PEG400 novel phase description in water

Nawal Derkaoui a, Sylvère Said b, Yves Grohens b, René Olier a, Mireille Privat a,∗

aUMR 6521, Chimie, Electrochimie Moléculaires et Chimie Analytique, Département de Chimie, Université de Bretagne Occidentale, 6 Avenue Le Gorgeu,
C.S. 93837, 29238 Brest cedex 3, France
bLaboratoire Polymères, Propriétés aux Interfaces et Composites, Université de Bretagne Sud, Centre de Recherche, Rue Saint Maudé,
BP 92116, 56321 Lorient cedex, France
Received 13 July 2006; accepted 2 October 2006
Available online 7 November 2006

Abstract
The behavior of hydroxyl-terminated PEG400 in water was investigated by surface tension measurements and 13C NMR as a function of concentration and temperature. PEG400 exhibited a critical aggregative concentration (cac) that evidenced both its amphiphilic character and its aggregation capacity. Moreover, the chemical shifts of the different carbons of the PEG were followed by NMR versus concentration at various temperatures. We observed a plateau between 20 and 35 ◦C at concentrations above 0.2 mol L-1 and ascribed it to the aggregation process. A good correlation was found between the NMR spectra in the region of aggregation and the cac region in the phase diagram. Our investigations were also focused on the solid–liquid region of the phase diagram at lower temperatures. These experimental data, together with conclusions available in the literature, led us to propose explanations for the conformation/hydration/aggregation in the PEG400–water solutions phenomena.
© 2006 Elsevier Inc. All rights reserved.
Keywords: PEG; Aqueous solutions; Aggregation; Helical structure

1.Introduction

Assembly properties of polyethylene glycol (PEG) and wa- ter mixtures are widely used (i) to induce more or less marked gelation processes, (ii) to organize surfaces for elaboration of composite materials, and even (iii) as agents promoting crys- tallization. However, despite the abundant literature about these mixtures, comprehensive views of such systems are very scarce.
This study was undertaken to get, first, additional knowledge about the bulk behavior of PEG400. The polymeric chain of any PEG is alternately hydrophilic and hydrophobic. It exhibits spe- cific interactions with water leading, for example, to a helical structure that appears under certain conditions of temperature and concentration. Interactions governing bulk phase behav- ior similarly act in surfaces and in the colloidal states used for practical purposes. This work is the first part of a work aiming at understanding surface phase diagrams of a model system of polyethylene glycol (400 g mol-1), PEG400, and fumed silica.

* Corresponding author. Fax: +33 2 98 01 70 01.
E-mail address: [email protected] (M. Privat).

0021-9797/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2006.10.008
A comprehensive picture of the structure of PEG400 in water is a prerequisite step. However, despite the proliferation of exper- imental and theoretical studies carried out both at macroscopic and molecular levels, the physicochemical behavior of poly- ethylene glycols in aqueous solutions is still very puzzling [1].
Structural evolutions, modifications of intramolecular or in- termolecular interactions (such as hydration), and subsequent changes in macroscopic properties have been evidenced by IR [1,3,4] and NMR [1,2,5,6] investigations and ultrasonic and photocorrelation spectroscopy experiments [1]. The factors leading to these evolutions are mainly polymer concentration [3–6] and temperature [1,2]. The coupling of molecular dy- namic simulations [7] with spectroscopic data has produced more accurate images of PEG intramolecular configuration and of the resulting intermolecular arrangements as a function of temperature and concentration in water.
Among about 10 potential [8] interaction sites per monomer, 2 or 3 are considered as being active in the hydration process and are involved in the formation of the first layer of bonded wa- ter molecules. H-bond complexes between water and PEG have to adapt to packing conditions with the nearest-neighbor repeat units [9,11], conditions that depend on the PEG conformation.

Thus, the flexibility of the PEG molecule allows many con- formers with the possibility of forming locally ordered helical structures with ttg sequences [7]. Since the chemical structure of PEG includes hydrophobic ethylene units and hydrophilic oxygen, the three-dimensional arrangement of the repeat units is of particular interest. However, despite evidence of multi- stranded helices from atomic force microscopy on graphite [7], helix aggregation in water is still quite unclear.
The clustering effect for low-molecular-weight PEG is still controversial, since some authors have claimed that it is caused by impurities [8], whereas others have attributed it to a chain end effect [12]. This topic is still an open question, which is especially crucial for a better understanding of PEG surface be- havior. Numerous studies have been undertaken throughout the world to figure out the phase diagrams of PEG–water systems and to achieve good predictions through relevant models. Good agreement between literature data seems to be obtained about the lower critical solution temperature (LCST) and the closed- loop high-temperature two-phased region, not in concern here, since it exists only with high-molecular-weight components. On the other hand, within this temperature range (100–120 ◦ C) and for low-molecular-weight PEGs, the aggregation phenom- ena are still unclear [7].
To gain more insight into the changes in conformation and aggregation of a low-molecular-weight polymer chain in solu- tion in water, we undertook an extended and complementary study of the physicochemical properties of PEG400 solutions versus temperature and concentration.
We thus performed tensiometry experiments with liquid mixtures close to room temperature. A careful determination of the variations in critical aggregation concentration (cac) with temperature produced part of the general bulk phase di- agram between 20 and 50 ◦ C. This was further supported by a 13C NMR study carried out on different concentrations of the same solutions and varying temperatures over the same range; through the micellization diagram, it showed a change in the evolution of the chemical shift of some carbons in the micelle zone. Besides, from calorimetric investigations at tem- peratures below 0 ◦ C, the classical V-shaped solid–liquid equi- libria appeared from the melting point determinations. These data clarified the cooperative behavior of PEG400 and were in good agreement with previous observations made by other au- thors [10].

2.Experimental

2.1.Chemicals

PEG400 was supplied by Aldrich and said to be such that Mn , Mw = 400, so the polydispersity was equal to 1. Water was
purified on a MilliQ device from Millipore.

2.2.Methods

Critical aggregation concentrations were determined at dif- ferent temperatures by tensiometry using a LAUDA TC1 ten- siometer with a frosted platinum plate previously heated at a

high temperature to remove all the surface impurities. To let the system reach temperature and adsorption equilibria, the mea- surement was started 20 min after the solution had been poured into the cell. The rate of plate immersion in the solution was 0.5 mm min-1. As the platinum surface was frosted, the plate was totally wetted by the solution, and the force exerted by the surface tension was measured as soon as the contact was real- ized, because the force stabilized very quickly before buoyancy became perceptible. To check for the lack of any adsorption ki- netics, the measured value was controlled, first, 20 min after the first measurement, and then twice. In doing so, surface ten-
sions were known at ±0.35 mNm-1 in the most concentrated studied solutions, where the high viscosity made the measure-
ments difficult, and at ±0.1 mNm-1 in dilute solutions. Crit- ical aggregation concentrations were known at ±0.04 molL-1 (maximum error).
The solid–liquid demixing was determined by DSC with a Mettler DSC-820 device. The thermal scans were started at room temperature with a 50 K min-1 decrease of temperature down to -100 ◦ C. The second scan was started at -100 ◦ C at
a heating rate of 10 K min-1 up to 90 ◦ C, and the melting point of the mixture was determined at ±1 K from the thermogram.
All NMR spectra were acquired on a Bruker DRX Avance 500 spectrometer, with a 5-mm inverse triple-resonance high- resolution probe (Bruker TBI 1H {BB}13C). To lock the spec- trometer, three D2O drops were added to each sample prior to experiment. Typical conditions for 13C spectra acquisition were a 30◦ excitation pulse, a 0.52-s acquisition time, a 2.0-s relax- ation delay, and 64-K data points. The data were processed with a minimum of 60 transients and then apodized with a line-broadening factor of 1.0 Hz prior to Fourier transform and phase correction. The spectral window was 31–446 Hz.
The observed bands were assigned from the relative amounts of the different nuclei and HMQC and HMBC sequences on a PEG400 sample in CDCl3.
The HMQC and HMBC data sets consisted of 1024(F2) 128(F1) complex data points, averaged on 48 scans with× 17,606- and 964-Hz spectral windows in F1 and F2, respec- tively. The chemical shifts were measured by reference to the external standard TSP (2,2,3,3-tetradeutero-3-(trimethylsilyl)- propionic acid sodium salt), Uvasol quality from Merck; the Si(CH3)3 resonance frequencies were set to 0.0000 ppm. Tem- peratures were calibrated with a reference sample (80% ethyl- ene glycol in dimethyl sulfoxide—DMSO) and monitored with a BVT 3003 Bruker device.

3.Results and discussion

3.1.Surface tensions

Fig. 1 shows surface tension curves at 20, 25, 35, and 40 ◦ C. At the highest PEG concentrations all of them exhibit a marked minimum, only shown, in the figure, at 20 and 40 ◦ C, because of the choice of abscissa scale. These minima are very difficult to experimentally determine and correspond, in fact, to a very viscous liquid observed in the mixtures. It is worth mentioning that, despite the lack of rheological investigation in this study,

(a) (b)

(c) (d)
Fig. 1. Variations of surface tensions of PEG400–water mixtures at (a) 20, (b) 25, (c) 35, and (d) 40 ◦ C.

it was proven elsewhere [9] to be a gel-like state. The above results at high concentration will not be further discussed, but the transition to gel state will be noted on the final phase di- agram. It is also worth specifying that obtaining the data has been, at any place, a difficult task and that it is necessary to ex- plore a wide concentration zone lead to get a great number of points. The number of points on every plateau corresponding to the beginning of aggregation may seem small, but a sort of self-consistency of the set of these points at different tempera- tures made it possible to build a rather satisfactory “aggregation curve” later shown in Fig. 2. These data could be also perti- nently interpreted using a model of surface tension such as the one of Ref. [13]. However, this paper being dedicated to bulk study, this has been put back.
The slopes at the start of curves can be used to calculate an approximate value of adsorption at the liquid/vapor interface using the so-called Gibbs isotherm formula,
where ΓPEG,w is the relative adsorption of PEG with respect to water, γ is the surface tension, γc is the activity coefficient rel- ative to molarity in an unsymmetrical system of reference, and cPEG is the molarity of PEG. In the following, it has been as- sumed that γc did not vary much in the zone where calculations were made, which leads to the formula

ΓPEG,w = -(1/RT )(dγ/dln cPEG)T. (1′) As the order of magnitude of the aggregation concentrations does not vary much with temperature (∼0.1 molL-1 ), the er- ror possibly generated in this way is also of the same order of magnitude for each determination.
Then, when relation (1) is applied to the curves obtained at different temperatures, just before the surface tension becomes constant, it leads to 7 × 10-7 < ΓPEG,w < 8.9 × 10-7 molm-2 , which means that it is statistically temperature-independent, and, as a consequence, to a mean molecular surface area of

ΓPEG,w = -(1/RT )(dγ/dln γccPEG)T, (1) 2 nm2.

Fig. 2. Variation of the critical aggregation concentration versus temperature. The dashed lines indicate the concentrations at which NMR spectra were
recorded: (a) c = 9.75 × 10-2 , (b) c = 0.257, and (c) c = 0.295 molL-1 . 13 C spectra were recorded at these concentrations as well as at (d) c = 1.34 and (e) c = 1.99 molL-1 (not shown). The vertical lines show the interval of temperatures where the chemical shift of carbon 3 remains constant at the considered concentration (see below). The grey schemes indicate the possible PEG400 configurations in the different zones in accordance with the discussion developed in Sections 3.2.2 and 3.2.4.

A plateau in the surface tension is observed at intermediate concentrations. It was also observed at 45 and 50 ◦ C, not shown. Such shapes in surface tension curves are commonly explained by the formation, in bulk, of aggregates, which takes place af- ter the water/air interface saturation by PEG. The critical ag- gregative concentration (cac) is the most dilute concentration at which aggregates appear; it thus corresponds to the start of the plateau. The graph presented in Fig. 2 was drawn by linking cac values versus temperature; the resulting cac lines delimit a zone characterized by a low concentration and a temperature higher or lower than 35 ◦ C, where PEG molecules are unable to aggre- gate; on the other hand, the triangle-shaped zone corresponds to the area where aggregates exist. This behavior is very striking because a line drawn at constant concentration, on the graph, goes through two different aggregation temperatures. The lower and upper zones both correspond to the formation of aggre- gates, but on condition to respectively elevate or lower tem- peratures. A temperature of 35 ◦ C seems to be a critical value at which aggregation points are superimposed; moreover, the corresponding concentration is the lower one at which aggrega- tion can be observed. A concentration lower than 0.025 mol L-1 may correspond to total solubility of the PEG400 in water what- ever the temperature.
Then, it is tempting to attribute the different zones of Fig. 2 to different solvation states of the polymer entities in relation to temperature and concentration conditions, i.e., clusters in bad solvents and extended coils in good solvents. The former, at low concentrations, would favor a non-aggregative behavior, whereas the latter, at higher concentration and around 35 ◦ C, would favor the formation of aggregates. However, this analysis has to be modified on taking into account the specific properties of PEGs.

Simply considering the alternate hydrophobic ethylene groups and hydrophilic oxygen bridges present all along the PEG chain, it is not sensible to imagine that this chain folds up on itself to flee water contact. In any case, contact with water may be possible in a more or less subtle way.
According to Faraone et al. [1], in the temperature–concen- tration (T –C ) plane, the solvation properties of water drasti- cally change for PEG 600 with respect to the T –C coordinates, as probed by the NMR-measured diffusion coefficients. Thus, at low concentrations and at any temperature, water is claimed to be a good solvent for PEG; i.e., more water molecules are linked to PEG molecules. However, by elevating the concen- tration, expressed in volume fraction, up to a value between 0.2 and 0.35, one can observe, according to the same literature, a zone of transition solvent for PEG at 20 and 40 ◦ C (the theta solvent at 97 ◦ C, which is defined as the temperature for which the solvent–solvent interactions equal the solvent–monomer in- teraction, is far above the temperature region we are dealing with). Finally, above volume fraction 0.4, the measurements in- dicate poor solvent conditions; this simply indicates that fewer water molecules stick to PEG ones. The range of concentrations for good solvent conditions is extended when the temperature is increasing, but above 40 ◦ C there is another bad solvent zone. It is worth mentioning that the conclusions drawn from our cac measurements are consistent with these literature data, in the sense that zones of different types of interaction with water have been observed. Our transition solvent region should be consid- ered as a region in which PEG–water interactions are strong inside the helix but the whole helix behavior yields aggregation.
To gain more insight into the molecular behavior of PEG400 in water, we carried out several 13C NMR experiments.

3.2.NMR study

3.2.1.Spectra
Fig. 3 shows a typical 13C NMR spectrum recorded at c 0.29 molL-1 and t = 40 ◦ C. One should note that the =carbon labeled 1 is the closest to the labile lone-electron pairs of the oxygen atom of the alcohol group; as a result it is also the most shielded. The different peaks were assigned according to the most probable shielding for carbon nuclei and from the rel- ative numbers of the different nuclei. The assignments were confirmed from HMBC and HMQC spectra.

3.2.2.Chemical shifts
Spectra similar to those in Fig. 3 were recorded at different temperatures from a PEG/water solution at five different con- centrations. A plot of chemical shift variations versus tempera- ture for the three types of carbon atoms present in the solution highlighted several behaviors.
Fig. 4 illustrates that, at c = 0.257 molL-1 , carbon 1 was unshielded by the rise of temperature likely further to changes in the interactions between the –OH group and water due to the mobilization of the lone electron pair. This effect was unaltered by changes in concentration and temperature. The variations of chemical shift result from a modification of the electronic en- vironment of the probed carbon. The continuous deshielding

Fig. 3. 13 C NMR spectrum at 40 ◦ C and c = 0.29 molL-1 and line assignment.

PEG400 chain, which protects carbon 3 from any modification in the water environment at a given temperature.
This zone width was reported as vertical lines for three con- centrations in Fig. 2 to allow direct comparison with cac mea- surements.
Fig. 5 shows that aggregation affects the variation of the chemical shift versus temperature: at a given concentration, the chemical shift remains constant in the zone of temperatures where surface tension data suggest the existence of clusters. Ac- cording to Refs. [1,2,11], the hydration number is lowered by elevating the temperature; this number is related to PEO confor- mational changes and implies stronger entropic restrictions for the H-bonds between PEO and water. It affects carbon shield- ing, in agreement with the observations made on Figs. 4 and 5. The zone of temperature considered as the one where water is the better solvent [1] (30–35 ◦ C) is the one where we observed

Fig. 4. Displacement of carbon 1 chemical shift versus temperature at c 0.257 molL-1 . =

process shown by carbon 1 suggests a temperature-dependent evolution of the PEG400 chain end hydration, which can be as- cribed to a change in the number and strength of the H-bonds built by –OH chain ends. By inducing variations of the delo- calization potential of the free electron doublets of the oxygen atom, this modification of H-bonds in number and strength af-
fects the carbon neighbors. However, at c = 1.99 molL-1 , the chemical shift decreases with temperature (not shown), show- ing a shielding of carbon 1. This different behavior appears at the concentration where gel formation has been previously identified [9].
Fig. 5 displays the variations of the chemical shift versus temperature for the carbon denoted 3. The changes observed in the case of carbon 2 are quite similar and not shown here. For both carbons the increase of temperature yielded a decrease of the chemical shift, which we ascribed to a shielding process. It results from change in hydration, i.e., in the number and arrangement of the water molecules around the carbons.
Figs. 5a–5c (molarities in the range 0.097–0.29) show a zone of temperatures free from chemical shift change; this plateau becomes larger when the concentration is elevated.
The plateau observed for carbon 3 (or 2) can be explained by a stop of the shielding process caused by aggregation of the
aggregation in a larger concentration zone. This finding may appear puzzling, because aggregative behavior is generally con- sidered a way for a system to solve a conflict between a solute and a solvent; this conflict should be less marked in a good solvent where polymer molecules are extended. However, the particular structure of PEG chains calls for care in considering this kind of reasoning.

3.2.3.Intensities
Fig. 6 highlights the changes in intensity of the carbon 3 peak. In the event of aggregation, carbon 3 should be embed- ded in the aggregative structure. Thus, in a given solution, the occurrence of aggregation likely induces a reduction of peak in- tensity because of the immobilization in aggregates of carbons, which become unable to respond in the liquid state NMR. This assumption was corroborated by Fig. 6; indeed, it shows, at two different concentrations, a decrease of intensity, which remains constant, given the reproducibility of data, in the zone of tem- peratures where aggregation is expected.

3.2.4.Possible aggregation process
According to many authors [12,14], PEG chains, which are characterized by short carbon links separated by oxygen atoms, can form helices, under the conflicting influence of both hy- drophilic and hydrophobic parts of the chain on water. Fig. 7 was inspired by Ref. [7], where molecular-dynamics simula-

(a) (b)

(c) (d)

(e)

Fig. 5. Displacement of carbon 3 chemical shift versus temperature at several concentrations: (a) c = 9.75 × 10-2 ; (b) c = 0.257; (c) c = 0.295; (d) c = 1.34; (e) c = 1.99 molL-1 . The dashed lines show the temperature range in which 13 C chemical shift is constant.

(a)

(b)

Fig. 6. Variation of carbon 3 peak intensity versus temperature: (a) c = 0.290; (b) c = 0.257 molL-1 .

tions at 300 and 373 K in aqueous solutions were presented. They showed that the most likely configuration of a rather short PEO (15 repeat units) was helical, with oxygen atom stuck in- side the helix, i.e., in fact, linked together by H-bonds through water molecules, and ethyl groups displayed outside the he- lix. We added to this PEO-inspired structure two hydroxyl end groups in free contact with water. It is worth emphasizing that Fig. 7 was built by assuming nine repeat units, which is the right structure for PEG400. These descriptions are consistent with IR studies [3], which show predominant gauche structures around the C–C bond, as well as twisting of carbon orbitals leading to particular structures of the polymeric chain. This spectrophotometric argument, as well as the amphiphilic argu- ments used below, make it not very likely for PEG to adopt the structure with oxygen bridges outside, as it has been possible to argue in other situations, namely in simulations at 0 K, very far from our experimental conditions [15]. Recent spectroscopic data [16], particularly Raman scattering, confirm the confor- mational change in crystalline polyoxyethylene glycol and PEG

Fig. 7. Simulation of the PEG chain in water (in the manner of Ref. [7]).

solutions. According to this study, the consequence of the popu- lation of gauche conformation being dominant around the C–C bond in aqueous solutions at PEG concentrations close to 20% is the unfolding of PEG chains and the formation of helices.
Fig. 7 shows that such helical PEG chains can be consid- ered as amphiphilic polymers, i.e., as constituted of an aliphatic chain terminated by hydrophilic groups at both ends. When the concentration of such items is large enough, interactions with water may lead to a lateral aggregation producing a bundle of stick-like structures that minimize the contact of aliphatic parts with water and let free the terminal hydroxyl groups. Hy- drophobic interactions may be involved in this process. We wondered about the aggregation process and assumed that a close examination of both the configuration of chains developed through interaction with water, including good solvent areas, and our NMR data should give better insight into this question. By considering Figs. 3, 4, and 5, it is obvious that carbon 1,
which holds the -OH group, is unaffected by the aggregative process, likely because aggregation affects neither this group, which lies outside the aggregates, nor its interaction with wa- ter at the origin of carbon 1 shielding. Carbon 1 chemical shifts vary linearly with temperature, like the hydration number of the chain ends. Conversely, there is no change in the environment of carbons 2 and 3, because they are both embedded within the cluster at a given concentration; as a result, a variation of tem- perature over some tens of degrees is likely devoid of effect upon the environment and, thus, upon the shielding. This the- ory explains the plateau observed in our figures, but also the further formation of gels (when concentration is raised), since most of these structures form from amphiphilic entities stuck together in ordered states. This explains why constant values of chemical shifts (or intensities) for carbons 2 and 3 are observed in the zone of “spread gels” or gel (Figs. 5c, 5d, 5e, and 6b). Some authors [1,12] ascribed the concentration effect to the ac- cessibility of the water molecules that makes the PEG structure less rigid in the concentrated state due to H-bonds. In the di- luted state, the effect of the temperature is enhanced, since the flexibility of the chain can strongly modify the number of PEG- interacting water molecules (second hydration shell).
Finally, this behavior may be interpreted as a miscibility gap, which is governed on the one hand by the conformation of the PEG and, on the other hand, by the expected increased role of the entropy term with increasing temperature, which would work in the opposite direction. In fact this is counterbalanced by the loss of entropy resulting form the local order introduced by the helices structuring.
This behavior is different from the well-known behavior of PEGs with much higher molecular weight, which exhibit in ad-

Fig. 8. Solid–liquid phase diagram of PEG400–water mixtures. (+, —) repre- sent our own data. Other points are taken from Ref. [10].

dition a liquid–liquid demixing at very high temperatures [7,17, 18].

3.3.Solid/liquid demixing

Finally, we determined the phase equilibria at low tempera- tures. In Fig. 8, the curve has been drawn from our own data (crosses), whereas other points were drawn from Ref. [10]. Fig. 8 highlights that the solid–liquid diagram of PEG400– water mixtures is single-eutectic. Glassy transitions prevent one from observing the eutectic itself, but its composition is ob- viously about 60 wt%, i.e., 15 water molecules per PEG unit, which is consistent with the literature [1,12,17,19]. From this average number of water molecules forming the eutectic com- pound it is obviously difficult to determine whether they are homogeneously distributed along the PEG chain in the crys- tal structure or mainly bonded to the chain ends. However, it is worth noting that this eutectic composition is almost the one where, at higher temperatures, a gel appears. This phase dia- gram at low temperature is consistent with literature data [10, 19].

3.4.Complete phase diagram

Fig. 9 illustrates the different physical states observed in PEG400 aqueous solutions with respect to concentration and temperature. Reports about the crystallization process at low temperatures [8], the occurrence of gelling [9] at high concen- trations, and variations of the solvent properties of water with temperature [1,9] are available in the literature. These proper- ties have been established by very different physical methods, including spectroscopy and viscosity in PEG with various num- bers of repeat units (from PEG400 up to PEG4000). The data in Fig. 9 are specific to PEG400; they evidence, for the first time, a zone of aggregation of micelle-like forms. It constitutes de facto a local scale demixing area, which manifests itself in another form for PEG of higher molecular weight, but also pre- figures more extended structures further constituting gels. The dashed line in the upper part of the diagram limits the zone of aggregates determined by the end of the plateaus on surface ten-

Fig. 9. Complete phase diagram. (♦) cac line; (—) crystallization lines; (- – -) end of aggregate zone; (—) beginning of gel zone; (···) path for NMR mea-
surements versus temperature.

sion isotherms. A second and thick line indicates the occurrence of a gel, already mentioned in the literature [9], which, appar- ently, corresponds to the minima on surface tension isotherms. The spread-gel term, used for the fluid state observed between the aggregate zone and the well-established gel zone, is used by analogy with amphiphile phases in these intermediate states. We observed there an increasing viscosity for our mixtures, but we did not study these more precisely. Indeed, the complete diagram, considered in its whole, presents the main features of an ordinary amphiphile; it could have been produced by a soap. This surprising result was previously explained by the he- lical structure spontaneously adopted by the polymeric chain in a good aqueous solvent. Indeed, this structure is obviously amphiphilic. It would be worth investigating the role of this structure in the formation of gels and in the solid structure ob- served at low temperatures. We qualified as flock the probably floppy structure chains adopted outside the helical zone.

4.Conclusions

After this study, our knowledge of the bulk phase diagram of PEG400–water mixtures gained in clarity and precision. A new demixing area was shown, and its existence was correlated with helix formation by PEG molecules. These helices seem to be the driving force for the aggregation, but also, considering the diagram in its whole, for gel formation at room temperatures and probably for solid formation at low temperatures. Evolu- tion of molecular structures, as well as phase diversification, is an interesting hint to better understand how to use this kind of compound in several applications.
This study was based on original investigations by sur- face tension measurements and NMR analysis of PEG400– water solutions as a function of concentration and tempera- ture. Hydroxyl-terminated PEG exhibited amphiphilic behav- ior, which led to clustering above 0.2 mol L-1 and around 30 ◦ C.
We found some fascinating correlations between critical ag- gregation concentration measurements and the evolution of the chemical shift of the PEG400 carbons. These experimental data

allowed us to determine the phase diagram that highlighted the aggregative behavior of the PEG400 molecules.

References

[1]A. Faraone, S. Magazù, G. Maisano, P. Migliardo, E. Tettamanti, V. Villari, J. Chem. Phys. 110 (1999) 1801.
[2]C. Branca, S. Magazù, G. Maisano, P. Migliardo, E. Tettamanti, Physica B 270 (1999) 350.
[3]R. Begum, T. Yonemitsu, H. Matsuura, J. Mol. Struct. 447 (1998) 11.
[4]M. Rozenberg, A. Loewenschuss, Y. Marcus, Spectrochim. Acta Part A 54 (1998) 1819.
[5]K.J. Liu, Polym. Sol. 1 (1968) 213.
[6]M.M. Hoffmann, M.E. Bennett, J.D. Fox, D.P. Wyman, J. Colloid Inter- face Sci. 287 (2005) 712.
[7]K. Tasaki, J. Am. Chem. Soc. 118 (1996) 8459.

[8]M. Polverari, T.G.M. van de Ven, J. Phys. Chem. 100 (1996) 13687.
[9]A. Vergara, L. Paduano, G. D’Erico, R. Sartorio, Phys. Chem. Chem. Phys. 1 (1999) 4875.
[10]R.L. Davidson (Ed.), Handbook of Water-Soluble Gums and Resins, McGraw–Hill, New York, 1980.
[11]E.E. Dormidontova, Macromolecules 35 (2002) 987.
[12]R. Kjellander, E. Florin, J. Chem. Soc. Faraday Trans. 77 (1981) 2053.
[13]V.B. Fainerman, R. Miller, H. Möhwald, J. Phys. Chem. B 106 (2002) 809.
[14]B. Hammouda, D.L. Ho, S. Kline, Macromolecules 37 (2004) 6932.
[15]R. Gaudreault, T.G.M. van de Ven, M.A. Whitehead, Mol. Simul. 32 (1) (2006) 17.
[16]M. Kozielski, M. Mühle, Z. Blaszczak, M. Szybowicz, Cryst. Res. Technol. 40 (2005) 466.
[17]G. Karlström, J. Phys. Chem. 89 (1985) 4962.
[18]T. Sun, H.E. King Jr., Macromolecules 31 (1998) 6383.
[19]L. Huang, K. Nishinari, J. Polym. Sci. Part B 39 (2000) 496.