Unique Behavior of Poly(propylene glycols) Confined within Alumina Templates Having a Nanostructured Interface

Herein we show that the nanostructured interface obtained via modulation of the pore size has a strong impact on the segmental and chain dynamics of two poly(propylene glycol) (PPG) derivatives with various molecular weights (Mn = 4000 g/mol and Mn = 2000 g/mol). In fact, a significant acceleration of the dynamics was observed for PPG infiltrated into ordinary alumina templates (Dp = 36 nm), while bulklike behavior was found for samples incorporated into membranes of modulated diameter (19 nm < Dp < 28 nm). We demostrated that the modulation-induced roughness reduces surface interactions of polymer chains near the interface with respect to the ones adsorbed to the ordinary nanochannels. Interestingly, this effect is noted despite the enhanced wettability of PPG in the latter system. Consequently, as a result of weaker H-bonding surface interactions, the conformation of segments seems to locally mimic the bulk arrangement, leading to bulklike dynamics, highlighting the crucial impact of the interface on the overall behavior of confined materials.

Anodic alumina membranes (AAOs), both const-AAO and modul-AAO, were prepared via two-step anodization of aluminum in acidic electrolytes. A high purity aluminum foil (99.999%, GoodFellow) was degreased in acetone and ethanol and electrochemically polished in a mixture of ethanol and perchloric acid (3:1 vol.) at 20 V and 0 ºC, for 2 min. Two types of AAO membranes were prepared with (i) a straight pore diameter (const-AAO) and (ii) modulated pore diameter (modul-AAO). The anodization of aluminum was performed in a two-electrode electrochemical cell in an (i) 0.3 M H2C2O4 and (ii) 0.3 M H2SO4 aqueous solution at a constant voltage of (i) 45 V and (ii) 25 V, and the temperature of (i) 25 ºC and (ii) 5 ºC. The duration of the first step was (i) 1 h and (ii) 12 h. After that, the unordered porous layer was etched away in a mixture of 6 wt% H3PO4 and 1.8 wt% H2CrO4 for 1 h at 60 ºC. The second step of anodization was performed at a constant (const-AAO) or pulse (modul-AAO) potential modes.
The anodization at a constant potential (const-AAO) was performed at the same conditions as the first step, except anodization duration, which was t = 4 h. While during the pulse anodization (modul-AAO), a series of 50 potential pulses comprising a mild anodizing (MA) pulse (25 V for 3 min) and hard anodizing (HA) pulse (35 V for 1 s) was applied.

Samples preparation
As first, the applied AAO templates were dried in an oven at T = 423 K under a vacuum (10 -2 bar) for at least t = 24 h to remove any volatile impurities from the nanochannels. After cooling, they were used as a constrain medium. For that purpose, AAO templates were placed in a small glass flask containing PPG. The whole system was maintained at T = 298 K in a vacuum (10 -2 bar) for t = 24 h to let all compounds flow into the nanocavities. After completing the infiltration process, the surface of the AAO membrane was dried, and the excess sample on the surface was removed by the use of a metal blade and a paper towel. Note that although the imbibition process of the modul-AAO would not follow the classical Lucas-Washburn equation 2,3,4 the infiltration time was set as t = 24 has in case of const AAO. After that time, we found that the mass of both kinds of membranes was constant, indicating the imbibition process to be finished. Before the measurements, samples were annealed at T = 353 K for t = 1 h and 12 h. Importantly, there was no impact of the time of annealing on the segmental and chain dynamics of PPGs incorporated within the porous alumina templates of constant and modulated pore sizes. For the AFM experiments, samples were fractured, and produced cross-section was investigated. In the experiment, we used membranes of comparable pore diameters, Dp ~ 30 nm.

Broadband Dielectric Spectroscopy (BDS).
Isobaric measurements of the complex dielectric permittivity ε*(ω) = ε'(ω)-iε"(ω) were carried out using the Novocontrol Alpha dielectric spectrometer over the frequency range from 10 -2 to 10 6 Hz at ambient pressure. The temperature stability controlled by Quatro Cryosystem using a nitrogen gas cryostat was better than 0.1 K.
Dielectric measurements of bulk PPG were performed in a parallel-plate cell (diameter: 10 mm, gap: 0.1 mm) immediately after preparation of the amorphous sample. AAO membranes filled with PPG were also placed in a similar capacitor (diameter: 10 mm, membrane thickness: 0.05 mm) 5,6 . Nevertheless, the confined samples are a heterogeneous dielectric consisting of a matrix and an investigated compound. Because the applied electric field is parallel to the long pore axes, the equivalent circuit consists of two capacitors in parallel composed of ε*compound and ε*AAO. Thus, the measured total impedance is related to the individual values through 1/Z*c=1/Z*compound +1/Z*AAO, where the contribution of the matrix is marginal. The measured dielectric spectra were corrected according to the method presented in Ref. 7 .
To determine the mean relaxation times of the normal and segmental modes, τNM and τα, the obtained experimentally dielectric spectra were further analyzed by the superposition of two Havriliak-Negami (HN) functions with an additional term related to the dc conductivity 8 : where αHN and βHN are the shape parameters representing the symmetric and asymmetric broadening of given relaxation peaks, Δε is the dielectric relaxation strength, τHN is the HN relaxation time, ε0 is the vacuum permittivity and  is an angular frequency ( = 2πf). Note that τNM and τα were estimated from τHN accordingly to the equation given in Ref. 9 .
To determine the glass transition temperatures of all examined materials (bulk and confined), the estimated segmental relaxation times were fitted by a combination of the Vogel-Fulcher-Tammann (VFT) and Arrhenius equations. It should be mentioned that both functions were applied only for the PPG incorporated into const AAO membranes due to the observed deviation in the slope of τα(T)-dependences at Tg,interfacial. Note that the VFT equation was used only for an accurate determination of a point (temperature), at which the slope changes; while, the glass transition temperatures of confined samples (in this case of the core molecules) were estimated from the extrapolation of Arrhenius fits. In turn, for the bulk samples and PPG infiltrated into const-AAO, only a single VFT equation was applied 10,11,12 : where τα, DT and T0 are fitting parameters. The Arrhenius equation was used as follows: where kB is the Boltzmann constant, and E is the activation energy. Determined values of Tg were added to Fig. 4. Note that Tg is defined as a temperature at which τα = 100 s and the glass transition temperature of core molecules of examined confined materials, determined from the extrapolation of the Arrhenius fit, is denoted as Tg,core.  The AFM technique can be directly implemented for the measurement of the adhesion force between the sample and the cantilever tip 14 . To do that, the tip is brought into contact with the sample at a constant velocity until the maximum load is obtained. Afterwards, the movement is reversed, and the sample and the tip are separated. However, due to the interaction between the tip and the surface, a certain force has to be applied to separate them again. This force is described as the adhesion force in our experiments, and it can be defined as the total force that is exerted by a liquid on the AFM tip. This force depends on several parameters. The most important one is a linear function of the liquid surface tension γL 15  It is crucial to mention that the presented adhesion map is a result of the analysis of a number of curves, 65536 per image (typical force-distance curve is shown in Fig. S6).

Fig. S6.
Typical force-distance curve measured on the empty AAO membrane.
One can notice a slight correlation between the topography and adhesion force maps most probably stemming from the increase of the adhesion inside the pore due to the presence of residual water. This fact shows that to obtain reliable values of the adhesion between the AFM tip and the PPG other factors, such as the presence of water, has to be factored. Thus, the data from tens of images for each sample was gathered (more than 300k curves) and compiled into We have gathered data from each revealed pore centre and averaged it over tens of pores; a typical path is showed in Fig. S8.
The results can be found in Table S1.