jnumal of Colloid and lnterfacc Science III], 8-1".-' [I999]
`Article lDjcis.1993.5'.I'T6_ available online at hup:.=';'mnv.idealibrnrycom unl II E
`
`®
`
`Aggregation Behavior of Tyloxapol, a Nonionic Surfactant
`Oligomer, in Aqueous Solution
`
`.-'.t‘r'aef: and i‘.-’rr.s'!il'rr.' C. Srra"rort l'CNRS).
`‘Department! of C}rt'rrrr't‘rrl' Errgr'rrer‘r’r'rt_r;. Berr—Gm‘r'on Urrit-1=r‘.ril_1'. 8-Hflfi Brershe-\‘a.
`6 me Brnrssingarrlr.
`t‘i?l’lr‘lu"l S::'osbo:rrg. Frame
`
`Oren Regev"' and Raoul zanar-'
`
`Received March 26. 1998; accepted July 29, I993
`
`The aggregation behavior of Tyloxapoi. a nonionic surfactant
`oligomer with a repeating unit close to Triton X-100 (TX10[l}. and
`a maximum degree of polymerization of about 7. has been inves-
`tigated i.n aqueous solution by means of fluorescence probing,
`time-resolved fluorescence quenching (TRFQ) and transmission
`electron microscopy at cryogenic temperature (cryo—TEM}. The
`plot of the pyrene fluorescence intensity ratio l',l'I3 against the
`Tyloirapol concentration shows no clear evidence of a critical
`micelle concentration contrary to TXIDB. Nevertheless. the fitting
`of these data. assuming a partition of pyrene between Tyloxapol
`aggregates and water. yields ctuc values in the micromolar range.
`i.e., about a hundred times lower than for the "monomer" TXIIJD.
`The values of IJI3 at high surfactant concentrations indicate that
`Tyloxapol micelles provide pyrene a less polar environment than
`TX100 micelles. The use of the viscosity-sensitive probe 1.3-
`dipyrenyipropane indicates that the rnicroviscosity of Tyloxapol
`micelles is quite high. three to four times larger than that for
`TXl[l0 micelles. and decreases rapidly with increasing tempera-
`ture. Also the rnicroviscosities of both TXIOIJ and Tyloxapol mi-
`celles are larger than those for the micelles of the nonionic ethoxy—
`lated surfactant CREE: The aggregation numbers of Tyloxapol
`and of TX 100 micelles measured using TRFQ increase with tem-
`perature. with the Tyloxapol micelles being smaller than the
`TXIDO micelles. Cryo-TEM shows that the Tyloxapol micelles
`remain spheroidal up to a concentration of about ll} wt%. At 15
`wt%. some regions of ordered elongated micelles are also observed
`which may be the precursors of the hexagonal phase known to
`occur at about 35 wt%.
`o 1999 Academic hm
`
`Key Words: nonionic surfactant oligomer: Triton X-100: Tylox-
`apol; aggregation in aqueous solution: cmc; micelle shape; micelle
`microviscosity.
`
`INTRODUCTION
`
`In recent years a new class of surfactants which can be broadly
`referred to as surfactant oligomers has attracted increasing interest
`{I-5}. Surfactant oligomers are made up of two or more am-
`phiphilic moieties connected at the level of. or very close to. the
`head groups by a spacer group. The surfactant dimers, also re-
`ferred to as gemini surfactants (2). have been the most investi-
`
`‘ To tvhurn correspondence should be addressed.
`
`l'K'I2I—9'i"l'.n":"}9 530 flfl
`Copyright ii [999 by Academic Press
`All rights of reproduction in any form reserved.
`
`and olig-
`surfactant dimers
`in
`interest
`gated. The
`omers is due to the fact that they are characterized by much lower
`critical micelle concentration (cmc) values and much stronger
`surface tension lowering action then the conesponding conven-
`tional (rnonomeric) surfactants ( 1-5). These two properties deter-
`mine most uses of surtactants in formulations. Surfactant trimers
`
`and tctramers are characterized by even lower cmc values than
`surfactant dimers (6-8). The work on surfactant dimers and
`oligomers has been recently reviewed (3-5). Thus far. the reported
`studies have essentially concerned cationic and anionic surfactant
`dimers. There has been no report on the properties of nonionic
`surfactant oligomcrs. probably because of difficulties in their
`synthesis. However. a nonionic surfactant oligorner. refened to as
`Tyloxapol, has been synthesized and described (9. 10) and can be
`obtained from chemical manufacturers. Tyloxapol is very close to
`being an oligomer of the much investigated Triton X—|00
`{TXlD0; see chemical structures in Scheme 1). The micelle for-
`
`mation by Tyloxapol and the phase diagrams of Tyloxapolfwater
`and Tyloxapol/TXl0Olwater mixtures have been investigated
`(I143). In view of our current interest in surfactant oligomers
`and the availability of Tyloxapol we decided to investigate its
`aggregation behavior in aqueous solution and to compare it to that
`of its monomer, TXIOO. For this purpose we made use of fluo-
`rescence probing using the probes pyrene (cmc. micelle polarity)
`and 1,3-dipyrenylpropane (micelle microviscosity} (i4—l6). We
`also used tirne-resolved fluorescence quenching (micelle agre-
`gmion numbers) (l?—2(l'} and transmission electron microscopy at
`cryogenic temperature (direct imaging of the aggregates in the
`surfactant solutions‘) (21). The results reveal important differences
`between the properties of the micelles of Tyloxapol and TXIOO
`and also between their aggregation behaviors.
`
`EXPERIMENTAL
`
`M’arer'r'als
`
`The sample of TXI00 (Aldrich Chemicals- Europe} was
`used as received. A sample of Tyloxapol {Sigma} was exten-
`sively clialyzed against water using dialysis bags with a cutoff
`of about 1500 Da and the surfactant was recovered by exten-
`sive vacuum rotatory evaporation first under 15 mm mercury
`
`{PD
`j‘‰‚“Ž‚ fyRQSQM q‚ˆ† R
`lnnopharma EX1020, Page 1
`
`

`
`AGGREGATION BEHAVIOR OF TYLOXAPOL TN AQUEOUS SOLUTION
`
`9
`
`
`
`Triton X-100
`
`x= 9-10
`
`SCHEME 1.
`
`Tyloxapoi
`
`J: = 8-10 It < 6
`
`Chemical structures of Triton X-100 and of Tyloxapol {EO= —CH,Cl-110-).
`
`then under 1 mm mercury, at 40°C. The water content of the
`samples of raw and dialyzed Tyloxapol and of TXIOO were
`measured using Karl—Fisher titration and found to be below I
`wr%. The concentrations given below have been corrected for
`that water content. The raw Tyloxapol was examined by size
`exclusion chromatography (solvent: tetrahydrofuran; four col-
`umns filled with beads of polystyrene cross-linked by divinyl-
`benzene, of diameter 10 um and of pore sizes 5, 10, 50, and
`100 rim; detection: UV absorption or index of refraction). The
`chromatogram is shown in Fig. 1. Using the method reported in
`(22), this chrornatogram was decomposed in a sum of four
`components corresponding to the monomer, dimer. trimer. and
`
`(a.u.)
`
`signal
`
`20
`
`22
`
`24
`
`26
`
`28
`
`30
`
`Elution volume (ml)
`FIG. 1. Chromatogram of a 3 M16 solution of raw Tyloxapol in tetrahy-
`d1'ofIJ1'an. The signal (ordinate scale) is proportional to the refractive index of
`the eluted solution. The chromatognun has been analyzed according to the
`method in Ref. (22). Curve E is the experimental chromatogmntt Curves l, 2.
`3. and 4 correspond to the monomer. dimer. trimer. and higher oligorner.
`respectively. The broken line curve is the sum of curves 1 to 4.
`
`higher oligomers with percentages in weight of about 4.4, [0,
`16.9, and 68.7, respectively (Fig. I). The nonionic surfactant
`nonaethyleneglycol monododecyl ether (CIZE9) was purchased
`from Flnka and used as received. The clouding temperatures of
`aqueous solutions of CHE9, TX100, and Tyloxapol at a con-
`centration of 3 wt% were found to be 83 1 0-5, 65.9 t 0.2, and
`90 : 1°C, respectively.
`The fluorescence probes pyrene (purified by zone melting)
`and 1,3-dipyrenylpropane (DPP, fi'om Molecular Probes, used
`without further purification) were the same as in previous
`studies (7). The quenchers of the pyrcne fluorescence used for
`aggregation ntnnber determinations were dimethyIbenzophe-
`none (DMBP) (23) and tetradccylcyanopyridinium chloride
`(TCNPC) (24).
`Water purifed using a Millipore Milli-R0 3Plus (resistance
`>18 int!) was used for the preparation of all solutions. The
`surfactant concentrations are expressed in wt% or in moi/L on
`the basis of molecular weights of 624 g/mo] for TXi00 and 636
`g/mol for the repeating unit ot"l'y1oxapol. These values assume
`an average of 9.5 ethoxy units per molecule or repeating unit
`(11-13).
`
`illerhods
`
`Spectrafluorometry. The pyrene emission spectra were re-
`corded using a Hitachi F4010 spcctrofluorometer and used to
`obtain the ratio I J13 of the intensities of the fir-st and third
`vibronic peaks in the emission spectra of pyrene solubilizcd in
`the micellar solutions. The excitation wavelength was set at
`335 nm. For conventional surfactants the variation of J’.H',
`with the surfactant concentration C permits one to obtain the
`mic. In addition, the value of I,z'I3 at C $ cmc gives a
`measure of the polarity sensed by pyrenc at its micelle solu-
`bilization site (14, 15).
`The emission spectra of DPP were recorded using the same
`spectrofluorometer in the range from 350 and 550 nm, using an
`excitation wavelength of 346 nm. These spectra showed an
`
`j‘‰‚“Ž‚ fyRQSQM q‚ˆ† S
`Innopharma EX1020, Page 2
`
`

`
`10
`
`REGEV AND ZANA
`
`emission due to the monomeric font: of DPP in the range
`360-400 nm with four vibronic peaks (IM, intensity of the first
`peak located near 378 nm) and a broad emission maximum
`arising from the excimer form of DPP, centered at 485-490 nm
`(intensity IE). The product of the intensity ratio IMIIE by
`fluorescence lifetime TE of the DPP excimer is proportional to
`the microviscosity of the surroimding medium (25, 26). The
`intensities [M and IE were measured using air-saturated solu-
`tions. The lifetimes in the same solutions were measured using
`a single photon counting apparatus (23, 24) at an excitation
`wavelength of 335 rim and monitoring the emission using a
`high-pass Kodak gelatin filter 4 (It > 460 nm).
`
`Time-resolved flhrorescence quenching. We attempted to
`measure the aggregation numbers of Tyloxapol and TXl00
`micelles in aqueous solution using the time-resolved fluores-
`cence quenching method (TRFQ) (15, 17-20) with pyrene as
`the fluorescent probe and DMBP or TCNPC as quenchers. The
`fluorescence decay curves were recorded using the same single
`photon counting apparatus which included a Twe 7450 EGG-
`Ortec multichannel analyzer with a time base of 1024 channels
`(23, 24). As usual, the pyrene concentration was adjusted to be
`of about
`1
`to 2% of the mioelle molar concentration, [M],
`whereas the quencher molar concentration, [Q], was adjusted
`to be close to
`(i.e., an average of one quencher per
`micelle). Prior to each experiment, the solutions investigated
`were deaerated by three successive freeze—pump—thaw cycles.
`The following fluorescence decay equations (17-20) were fit-
`ted to the experimental fluorescence decay curves of pyrene in
`the absence and in the presence of the quencher, respectively,
`using a weighted least-squares procedure (23, 24):
`
`!(r) = l’(0)exp(-1.x’-r)
`
`Hr) = !(0)exp{—A2t - Asil ~* exptl - AJ)]}.
`
`[1]
`
`[2]
`
`In these equations {(0) and 1(1) are the fluorescence intensities
`at time 0 and I, r is the pyrene excited state lifetime in the
`absence of the quencher, and A2, A 3, and A4 are three fining
`parameters. The micelle aggregation ntnnber, N, the rate con-
`stant for intramicellar quenching, Ito, and the rate constant for
`probe andlor quencher exchange between micelles, kg, may
`then be determined from
`
`N = Alllc — cmc)»’[Q]}[l + er
`
`to = Am + e]
`
`k, = A, - kg = A.,e.*'(1 + e),
`
`[3]
`
`[41
`
`[5]
`
`where C is the surfactant concentration in mole per (ll'|13 (15,
`11-20), with
`
`E = (A2 ' 'r")!’AsA.:-
`
`[6]
`
`
`
`10"‘
`
`10"
`
`10"
`
`10"
`
`10'
`
`c (awn
`FIG. 2. Variations of I,:’l’, with the surfactant concentration at 10°C (0,
`I), 25°C (El, I), and 40°C (A, L) for Tyloxapol (O, A, El) and TXIOO (O, A,
`I). The solid lines going through the data points are only guides to the eye.
`
`The overall error on the aggregation number is estimated to
`be of 110%.
`
`Cryo-transmission electron microscopy (cryo-TEAO. The
`specimens are prepared by blotting a 511.1 drop of the sample
`on a lacey carbon substrate (300 mesh, Ted Pella Inc-) in a
`controlled environment vitrification system (CEVS) (27),
`where temperature is controlled by a bulb and relative humidity
`is kept above 95% by a wet sponge to prevent sample evapo-
`ration. The specimen is blotted by a filter paper, resulting in the
`formation ofa very thin sample film suspended over the grid,
`which is then vitrified very quickly in liquid ethane maintained
`at the liquid nitrogen temperature, 83 K (— l90.l5°C). The
`CEVS operation retains the original composition of the sample
`so that the original microstructures remain unaltered in the
`vitrified specimen. The vitrified specimen is then transferred by
`means of a Gatan cryoholder into a JEOL l200EXII electron
`microscope operating at 100 H1’
`in the conventional TEM
`mode with nominal underfocus of about 4 pm. The images are
`recorded at 83 K on S0-I63 film. The low-dose mode of the
`
`electron beam is used to reduce radiation damage by the
`electron beam.
`
`RESULTS AND DISCUSSION
`
`The core and Micelle Micropofarity
`
`Figure 2 shows the variations of the 1.0, ratio with the surfbo-
`tant concentration, C, for Tyloxapol and TXl00 at 10, 25, and
`40°C. The 11/I3 vs C plots for the raw and dialyzed samples of
`Tyloxapol were determined at 25°C and found to be coincident.
`This is not surprising in view of the very low fraction of monomer
`(4%) contained in the raw sample and which may be eliminated
`by dialysis. All of the measurements below refer to the raw
`sample. In all instances the intensity ratio shows a sigmoida]
`
`j‘‰‚“Ž‚ fyRQSQM q‚ˆ† T
`Innopharma EX1020, Page 3
`
`

`
`AGGREGATION BEHAVIOR OF TYLOXAPOL TN AQUEOUS SOLUTION
`
`11
`
`decrease upon increasing C, revealing that pyrene goes fiom a
`hydrophilic to a hydrophobic environment. The latter is provided
`by surfactant aggregates when present
`in the solution with a
`sufficient volume fraction to solubilize pyrene. The two sets of
`results show important differences.
`First, for Tyloxapol, the concentration range where most of
`the decrease of IIII3 occurs is about 10-fold lower than that for
`TXl00, showing that this surfactant oligomer provides pyrene
`with hydrophobic microdomains at a much lower concentration
`than its monomer. The crnc range of TXl00 is seen to be
`around 0.01 Wt‘/o, i.e., 0.15 mM. The crnc of TXl00 can be
`obtained using two approaches.
`
`is taken as the concentration corresponding to the
`It
`1.
`intercept of the extrapolations of the high concentration part
`and of the rapidly varying part of the IIII3 vs C plot (28). The
`crnc values are then found to be 0.012, 0.024, and 0.024 wt%
`which is about 0.19. 0.38, and 0.38 mM at 10, 25, and 40°C,
`respectively.
`2. For several ionic and nonionic surfactants with the crnc in
`
`the 0.1 mM range the comparison of the electrical conductance
`or surface tension vs C plots and of the M13 ratio vs C plots
`showed that the crnc, obtained for the conductance or surface
`tension plots (true crnc), closely corresponds to the concentra-
`tion at the infiection point of the 1, {[3 ratio vs C plot (29-31).
`This concentration is also very close to that at mid-decrease of
`the fill, ratio. Using this procedure, the TXl00 crnc values
`were found to be 0.05, 0.10, and 0.12 wt% (that is, 0.08, 0.16,
`and 0.19 mM) at 10, 25, and 40°C. respectively.
`
`The crnc of TX I 00 has been reported to be between 0.25 and
`0.2? mM at 25°C (32-34), a value between the ones given
`above at this temperature.
`The second difference between Tyloxapol and TXI00 is that
`the decrease of [1313 is much less steep for the former, stretch-
`ing over nearly two decades of concentration for Tyloxapol
`against about one decade for TXl00. In addition, there is no
`really sharp change at the end of the decrease for Tyloxapol,
`contrary to TXIOO. This complicates the determination of the
`“crnc” for Tyloxapol. Indeed. pyrene reports a partly hydro-
`phobic environment already at a concentration of, say, 3 )4
`l0_‘' wt% (i.e., 5 X 10_° M). However, the cute, if it exists,
`may be located at an even lower concentration. Indeed at such
`a low amphiphile concentration the volume of the hydrophobic
`pseudo-phase is very small and a significant fraction of pyrene
`remains in the water, thereby reporting a partly aqueous envi-
`ronment even at concentrations above the crnc- The 1,113 plots
`for Tyloxapol look very much like those characterizing sys-
`tems with very low or zero crnc, where pyrene is progressively
`partitioned into hydrophobic aggregates, as the concentration
`of the amphiphile is increased (35).
`In such systems.
`the
`experimental fluorescence intensity ratio can be written in the
`form (35)
`
`Ii”: = (Il;‘r3)Ag ‘l’ lulflaiw ‘ (fr/Islsg]/[1 ‘l’ K(C' 91310)]:
`
`[7]
`
`I!
`
`0.4
`
`0.8
`
`0.6
`
`0.2
`
`10"
`
`C {%wt)
`
`FIG. 3. Variations of the normalized fluorescence intensity ratio R with
`the surfactant concentration for Tyloxapol at 10°C (0), 25°C (El), and 40°C
`(A). The curves in solid lines 1, 2, and 3 have been calculated using Eq. [8]
`with the values K = 1434, 934, and 480 (wt%)“, and mm = 1.1 X 10"‘,
`1.4 X 10"‘, and 0.7 X 10“ wt%, at 10. 25 and 40°C, respectively.
`
`where (1.r'I3)w and (1J13)A3 are the values of the intensity
`ratio in water and in the aggregate (that is, at C 3- cute), and
`K is an apparent binding constant of pyrene to the aggregates
`(Kf55.5 is in fact the partition coefficient of pyrene between
`the hydrophobic domains considered as a pseudo-phase and
`water). The data in Fig. 2 have been normalized and replotted
`in Fig. 3 as R against C using a serni-logarithmic scale for the
`concentration, with R given by
`
`R = ill”: “ UIH3lag]il(1IH3)w ‘ (-{lf"3.)Ag:i
`
`= 1."[l + K(C — crncl].
`
`[8}
`
`The quantity R varies between 1 and 0 upon increasing C. It
`represents the fraction of pyrene solubilized in the aggregates.
`its variation with log C is an S-shaped curve and R = 0.5 for
`C - crnc = UK. The R vs log C plots permit an easier
`comparison between surfactants of differing (I 1:’!3) Ag and also
`an easy visualization of the quality of the fit of Eq. [8] to the
`data. The plots 1-3 have been calculated using Eq. [8] with
`appropriate values of the binding constant and of the crnc given
`in the legend to Fig. 3. These plots are seen to provide a good
`fit for the Tyloxapol results at 25 and 40°C. The fit is somewhat
`less satisfactory at 10°C. The differences which may exist
`between the calculated curves and the data have been discussed
`
`and attributed to the large change of pyrene concentration in
`the aggregates at concentrations close to the crnc (35). The cure
`values thus obtained are very inaccurate, around 10-4 wt%,
`which is 1.6 ,uM, at the three temperatures investigated. This
`value is very low, much lower than that for TXIOO. However,
`our results do not pennit one to conclude whether the aggre-
`gates capable of binding pyrene in Tyloxapol solutions result
`from an aggregation of Tyloxapol around the pyrene mole-
`
`j‘‰‚“Ž‚ fyRQSQM q‚ˆ† U
`Innopharma EX1020, Page 4
`
`

`
`12
`
`REGEV AND ZANA
`
`~
`I
`E
`
`18.4 ma
`
`15.6
`
`152
`
`1-.
`3
`.1"
`E
`
`3.1
`
`3.2
`
`3.3
`
`3.4
`
`3.5
`
`1o’rr
`
`FIG. 5. Variations of the micelle microviscosity (solid lines) for Tylox-
`apol (El), TXIOO (O). and CHE, (£5), and corresponding variations of the
`lifetime of the DPP excimer (broken lines) for Tyloxapol (I). TXJOO (I), and
`C.2E., (A), with temperature.
`
`because of the extremely low values of the excimer intensity
`which reflect extremely large values of the microviscosity. The
`‘TEIMHE values for Tyloxapol solutions below 40°C must be
`considered lower bound values.
`
`Figures 4 and 5 show the variations of 1-EIM/IE with con-
`centration and temperature (T), respectively, for Tyloxapol
`and TX100. In all instances the microviscosity of Tyloxapol
`aggregates is larger than that of TXIOO. A similar difference
`has been found between the microviscosity values of ionic
`surfactant oligomets and their corresponding monomers (7).
`Figure 4 shows that the microviscosity is nearly independent of
`the TXl00 concentration but decreases somewhat as the T3110):-
`apol concentration is increased- Figure 5 shows that the uni-
`croviscosity of both Tyloxapol and TX100 aggregates de-
`creases steeply as the temperature increases. In the semi-
`logarithmic representation adopted in Fig. 5 (Int:-EIMIIE)
`against III‘) the plots are linear; that is, one can write
`
`TEIMHE = ("'£1'M”Elo°XP(‘E:NfTl-
`
`[9]
`
`The slopes of the plots in Fig. 4 yield the values 61.2 and 62.2
`Id/mol for the activation energy E: of the microviscosity for
`TXIOO and Tyloxapol micelles, respectively. Also represented
`in Fig. 5 are the semi-logaritlunic variations of the lifetime T5
`of the DPP excimer with UT. The plots are linear and yield the
`values 16.8 and 1 I .8 lclfrnol for the activation energy E of T5
`for TX100 and Tyloxapol.
`For the sake of comparison we have also represented in Fig.
`5 the variations of the values of 1'5 and of the micellar micro-
`viseosity for the nonionic ethoxylated surfactant CHE9. It is
`seen that the microviscosity of CHE; micelles is lower than
`that for TXIOU and Tyloxapol micelles in the entire range of
`temperatures investigated. The activation energies for ‘TB and
`
`j‘‰‚“Ž‚ fyRQSQM q‚ˆ† V
`Innopharma EX1020, Page 5
`
`cules, or reflect the solubilization or binding ofpyrene to single
`Tyloxapol molecules which would be able to form intramolec-
`ular micelles, or arise from a true intennolecular association of
`Tyloxapol molecules as for normal amphiphiles at above the
`cute. In the latter case our cmc value would be much smaller
`
`than that advanced by Westesen and Koch (12). Note that ionic
`surfactant oligomers have consistently been found to have
`much lower cmc values than the corresponding monomers. A
`similar behavior is expected for Tyloxapol with respect to
`TX] 00-
`
`The third difference between TXIOO and Tyloxapol is that at
`C > cine, as the temperature is decreased from 40 to 10°C, the
`value of I 1113 increases from about 1.32 to 1.44 for TX] 00, but
`remains nearly constant at about 1.32 for Tyloxapol. At 25°C
`the values of)’,/I3 are 1.32 and 1.4 for Tyloxapol and TXIOO,
`respectively, indicating that the micelles of Tyloxapol provide
`pyrene with an environment of lower polarity than TXIOO
`micelles (I4, 15). However, at 40°C, Tyloxapol and TXIOO
`micelles are characterized by about the same value of I 1:’!3.
`
`Microvismshy
`
`The values of the DPP excirner lifetime and of the ratio
`
`IMIIE were determined for Tyloxapol and TXl00 solutions at
`different concentrations and temperatures. We did not attempt
`to calculate absolute microviscosities from the results. Indeed
`
`this involves the use of a calibration plot (plot of TEIMIIE vs
`viscosity for a series of liquids of known viscosity) and the
`microviscosities thus obtained depend on the calibration curve
`used (36). Since we were essentially interested in comparing
`the micellar microviscosities of Tyloxapol and of its monomer
`TX100, all the results below concern the values of the product
`TEIMIIE which is referred to as microviscosity even though it is
`realized that it is only proportional to the rnicroviscosity. At the
`outset it should be stated that the measurements of {M/IE for
`Tyloxapol solutions at temperatures below 40°C were diflicult
`
` 0
`
`‘III
`
`2|!
`
`3|}
`
`6 (%wt)
`FIG. 4. Variations of the micelle microviscosity with the surfactant con-
`centration for Tyloxapol (O) and TXIOD (I) at 45°C.
`
`

`
`AGGREGATION BEHAVIOR OF TYLOXAPOL TN AQUEOUS SOLUTION
`
`13
`
`TABLE I
`
`Results of TRFQ Studies of Tyloxapol and TX100 Solutions
`
`Surfactant
`
`C (w't%)
`
`T(°C)
`
`Quencher
`
`-r(ns)"
`
`A;‘ (as)
`
`A,
`
`10"‘ .4, (ns)
`
`TXIUD
`TXI00
`Txtoo
`Tyloxapol
`Tyloxapol
`Tyloxapol
`Tyloxapold
`CIZEQ
`c,,E.,
`
`3.436
`3.436
`3.436
`3.418
`3.456
`3.418
`3.35
`3.2
`3.2
`
`25
`25
`40
`25
`40
`40
`55
`25
`40
`
`0.463 mM TCNPC
`0.528 mM DMBP
`0.523 mM DMBP
`0.570 M TPCNC
`0.590 mM nuns
`0.530 mM TPCNC
`1.236 mM TPCNC
`No quencher
`‘No quencher
`
`343.2 [24l.l]
`343.2 [241.1]
`329.2 [i91.t)"]
`349.5 [2s1.o]
`323.1 [225.4*]
`328.1 [225.4"]
`[£92.71
`345.6 [l8l.-1]
`321.3 [1353]
`
`317.1
`310.2
`268.4
`
`285.6
`292.1
`lS2.8
`
`0.750
`0.850
`1. 176
`
`0.427
`0.420
`1 .63
`
`4.13
`4.78
`4.57
`
`3.83
`420
`4.22
`
`N’
`
`106
`I04
`155
`
`63.8
`58.3
`99
`
`to" to {s‘ ')
`
`10"’ it, (s")
`
`3.95
`4.40
`4.07
`
`3.04
`3.46
`3.52
`
`2.3
`3.8
`5.0
`
`8.4
`7.4
`7.2
`
`" Values in brackets are for aerated solutions. The other values are for deaerated solutions.
`" Values at 41°C.
`" Values representing the number of surfactant hydrophobic moieties per rnicelle.
`“' This measurement has been performed on an air-saturated solution.
`
`for the microviscosity for C DE, are I17 and 49.8 ldlmol,
`respectively.
`Note that at a given temperature, the 1-,; values increase in
`the order CIZE9 < TX 100 < Tyloxapol, that is, in the order of
`increasing microviscosity,
`in the temperature range investi-
`gated. The values of the pyrene fluorescence lifetimes in aer-
`ated micellar solutions of the three surfactants listed in Table 1
`
`also increase in the order C DE‘, < TXIDO < Tyloxapo].
`However, Table I shows that identical values are found for the
`pyrene lifetime in deaerated micellar solutions,
`i.e.,
`in the
`absence of solubilized oxygen, of the three surfactants- This
`fact confirms that the differences in DPP excirner and pyrene
`lifetimes found for the aerated solutions of the three surfactants
`
`are due to a loss of efficiency in the quenching of the probe
`fluorescence by molecular oxygen, which increases with the
`micelle microviscosity, as previously discussed (7).
`The larger microviscosity of TX l 00 micelles with respect to
`micelles of ethoxylated nonionic surfactants C,,,E,, (E =
`ethoxy group) explains why abnormally low values of the
`micelle aggregation number were measured in a previous time-
`resolved fluorescence quenching study of TXIOO solutions
`(37). This study used the formation of excimer by pyrene (15).
`A later study (38) showed that the rate constant for inu'arr1i-
`cellar pyrene excimer formation is generally lower than the rate
`constant for pyrenc quenching by alkylpyridinium ions. This
`fact together with the high microviscosity of TXIOO resulted in
`decay plots where the linear part was not well developed. The
`fitting of Eq. [2] to the decay data yielded incorrect values of
`the fitting parameters and suggested a very rapid intermiccllar
`exchange of pyrene. In turn, the analysis of the fitting param-
`eters by means of Eqs.
`[2]—{S] yielded N values much lower
`than those obtained when using light scattering, for instance,
`particularly at low temperatures where the microviscosity of
`TX100 micelles is high.
`The microviscosity of the mixtures of TXIOO and Tyloxapoi
`and the DPP excinter lifetime in these mixtures have been
`
`measured at 45°C. The results are represented in Fig. 6 against
`the Tyloxapol weight fraction. The continuous increases of TE
`and -r,._1M/IF between pure TXIOO and Tyloxapol indicates the
`formation of mixed aggregates. This conclusion is supported
`by the results of Westesen er al. (13).
`
`Aggregation Numbers
`
`As discussed in the preceding paragraph the high tnicrovis-
`cosity of TX100 and Tyloxapol micelles makes it necessary to
`use very eificient quenchers of pyrene fluorescence in order to
`make reliable estimates of the micelle aggregation numbers of
`these two surfactants. In a first attempt we used the tetrade-
`cylpyridiniunt ion as quencher. With both TX100 and Tylen-
`apol the decay plots were nearly linear and did not permit any
`meaningful determination of aggregation numbers. We there-
`
`1600
`
`‘I400
`
`1200
`
`% 1000
`E:
`son
`000
`
`400
`
`200
`
`0.0
`
`0.2
`
`0.4
`
`0.8
`
`0.3
`
`140
`
`120
`
`E‘
`10!! :3
`
`E0
`
`.
`1.0
`
`Tyloxapol
`weight fraction
`‘DH 00
`FIG. 6. Variations of the mioelle microviseosity (O) and of the DPP
`excirner lifetime (0) for the TXl00-Tyloxapol n:1ixtures at 45°C, as a function
`of the Tyloxapol weight fraction in the mixture. Total surfactant concentration:
`EbO'l.Il 5 WW9.
`
`j‘‰‚“Ž‚ fyRQSQM q‚ˆ† W
`Innopharma EX1020, Page 6
`
`

`
`14
`
`REGEV AND ZANA
`
`1"
`
`"
`
`.'
`
`'
`
`-'.
`
`—'
`
`'
`
`-'
`
`.
`
`'.i.;“
`
`FIG. 7. Cryo-TEM micrograph of a 0.1 wt% Tyloxapol sample quenched from 25°C showing spheroidal micelles 5 nm in cliameter. Dark objects are frost
`panticies. Bar, 200 run.
`
`fore turned to the dirnethylbenzophenone (23) and the tetrade-
`cylcyanopyridinium ion (24) which are very efficient quench-
`ers of pyrene fluorescence. The resuits are summarized in
`Table I. The surfactant concentration used was fairly high in
`order to ensure that the probe was nearly completely parti-
`tioned in the micelles.
`
`The N values found for TXIOO with the two quenchers at
`25°C are identical within the experimental error. Also N in-
`creases much with temperature between 25 and 40°C, as for the
`nonionic surfactants C,,,E,, (3?—40). Our N values are in very
`good aggreement with those reported by Brown et at‘. (41), who
`used TRFQ with the pair pyrene—benzophenonc. They also
`agree with the value reported in a recent dynamic light scat-
`tering study (42).
`At 25°C the decay plots for Tyloxapol with the DMBP and
`TCNPC quenchers were close to linear and did not permit
`reliable estimates of aggregation number values. Recall that at
`this temperature the microviscosity of Tyloxapol micelles is
`still quite high and much larger than that of TXIOO. At 40 and
`55°C, however, the decays were sufficiently curved to permit
`reliable estimates of N. The N values thus obtained (see Table
`1) Show that N increases with temprature as for TX100. How-
`ever, the values for Tyloxapol are lower than those for TX 1 00,
`indicating smaller micelles. This result is at variance with that
`found for ionic surfactant oligoiners for which the micelle
`
`aggregation number, expressed as the number of alkyl chains
`per micelle, increased in going from the monomer to the dimer
`and trirner at a constant weight percent of surfactant (6, 43).
`However, Tyloxapol is an oligomer of larger oligomerization
`degree (up to seven) than those used so far. Also, as shown by
`the chromatogram in Fig. l, the sample used is polydisperse.
`Mixed micellization most likely takes place between the vari-
`ous oligomers and may result in spherical micelles. Recall that
`the mixture of dodecyltrirnethylarnmonium bromide (DTAB)
`which forms spherical micelles and of its dimer (denoted
`12-2-12) which forms giant thread~like micelles yields sphe-
`roidal micelles at a relatively low mole fraction of DTAB (44).
`The diameter of the hydrated Tyloxapol micelles has been
`calculated to be 6.3 nm, using an average value of 60 units for
`the aggregation number at 40°C, assuming a density of l for
`the aggregates (12), and an average hydration number of four
`water molecules per ethoxy group (45-49). A value of 7 nm
`was previously obtained (12) on the basis of distribution func-
`tion analysis (50) of small angle X-ray scattering data for
`Tyloxapol micelles at a concentration of 3 wt%.
`It is interesting to note that the values of the quenching rate
`
`constant, kc, for the Tyloxapol and TXIOO micelles are fairly
`close at 40°C (see Table 1). This is so because k9 is inversely
`proportional to the product of the microviscosity by the aggre-
`gation number (51-53). In the present case the higher micellar
`
`j‘‰‚“Ž‚ fyRQSQM q‚ˆ† X
`Innopharma EX1020, Page 7
`
`

`
`AGGREGATION BEHAVIOR OF TYLOXAPOL TN AQUEOUS SOLUTION
`
`15
`
`-
`-
`' 4"
`i‘
`_
`.
`.
`.
`
`FIG. 8. Cryo-TEM micrograph of a 15 w't% Tyloxapol sample quenched from 25°C. Dark objects are frost particles. Bar. 200 um.
`
`_
`'-$1-'-I
`
`s_..
`
`.
`
`.
`
`f.
`
`".-
`
`_‘_I‘
`
`-Q.
`
`'
`
`8-
`
`r ,_.
`
`micruviscosity of Tyloxapol with respect to TXIOO is largely
`compensated by the smaller rnicelle aggregation number of
`Tyioxapol with respect to TX100.
`The values of the exchange rate constant are fairly tour for
`both surfactants at 40°C. Exchange probably takes place
`through micellc collisions with temporary merging, as in mi-
`cellar solutions of CME,, surfactants (37-40).
`
`35 wt% Tyluxapol.“ The d spaL:i.ng calculated in this work
`
`(Fig. 2 of Ref. 12), 6 nm, is similar to the d spacing measured
`from our cryo-TEM image (Fig. 8).
`Below 0.] wt% we are unable to observe any form of
`aggregation in this system, most probably due to the low
`concentration of aggregates andfor their low aggregation num-
`her.
`
`Microsn-ucrure of the Solutions (Cr;va»1EM_)
`
`Cryo—TEM images of Tyloxapol at 0.1 wt% show spheroidal
`micelles 5 nm in diameter (Fig. 7). The diameter obtained
`confirms the preceding TRFQ results and also the small angle
`X-ray scattering (SAXS) results of Westesen and Koch (12).
`Note that in the cryo~TEM micrograph the micelles crowd in
`the thicker part of the vitrified film, close to the polymer
`support (21, '54, 55). Up to 10 wt% surfactant the micellar
`shape and size do not change much (not shown). However, at
`15 wt%, still in the micellar phase, some regions of ordered
`elongated micelles are observed (Fig. 8). It
`is tempting to
`believe that these regions are precursors of the hexagonal
`phase, appearing only at 35 wt% (1 1, I2). Indeed, this result is
`in line with Westesen findings from the evolution of peaks in
`SAXS spectra with increasing surfactant concentration (12)
`mentioned as “strong interparticle interference between 10 to
`
`For the sake of comparison and also because the microstruc-
`ture of solutions of TX100 has not yet been investigated by
`cryo-TEM, we h