Phase separation in paraffin blends by simultaneous small-angle scattering and calorimetry

Elliot Gilbert, David Sutton and Andrew Nelson (ANSTO), Nick Terrill (DIAMOND), Chris Martin (Daresbury), Jyotsana Lal and Ed Lang (Argonne). 

Simultaneous synchrotron small-angle x-ray scattering (SAXS) and differential scanning calorimetry (DSC) studies on paraffin blends of composition 2:1 C28H58:C36D74 have been performed as a function of cooling treatment from the melt (Figure 1).

In previous studies, we have shown that these blends phase separate in the solid state to form lamellar superstructures whose spatial dimension may be controlled by chain-length difference, molar composition, isotopic substitution and confinement.

Figure 1. Temperature-dependent SAXS and DSC from i) 100ºC/min, ii) 0.1ºC/min, iii) 63.2ºC-annealed and iv) 54.0ºC-annealed samples.

Six cooling protocols were investigated in taking the blend from 100ºC (approximately 30º above the molten state) to 30ºC (approximately 40º below) (Figure 2).

The annealing temperatures chosen (63.2 and 54.0ºC) represent the mid-point between the lowest temperature phase transitions in the pure materials and the mid-point between the chain mixing transition temperature and the lowest temperature C28H58 transition respectively.

Figure 2. Graph of cooling protocols used in simultaneous SAXS/DSC experiments.

SAXS measurements were performed on the new MPW6.2 station at the SRS, Daresbury and represent the first SAXS user experiments from the new facility. To minimise radiation damage to the sample a fast opening shutter was employed that was programmed to open for the collection of the SAXS data and was closed at all other times.

The snapshots were of 6 second duration taken every 30 seconds, corresponding to data collection in 0.5º intervals. SANS experiments were performed on the SAND instrument at IPNS, Argonne National Laboratory. The specially modified DSC pans are shown in Figure 3.

Figure 3. Schematic of modified DSC cell. The cell is composed of the aluminium pan and lid on the outside, followed by mica windows with the sample material in the centre.

DSC shows that the rapidly quenched sample gives rise to the most disordered superstructure; an additional endotherm is exhibited relative to both pure materials (not shown) whose transition is close to the monoclinic to rotator transition observed for C28H58. Based on our previous SANS studies, we may attribute the lower temperature endotherm to a mixing transition above which the demixed lamellar arrangement is destroyed (Gilbert et al. J. Phys. Chem. 1996;100:18201-18213.).

The 63.2°C-annealed sample, while not exhibiting this additional endotherm, shows a strongly asymmetric lower temperature transition. This is in contrast to the sharp endotherms observed for the slow cooled and 54.0°C-annealed samples indicative of more extensive superstructure formation.

For the rapidly cooled and higher-temperature annealed mixtures, a broad SAXS peak is observed at 30°C corresponding to 43.2 and 42.4 ? structures respectively. Conversely, the 54.0°C-annealed sample exhibits a single sharp peak at ca. 41 ? and the slow cooled sample a double peak with the lower angle feature also at ca. 41 ? (Figure 1). These numbers are in excellent agreement with Vegard's Law.

The larger spacings observed in the former two samples indicate the formation of quenched-in voids. The higher q peak in the slow-cooled sample is associated with a distance of ca. 40 ?. Strictly speaking, it is not possible to unambiguously assign this peak to a second phase or a superstructure as the behaviour of the higher order reflections has not been probed over the experimental q range.

One would expect however that for a separate phase, the separation, Dq, between the peaks would increase with q; for a superstructure, Dq would be constant with q. SANS is able to assist with this determination (see below).

The variation in SAXS with increasing temperature is complex, yet a consistent picture emerges. Below the mixing transition, the Bragg peak associated with the demixed structure shifts to larger d-spacing and reaches its maximum intensity at approximately the peak transition temperature. This may be attributed to a gradual mixing of the chains and a disruption of the lamellar boundaries leading to an increase in the number of voids. At the same temperature, a second higher q peak is apparent.

Its intensity increases with temperature, while its position also increases with temperature until reaching a plateau below the melting transition. The lower d-spacing peak indicates an apparent shortening of one or both of the chains either via the formation of trans-gauche defects or a rotator phase. While more extensive order is expected using a slower cooling rate, a comparison of the two annealed samples indicates that maintaining the temperature a few degrees above the mixing transition greatly assists in the more ordered superstructure formation.

The SAXS technique provides useful information on the 'average' lamellar chain structure but is far less sensitive to the modulated arrangement of chains. This is due to the minor difference in X-ray scattering length density between the components. The use of SANS, with one of the chains is isotopically labelled, enables the separation of deuterated chains to be easily determined.

Figure 4. SANS from 2:1 C28H58:C36D74 mixtures highlighting effect of cooling protocol: quenched 100ºC/min (crosses), slowly cooled at 0.1°C/min (open circles) and two-step quenching from 100°C to 61°C (annealed 11h) and 61°C to 25°C (closed circles).

Figure 4 shows the SANS from this system exhibits a characteristic spacing of approximately 120 ?. This is in excellent agreement with a hypothetical superstructure based on a repeating arrangement of two C28H58 chains followed by a C36D74 chain in their orthorhombic packing arrangement (d = 123 ?).

A comparison of the SANS data indicates more extensive formation of the demixed structure from both the scattering intensity and peak width in the slow cooled and low-temperature annealed samples. This study highlights the complementary use of neutron and X-radiation for these type of studies.