Poly (vinyl alcohol-vinyl acetate CO-) complex formation with anionic surfactants particle sizeRead more...
a b s t r a c t
Colloidal aggregates, the so-called “pseudo-micelles” or nanogels, in view of their size in the nanometer range, are present in aqueous solution of vinyl alcohol-vinyl acetate copolymers (PVA) as a consequence of hydrophobic interaction between acetate sequences. The characteristics of PVAs, which are of interest as polymeric emulsifiers in suspension polymerization, were examined by dynamic light scattering (DLS). Different samples of average hydrolysis degrees (DH) from 73 to 88 mol% were studied. These nanogels corresponding to volume fractions up to 30% and in the size range of 30–40 nm, could be disaggregated by complex formation with sodium dodecyl sulfate (SDS). The evolution of the particle size was determined as a function of SDS concentration and temperature. Relative small amounts of SDS, typically 1–5% with respect to PVA, induce the disaggregation of the nanogels. By complex formation, the cloud point of PVA is shifted to higher temperatures, which are adjustable by the SDS concentration.
At present it is well established that the interactions between polymers and surfactants in aqueous media give rise to the formation of association structures, thereby modifying the solution and interfacial properties [1–3].
These interactions, studied by various techniques such as NMR, SLS, DLS, fluorescence spectroscopy, etc., are of importance to many applications including emulsion and suspension polymerization processes as reported by Gosa and Donescu .More recently minisuspension polymerization of styrene in the presence of PVA–SDS colloidal stabilizers was described by Ramirez et al. .
Poly(vinyl alcohol-co-vinyl acetate) copolymers, currently referred as PVA, that are obtained by partial hydrolysis of poly(vinyl acetate) PVAc, are also well known as stabilizers for the suspension polymerization process of vinyl chloride monomer (VCM).
The molecular characteristics of PVAs, mainly the average degree of hydrolysis (DH) and their polymerization degrees (DPn, DPw), have a major influence on the monomer droplet size and on the properties of the final PVC resin [6,7].
An additional feature is that small gel particles, so-called microgels or “pseudo-micelles”, which are generally present in PVAs may have an influence on their emulsifying efficiency. These PVA microor nanogels are formed either by intermolecular paracrystalline domains and/or by hydrogen bondings for PVAs with DH values close to 100 mol% or by hydrophobe–hydrophobe interactions for PVAs with DH around 70–90 mol%.
The first type of microgel formation for PVA aqueous solutions was reported in the early sixties by Matsuo and Inagaki . By static light scattering (SLS) experiments, these authors proved that the stable microgel particles, formed with PVA of high hydrolysis degree (DH ≈ 99–100 mol%) have a typical paracrystalline structure which is not disintegrated by heating the solution at 100 ◦C.
This observation was confirmed later on by different authors [9–11], who demonstrated that, in addition to formation of paracrystallin domains, intermolecular hydrogen bonding is involved in this association mechanism. By viscometry and SLS experiments Wang et al.  showed that for PVAs (DH = 100 mol%) the size of the colloidal particles is in the range of 6.3–26 nm, with increasing polymerization degrees from 500 to 8000.
The self associations of PVAs is not limited to those polymers of high DH, it could even be more important for PVAs which are obtained by partial hydrolysis of PVAc. Typically water-soluble copolymers with remaining vinyl acetate (VAc) contents of 10–30 mol% (DH = 70–90 mol%) have a strong tendency to form colloidal aggregates, still by hydrogen bonding but mainly by hydrophobic–hydrophobic interactions between the VAc sequences of the copolymers. This micellar type of association was well illustrated by different authors [13–15], in particular by Lewis and Robinson  and by Aladjoff et al. .
These authors, as well as Meehan et al.  showed that these colloidal aggregates, but not the paracrystalline structures, are disrupted by complex formation with anionic surfactants, such as sodium dodecyl sulfate (SDS) or ammonium laurate. Lewis and Robinson  came to this conclusion by viscosity measurements, whereas Aladjoff et al. [17,19] and later on Meehan et al.  confirmed this effect by SEC and HPLC techniques. The approach of these authors was based on the well-known concept of complex formation between anionic surfactants and water-soluble polymers which was investigated, for PVA and PVA based copolymers, over the last decades by quite a number of authors [20–26].
By different techniques, mainly by viscometry and light scattering, it could be shown that in aqueous solutions SDS binds readily to PVA preventing the intermolecular interactions and leading to the dissociation of the multimers.
Most of these authors were more specifically interested in the surfactant micellization behaviour induced by the presence of PVA. However, to the best of our knowledge, no detailed investigation was reported on the size characteristics of the existing micro- or nanogel particles and their evolution in the presence of SDS.
The interest of water soluble polymers–surfactant complexes in a large variety of applications, and in particular in emulsion and suspension polymerization, prompted us to examine the colloidal aspects of PVA micro- or nanogels, designated afterwards by nanogels in view of their size in the nanometer range, and their disaggregation in the presence of anionic surfactants.
The objective of this study was to investigate by dynamic light scattering (DLS) the size characteristics of PVA nanogels and their disaggregation by complex formation with SDS. This disaggregation phenomenon was examined for a series of PVAs of different molecular weights with DH values in the range of 70–90 mol%, which are generally used in the industrial production processes of PVC, and for SDS concentrations below the CMC value of this surfactant.
In parallel, the influence of temperature was checked in order to confirm that even at low SDS concentrations the cloud-point of these PVA–SDS systems can be shifted to higher temperatures, which is of practical importance for the PVA applications as stabilizers in emulsion and suspension polymerization.
The PVAs examined in this study were supplied by Kuraray, Nippon Gohsei and Synthomer. These samples, hereafter used without further purification, are identified by their average DH and DPw, as for instance PVA-73-700. The main characteristics of the PVAs, determined by 1H NMR and SEC, are summarized in Table 1.
The average hydrolysis degree, DH with a precision of ±1mol%, was determined using 1H NMR (Bruker AC-400F operating at 400 MHz) in dimethylsulfoxide (DMSO)-d6 at 70 ◦C according to van der Velden and Beulen . These characteristics were confirmed by 13C NMR spectroscopy of the polymers solubilized in a 50/50 (v/v) D2O and deuterated acetone mixture. This technique gives in addition access to the average sequence lengths of vinyl acetate nVAc 0 and vinyl alcohol nVOH 0 respectively, defined by Moritani Moritani and Fujiwara . It may be noticed that sample PVA-73-700 has the highest VAc sequence length nVAc 0 .
The SEC measurements were carried out with a Shimadzu LC- 20AD liquid chromatograph equipped with two Varian PL gel 5m MIXED-C columns (column, injection and refractometer temperature: 30 ◦C; injection volume: 100L; solvent: tetrahydrofuran at 1mLmin−1) and a refractive index detector (Shimadzu RID-10A). The PVA samples were at first reacetylated as recommended by Bugada and Rudin  and the “universal calibration technique” with polystyrene standards was applied for the calculation of Mn, Mw and the polydispersity index PI = Mw/Mn, the highest values are those for samples PVA-73-1250 and PVA-78-720.
Sodium dodecyl sulfate (SDS), obtained from Acros Organics with a purity of 99%, was used without any further purification.
2.2. Sample preparation
The PVA solutions were prepared by dissolving under agitation the required amounts of PVA in triple distilled and filtered (0.22m Millipore filter) water at room temperature for 24 h, for samples with DH = 73–78 mol%. For the PVA samples with DH = 88 mol% the aqueous solutions were heated at 80 ◦C for 30 min for total dissolution of polymer. The solutions are then agitated for 24 h at room temperature.
An alternative method for the preparation of the PVA solutions was to dilute a 2 wt% “stock solution”, prepared as aforementioned, to the required concentration in order to demonstrate that the nanogel formation is independent of the dissolution process.
For the cryogelation experiments, the PVA solutions at a given concentration were placed in a freezer at −20 ◦C for 20 h and followed by a 4 h thawing step at room temperature.
The SDS solutions were prepared by dilution of a 1 wt% “stock solution”. To avoid the hydrolysis of SDS, all surfactant solutions were used in within 24 h.
For the preparation of PVA/SDS solutions two methods were used.
In the first procedure, and if not otherwise stated, PVA is directly solubilized in the aqueous SDS solution at the required concentration, the solubilization being achieved, either at room temperature or at 80 ◦C according to the aforementioned procedure for the PVA samples having a DH value of 88 mol%. The second procedure, that would avoid the hydrolysis risk of SDS at 80 ◦C, consists in dissolving SDS in the previously prepared PVA dissolutions. In both cases, the size determinations are carried out after agitation during 24 h at room temperature.
Before use, all solutions were filtrated over 0.45m Chromafil Xtra MV-45/25 filter.
2.3. Dynamic light scattering
DLS measurements were carried out on a Malvern Nano-ZS6 Zetasizer equipped with a 4mW He–Ne laser operating at a wavelength of 532 nm. The measurements were made at a scattering angle = 173◦ at a fixed temperature in the range from 20 to 60 ◦C, by taking into account the viscosity variations as a function of temperature. Quartz cuvettes were used for all the experiments.
The data were acquired with the Malvern’s Dispersion Technology Software version 4.20. The software package of the instrument calculates, by using the Stokes–Einstein equation, the hydrodynamic diameter (volume average) Dv, the Z-average, which is an intensity weighted size average and the polydispersity index PDI of the sample.
The PDI is given on a scale from 0 to 1, where 0 represents the highest level of monodispersity. In addition to Dv, and for comparative studies the Z-average was taken as representative of the hydrodynamic diameter and size distribution of the samples, as long as PDI < 0.5 .
To determine the diameter of the particles, the data were collected in automatic mode, typically requiring a measurement duration of 70 s. The “data quality report” incorporated in the software indicated “good quality” for all the obtained data. For each experiment at a given temperature, the average of 5 consecutive measurements is indicated in the tables and figures. For all the samples characterized by their Z-average, the PDI was in the range of 0.2–0.4.
2.4. Cloud point determination
The starting PVAs and their complexes with SDS were characterized by their lower critical solution temperature (LCST). The onset of precipitation as a function of temperature was determined by monitoring the turbidity of the solution at a concentration of 1 wt% as the temperature was increased stepwise at a rate of 1 ◦Cmin−1. The cloud point is given by the temperature corresponding to the onset of the turbidity curve.
2.5. Viscosity measurements
Viscosities of thePVAsolutions were determined using anAMVn Automated Micro Viscometer from Anton Paar. The measurements were carried out as a function of the polymer concentration and the intrinsic viscosity  was determined by extrapolating the reduced viscosity to zero concentration.
2.6. Aqueous SEC measurements
GPC profiles were obtained on a gel permeation chromatograph (PL-GPC 120, Polymer Laboratories) equipped with three PL Aquagel-OH 20/40/60, 8m, 300mm×7.5mm columns, a IR detector and a high resolution data acquisition system (PL Data Stream).
All the measurements were made at 20 ◦C either with water or SDS aqueous solutions as eluent, with aflowrate of 1ml/min and an injection volume of 100l. Calibration curves were generated using PEO standards in the molecular weight range of 420–895 500 mol/g.
3. Results and discussion
3.1. Viscometry and SEC of PVAs without SDS
In order to complete the characterization of the PVA samples in the absence of SDS, their intrinsic viscosity , was determined in aqueous solution at 20 ◦C. The corresponding values are listed in Table 2, with the linear correlation coefficients R2 and the Huggins constants, kH, given by sp/c =  + kH2c + . . ..
The intrinsic viscosity being an indication of the hydrodynamic volume occupied by a single chain, the critical overlapping concentration, can be computed as C* = 1/. A similar approach was indicated by Hong et al.  for fully hydrolysedPVAand by Budhall et al.  for partially hydrolyzed PVAs.
From Table 2, it appears as expected that  increases with increasing Mn or Mw values. Although, water at room temperature is not a “poor solvent” for PVA, the Huggins viscosity coefficient, kH, are typically higher than 0.5, especially for samples PVA-73- 700 and PVA-73-1250 having the lowest DH values. According to Lewandoska et al. , who presented a compilation of Huggins constants for PVAs, high kH values are a typical indication of interand intramolecular associations. For PVAs of DH values around 88 mol% no intermolecular associates could be detected according to Budhall et al. .
On the contrary for PVAs with DH of 70–80 mol%, the aqueous systems are more complex; they contain a certain fractions of nanogels in the presence of molecularly dispersed species as clearly demonstrated by the SEC experiments of Aladjoff et al. . These SEC experiments were confirmed for two typical PVAs of our series. For PVA-88-2200, with aDHvalue of 88 mol%, a chromatogram with amonomodal distribution could be obtained with a peakmaximum at 21.33 min, whereas those of PVA-73-700 with DH values of 73%, have distinctly a bimodal distribution. A first peak of high molecular weight appears for PVA-73-700 at an elution time of 18.65 min with a second peak at an elution time of 25.75 min. For the very high molecular weight species of PVA-73-700, the relative volume fraction, estimated by extrapolation to zero concentration, turns out to be around 26%.
A complete quantitative evaluation of its molecular weight was however not possible as the elution time of these nanogel fractions was beyond the largest PEO standards, that ofM= 895 500 available for these experiments.
In addition, for these PVAs having a polydispersity in molecular weight and composition, the peak areas determined with a differential refractive index detector may not be proportional to the actual volume fractions, as demonstrated by Dawkins et al. .Moreover, these copolymers nanogels, in the size range of 20- 50 nm,might have a strong tendency to be adsorbed in the columns, as pointed out by Tuzar and Kratochvil  for copolymer micellar systems.
3.2. Dynamic light scattering of PVAs without SDS
The presence of a nanogel fraction in the PVA samples of lower DH values having been demonstrated by SEC, it was of interest to approach this problem in a systematic way by DLS. This technique has the advantage to lead to quantitative information concerning the volume fraction and the size of the nanogels, without disturbing the unimer/aggregate equilibrium .
The results showing the presence of aggregates in form of nanogels for PVAs of lower DH and their absence in PVAs having a DH value of around 88 mol% are therefore consistent with those obtained by SEC. They are further in agreement with the conclusions of Budhall et al.  who have shown by DLS that PVAs with DH values of 88 mol% lead to aqueous solutions without aggregates.
Table 3 summarizes the volume fractions and the size characteristics given by Dv of the different PVA samples, and for various preparation procedures, at a concentration of 1 wt% in water.
Nanogels in the size range of 36–47nm can be noticed for the samples with low DH values, their relative amount decreases, from around 20 to 10 vol%, with a DH increase from 73 to 78 mol%. These proportions are in fair agreement with the SEC results of Aladjoff et al. .
It also appears that, within the experimental error limits, neither the volume fraction nor Dv are influenced by the preparation procedure, such as dilution or prior cryogel formation (see details in Section 2.2). No nanogel formation could be observed for the two PVA samples with a DH of 88 mol%. For these monomodal systems, the Dv values are typical for hydrodynamic diameters of “free chains” in agreement with Budhall et al. . According to Damas et al.  the average hydrodynamic radius Rh as determined by DLS should also correspond to that calculated from the intrinsic viscosity  and the molar mass M, which is defined by Rh = (3M/10Na)1/3, where Na is the Avogadro number and M is either Mn or Mw.
This correlation between the hydrodynamic radius (Dv/2) extrapolated to zero concentration and the one calculated with  Mn and  Mw is shown in Table 4 for the monomodal systems corresponding to samples PVA-88-450 and PVA-88-2200.
From these results it turns out that Dv/2, the average hydrodynamic radius obtained by DLS measurements is in between the Rh values calculated by taking into account the intrinsic viscosity in combination with Mn and Mw respectively. A similar calculation is not possible for the systems having a bimodal size distribution as their intrinsic viscosity includes the contributions of the nanogels and the “free chains”. It is however interesting to notice that their Dv values of the “free chains”, determined by DLS, are in the same size range as those of the monomodal systems (see Table 3). In a first approximation it might therefore be concluded that this type of chains, present with the nanogels, are actually non-associated species.
In order to take into account possible viscosity effects of the medium, which may influence the hydrodynamic size character-istics of the dispersed species, the following of our study will be focused on dilute PVA solutions up to 2wt%. The extrapolated values to zero concentration of the size and volume fractions, as well for the nanogels as for the “free chains”, are listed in Table 5. All linear correlation coefficients R2 of these extrapolations were higher than 0.910.
It is worth noting that the PVAs of lower DH, from 73 to 78 mol%, have a volume fraction of “free chains” in the range of 70–80% with a Dv from 8 to 12.5 nm.
For the nanogels one can notice that for the PVAs of DH = 73 mol% their volume fraction is of 27–30% with Dv values of 30–35 nm. As well the volume fraction as the size are decreasing for sample PVA-78-720 having a DH of 78 mol%.
From these SEC and DLS results it appears clearly, in agreement with Aladjoff et al. , that the aqueous PVA dispersions of lower DH values contain a certain amount of nanogel and that their volume fraction decreases from around 30% to 20% with increase inDH from 73% to 78 mol%. For PVAs of higher DH, those of 88 mol%, no nanogels could be detected, which means, in agreement with Budhall et al. , that these samples can be considered as molecular dispersions in aqueous medium at 20–25 ◦C.
As well established by chromatographic and fractionation techniques, the nanogel particles have their origin in the polydispersity in molecular mass and in composition of the PVA copolymers, especially for the DH range of 70–90 mol% where these products are water dispersible [17,19]. In fact, as the water solubility decreases with decreasing DH value, the fraction with the lowest DH approaches its limit of solubility. Thus with increasing vinylacetate (VAc) contents, hydrophobic intra- and intermolecular interactions are becoming predominant with formation of associated species.
3.3. PVA/SDS interaction
Already in the midst of the last century, it was shown by viscometry that partially hydrolyzed PVAc forms polyelectrolyte complexes by hydrophobic interaction with anionic surfactants such as SDS [21,22,37]. These conclusions were confirmed in a systematic study carried out by Lewis and Robinson , who could further demonstrate by viscometry that PVA polymer aggregates are disaggregated by binding of SDS. Mention should also be made that several authors were interested in the micellization mechanism of SDS itself in the presence of PVA [25,26].
It is Aladjoff et al. , and later Meehan et al.  who have examined in more detail by SEC and HPLC this type of interaction between SDS and PVAs of various DH values. These authors could demonstrate that the aggregates formed in aqueous medium are disaggregated in the presence of SDS.
Our objective was at first to confirm by SEC that the nanogels formed by hydrophobic interaction of the vinyl acetate units (VAc) are disaggregated in the presence of SDS. Our main interest will be focused on monitoring by DLS the change in particle size of the nanogels induced by SDS at different concentrations below its critical micellar concentration in water (CMC).
The SEC experiment performed on the different PVA samples of lower DH (PVA-73-700, PVA-73-1250 and PVA-78-720) show clearly the disappearance of high molecular weight peak corresponding to the nanogels, even in the presence of relatively low concentration of SDS. For sample PVA-73-700, for instance, the nanogel peak at 18.65 min was no longer detectable at a concentration of 0.01 wt% SDS in the aqueous phase (which corresponds to 1 wt% with respect to PVA). In parallel, it could be noticed that the initial peak at 25.75 min, attributed to the “free chains” is shifted to a slightly higher elution time of 25.88 min. This shift is increased to 26.90 min for a concentration of 0.1 wt% SDS in the aqueous phase which corresponds to 10 wt% with respect to PVA.
On the contrary, for samples of higher DH and containing no nanogel, such as PVA-88-2200 with a DH value of 88 mol%, the elution time shift, from 21.23 to 21.25 min, is almost negligible at low SDS concentration.
DLS experiments show even more distinctly the disaggregation of the nanogels in the presence of SDS. This behavior is illustrated in Fig. 1 for a typical example that of PVA-73-700 and the data for all the samples as a function of the SDS concentration, are summarized in Table 6.
In the absence of SDS, samples PVA-73-700, PVA-73-1250 and PVA-78-720 have typical bimodal distributions corresponding to nanogels and “free chains” respectively, whereas a monomodal distribution is observed for PVA-88-2200 and PVA-88-450. For PVA-73-700 as shown in Fig. 1, the peak attributed to nanogels has its maximum at 40.3 nm, that for “free chains” being situated at 14.6 nm. The peaks being well-separated and at least by factor 2 as required by Malvern’s instruction  each population can thus be evaluated separately. One has therefore access to the corresponding volume fractions and average volume hydrodynamic diameter Dv listed for different PVAs in Table 6.
On addition of SDS, at a concentration of 1wt% with respect to PVA, the peak attributed to nanogels has partially disappeared. The monomodal peak obtained at this SDS concentration shows a remaining “tail” of a certain fraction of nanogels that are still present, however of lower size as the initial ones. It is interesting to notice that on the other side of the peak, a lower size fraction has been formed by disagregation of the nanogels. This seems to be an indication that the polymer chains associated in the nanogels are of lower molecular weight than those corresponding to the “free chains”.
At this point it is no longer possible by DLS to determine the size characteristics, such as volume fraction and size of each population. Only a mean value of Dv becomes accessible, although a certain amount of nanogel is still present.
Onfurther addition of SDS the whole distribution curve is shifted to lower size values, with amore or less important “tail”, at around 10nm corresponding to the complex of “free chains” and SDS.
For PVA-73-700 it can be admitted from the DLS distribution curves that the nanogels are almost disaggregated at a SDS concentration around 5wt% with respect to PVA. For sample PVA-73-1250 and PVA-78-720, these values are slightly higher, at SDS concentrations between 5 and 10 wt%.
The Dv values given in Table 6 for samples PVA-73-700, PVA- 73-1250 and PVA-78-720 at a 10% SDS concentration with respect to PVA might be attributed to the shorter chains liberated from the nanogels and/or to a clustering effect leading to the coil shrinkage as shown by Nilson , Liu and Hoffmann . In fact, the formation of SDS micelles as such may be excluded at this concentration of 10% with respect to PVA, which corresponds to 0.1 wt% in aqueous media, a value far below the C.M.C. of SDS, given as 0.24 wt% at 20 ◦C .
For multicomponent and more complex systems such as the present one, comprising nanogels, “free chains”, SDS complexes, it is recommended to use the Z-average particle size value, which as an intensity average allows tomonitor the evolution of the larger size particles of systems such as the nanogels .
Fig. 2. Schematic representation of mechanism suggested for the disaggregation of PVA nanogels. (a) At zero SDS concentration, in aqueous media the PVAs with a DH between 70 and 80 mol% contain nanogels, formed by hydrophobic interactions of VAc sequences, and “free chains”. PVAs with a DH between 80 and 90 mol% contain only “free chains”. (b) For the PVAs with a DH in the range of 70–80 mol% the addition of very small amounts of SDS (0–0.5 wt% with respect to PVA) corresponds to the beginning of the surfactant binding to the VAc sequences and to the disaggregation of nanogels with its size reduction. In the same range of SDS concentration, the interaction with PVA of higher DH value (88 mol%) is less pronounced as only minor changes in size were observed. (c) If [SDS] > 5% with respect to PVA all the nanogels are disaggregated and for all the PVAs a shrinkage of the polymer coil can be observed.
From the Table 6 it becomes possible to have an estimation of the SDS amount, where it can be assumed that all the nanogels are disaggregated. From these SDS concentrations and those of VAc units of the different PVAs, one has the possibility to evaluate on a molar basis the average number of SDS molecules per VAc unit. For PVA-73-700, PVA-73-1250 and PVA-78-720 these values are in the range of 12–20 mol.
At a given DH value of 73 mol% (samples PVA-73-700 and PVA- 73-1250) the difference in behavior might be attributed, either to a difference in polydispersity and/or to the higher “blockiness” of PVA-73-700 .
A plausible mechanism (Fig. 2) involved in the disaggregation of PVA nanogels as a consequence of complexation with SDS could be similar to that suggested by Nilson  for the system SDS/hydroxypropyl methyl cellulose (HPMC).
According to this author, as well as to Yan et al. , there is a so-called critical aggregation concentration (CAC) that corresponds to the beginning of the surfactant binding to the water-soluble polymer. It was shown by different techniques, including NMR, calorimetry, fluorescence spectroscopy, etc., that the onset of aggregation between SDS and polymer, such as PEO or poly(vinylpyrrolidone), is definitely below the CMC value of the surfactant in the absence of polymer [43,44].
For PVAs, with DH values in the range of 70–80 mol% the nanogels are formed by the most hydrophobic fraction of the samples. A reasonable hypothesis in this case may be that the CAC values are very low and that complexation starts at relatively low SDS concentration typically below 0.5 wt% with respect to PVA. The nanogel and the already present “free chains” participate at their turn in the complex formation with SDS. At increasing SDS concentrations, an intramolecular clustering may occur, leading to the shrinkage of the polymer coil, as shown by Yan et al.  for awater soluble acrylic copolymer and by Nilson  in the case of HPMC.
For PVAs with DH value of 88 mol%, containing predominately “free chains” and where only minor size change was observed by DLS (see Table 6) it might be admitted that the CAC starts at much higher SDS concentrations, most probably in the range of 5–10 wt% SDS with respect to PVA.
3.4. Cloud point determination and influence of temperature
The PVA samples in aqueous medium were characterized by their lower critical solution temperature that is determined by the onset of precipitation as a function of temperature [17,45].
The cloud points for the PVA samples with DH values of 73 mol% are listed in Table 7, keeping in mind that those of PVAs of higher DH are not directly accessible under our experimental conditions limited at 80 ◦C.
This table illustrates the cloud-point shift to higher temperatures with theDHincrease of the sample and the SDS concentration. For PVA-73-700, it is worth noting that even very low SDS concentrations, such as 0.5 wt% or 1 wt% with respect to PVA, lead to a cloud point shift of 20 ◦C and 46.5 ◦C respectively.
It is also of interest to compare sample PVA-73-700 in the presence of 0.5% SDS (with respect to PVA) having a cloud point at 48.5 ◦C, which is almost the same, e.g. 47 ◦C, as for PVA-78-720 with no SDS. This is further evidence that complex formation with SDS corresponds to an enhancement of hydrophilic characteristics of the polymer.
Moreover, at a given SDS concentration, as for instance 1 wt% with respect to PVA, the cloud point shift is the most pronounced for PVA-73-700, as compared to PVA-73-1250 and PVA-78-720, having on the average shorter VAc sequences (nVAc 0 values given in Table 1). A similar observation, concerning the influence of the sequence distribution on the cloud point shift, was reported by Benkhira et al.  for the complex formation of ethylene oxide-methylene oxide copolymers with SDS.
In completion of the cloud point determination, it was of interest to check the evolution of particle sizes as a function of temperature, at first in the absence of SDS, then in the presence of increasing amounts of SDS.
The influence of temperature on Z-average at various SDS concentrations with respect to PVA is illustrated by Fig. 3 for PVA- 73-700 at a concentration of 1 wt%.
From this figure it can be noticed that the Z-average particle size is increasing with temperature, especially near the cloud point. This behavior may be explained by a slight aggregation of particles mainly if the nanogels are still present in the starting system. The size variation with temperature is however less pronounced at higher SDS concentrations.
Similar behaviors were observed for PVA-73-1250 and PVA-78- 720 and the reversibility of the phenomena was demonstrated.
In this study we confirmed by DLS, and in agreement with different authors [16–19], that PVAs of lower DH, e.g. in the range of 73-78 mol% form nanogels by hydrophobic interactions between VAc sequences. The reversibility of this self-association process could be assessed.
In addition to the existing data of the literature, the volume fraction and the particle size of the nanogels could be evaluated as a function of the PVA concentration, itsDHvalues and polydispersity, as well as in function of temperature. It turned out that their size and volume fraction decrease with increasing DH values. On the contrary, their size and volume fraction increase with temperature and polydispersity in molecular weight.
According to Aladjoff et al. , it can be assumed that due to the polydispersity of the starting PVAs, the nanogel particles are formed especially with those fractions of higher VAc content and most likely with those of lower molecular weights.
No nanogels could be detected for PVAs with DH values of 88 mol% in agreement with Budhall et al. .
By DLS there is evidence that relatively small amounts of SDS, typically 1–5 wt% with respect to PVA, induce the disaggregation of the nanogels.
This interaction between SDS and the VAc sequences of PVA has furthermore a strong influence on the LCST values, as a dramatic shift to higher cloud point temperatures is observed. In fact, the SDS/PVA complexes are more hydrophilic that the starting PVA.
For PVAs with DH values in the range of 73–78 mol% it appeared that a complete disaggregation of the nanogels could be achieved with SDS solutions at concentrations below the CMC and for molar ratios of 10–20 mol SDS per unit mole VAc of the copolymer.
The progressive disaggregation of the nanogels with increasing SDS concentrations leads to “free chains” and to their shrinkage as a consequence of the complex formation with SDS, which induces a global shift to lower particle sizes. In addition to the shrinkage of the PVA/SDS polymer coil, it can also be assumed that the initial nanogels might be formed, not only by the PVA fraction of increased VAc content, but also of lower molecular weight than the average Mw value of the sample.
The self association process with formation of nanogels particles has in consequence to be considered as a fractionation in composition and molecular weight.
From amore practical point of view, these results are of interest for the application of PVAs in suspension polymerization processes. In fact preliminary tests, on chlorobutane/water model emulsions for VCM/water systems in the presence of PVA, have already shown that the droplet size and the stability of such emulsions is strongly influenced by the presence of nanogel and PVA/SDS complexes. Further work is also in progress concerning the Pickering type stabilization of emulsions with PVA nanogels.
The authors would like to thank T. Lasuye, B. Stasik and S. Michel for their interest in this study, as well as the various PVA suppliers. They also express their thanks to Dr. Doina Hritcu for her assistance with the aqueous SEC experiments.