ienti� work is done in solitude. Three mi ros opists are responsible formy newly found appre iation for (abstra t) art: Dr. Jean-Mar Verbavatz (CEA,Sa lay), Dr. Mar us Dre hsler (University of Bayreuth), and Prof. Dr. Ishi Talmon(Te hnion-Israel Institute of Te hnology). Thank you for your wonderful pi turesand informative dis ussions.I thank my oworkers that transformed the `supposed-to-be' serious work en-vironment into a fun pla e (although it sometimes resembled more a loony-bin):Alina for showing me the ropes from the �rst day on, Chloé for her ability to turnba k time and make me feel like a kid whenever I'm in her ompany, Geli for thesmile I was greeted with every morning, Jeremy for his onversations on and o� theOrbiTrek.To my traveling ompanions: Moj a & Neza. After the adventures we had, oming ba k to work seemed like va ation. I am looking forward to all the itiesthat still await us.Thank you to Jorge Cham, the reator of PhD omi s, for helping me �ghtinsanity with irony.To Daniel: Thank you for reading my thesis, listening to my endless omplaining,helping me pi k out my interview suit, wiping the tears away when the reviewers(wrongfully, of ourse!) reje ted my manus ripts and nursing me after my eye oper-ation. You are my endless sour e of motivation and I am lu ky to have you to openmy eyes (despite my newly-found 150% vision) when needed.To my father: thank you for en ouraging me on all my paths. I am still waitingfor that Vla hy & Vla hy arti le however . . .

162

List of Publi ationsPublished:1-N. Vla hy, J. Dolen , B. Jerman, K. Kogej, In�uen e of Stereoregularity of thePolymer Chain on Intera tions with Surfa tants: Binding of Cetylpyridinium Chlo-ride by Isota ti and Ata ti Poly(metha ryli a id). J. Phys. Chem. B 2006 110,9061-9071.2-A. Renon ourt, N. Vla hy, P. Bauduin, M. Dre hsler, D. Touraud, J.-M. Ver-bavatz, M. Dubois, W. Kunz, B. W. Ninham, Spe i� alkali ation e�e ts in thetransition from mi elles to vesi les through salt addition. Langmuir 2007 23, 2376-2381.3-N. Vla hy, D. Touraud, K. Kogej, W. Kunz, Solubilization of metha ryli a idbased polymers by surfa tants in a idi solutions. J. Colloid Interf. S i. 2007 315,445-455.4-N. Vla hy, M. Dre hsler, J.-M. Verbavatz, D. Touraud, W. Kunz, Role of the sur-fa tant headgroup on the ounterion spe i� ity in the mi elle-to-vesi le transitionthrough salt addition. J. Colloid Interf. S i. 2008 319, 542-548.5-N. Vla hy, A. Renon ourt, M. Dre hsler, J.-M. Verbavatz, D. Touraud, W. Kunz,Blastulae aggregates: New intermediate stru tures in the mi elle-to-vesi le transi-tion of atanioni systems. J. Colloid Interf. S i. 2008 320, 360-363.A epted:6-N. Vla hy, A. F. Arteaga, A. Klaus, D. Touraud, M. Dre hsler, W. Kunz, In�u-en e of additives and ation hain length on the kineti stability of supersaturated atanioni systems. Colloids Surf., A 2008.7-N. Vla hy, C. Merle, D. Touraud, J. S hmidt, Y. Talmon, J. Heilmann, W. Kunz,163

Determining the delayed ytotoxi ity of atanioni surfa tant mixtures on HeLa ells. Langmuir 2008.Submitted:8-N. Vla hy, M. Dre hsler, D. Touraud, W. Kunz, Anion spe i� ity in�uen ingmorphology in atanioni surfa tant mixtures with an ex ess of ationi surfa tant.Comptes rendus Chimie A adémie des s ien es 2008.9-N. Vla hy, B. Jagoda-Cwiklik, R. Vá ha, D. Touraud, P. Jungwirth, and W. Kunz,Hofmeister series of headgroups and spe i� intera tion of harged headgroups withions. J. Phys. Chem. B 2008.10-N. Vla hy, D. Touraud, J. Heilmann, W. Kunz, Determining the delayed yto-toxi ity of atanioni surfa tant mixtures on HeLa ells. Colloids Surf., B 2008.

164

List of Oral and Poster PresentationsOral Presentations:09/2007 21th Conferen e of the European Colloid and Interfa e So iety, Genève,Switzerland. Blastula vesi les: Formation of regular patterns through se ondaryself-assembly of atanioni vesi les.Poster Presentations:08/2005 29. International Conferen e of Solution Chemistry, Portoroz, Slove-nia.Mixed Solutions of Isota ti Poly(metha ryli a id) and Cetylpyridinium Chloridein water.08/2006 1st European Chemistry Congress, Budapest, Hungary.Solubilization of Hydrophobi Polymers with Surfa tants.09/2007 21th Conferen e of the European Colloid and Interfa e So iety, Genève,Switzerland.E�e t of �hydrophobi � ations on vesi ular self-assembly.11/2007 Formula V, Potsdam, Germany.In�uen e of additives on the stability of (eute ti ) supersaturated atanioni sys-tems.

165

Herewith I de lare that I have made this existing work single-handed. I have onlyused the stated utilities.Regensburg, June 2008Nina Vla hy

166