Enhance understanding of rhythmic crystallization in confined evaporating polymer solution films: from environment to solution film and then to one period

Abstract

It has been shown that rhythmic-crystallization-caused concentric ringed spherulites in polymer solution films can only be observed upon a confined slow evaporation. Here, by analysing the physical principle of evaporative crystallization, the special effects of experimental conditions on such rhythmic crystallization in a confined environment are explored via optical microscopy and atomic force microscopy observations. It is unveiled that the coupling of the confined conditions and extra solvent provides a suitable and stable solvent partial pressure that ensures the generation and continuation of rhythmic crystal growth. By slanting the substrate, the unusual dependences of radial growth rate and ring periodicity upon the film thickness are also illustrated. Finally, the growing front evolution and the melt-like growth in the induction time before the next period are demonstrated. These present findings significantly enhance our understanding of the rhythmic crystallization in evaporating polymer solution films.

---

Introduction

Ring-banded structures are frequently encountered in polymer spherulites formed either in the bulk or from a concentrated solution. Lamellar twisting has now became the most popular explanation for the formation of extinction banding in polarized light, and the uneven stresses upon opposite folded surfaces are generally believed to be the mechanical origin for the periodic twisting of lamellae.^1–9 However, concentric ringed morphology that shows the opposite optical features with extinction bands, i.e., a poor contrast under polarized light but a clear appearance in unpolarized light, has also been observed in polymer thin film systems, and twisted lamellae are absent.^10–13 Although it is now recognized that such a periodic structure is the manifestation of a structural discontinuity induced by rhythmic crystallization, the underlying cause has been rather elusive. Here we focus on exploring the rhythmic crystallization behaviour in evaporating polymer solution thin films under a confined environment.

Actually, rhythmic crystallization often occurs in small molecular systems, and well known is the occurrence of Liesegang rings.¹⁴ This mechanism has even been introduced into the macromolecular systems by Keith and Padden to interpret the formation of extinction banding in polymer spherulites as early as 1950s, but it proved unfounded and was quickly abandoned with the introducing of chain folding.¹⁵ Then, Brumberger observed concentric ringed spherulites that exhibited a wave-like topography in a poly-L-alanine solution film on slow evaporation, and ascribed their emergence to rhythmic crystallization.¹⁶ Based on simulation results, Kyu et al. suggested that ringed spherulites in poly(vinylidene fluoride) and poly(vinyl acetate) blends are attributed to rhythmic crystal growth.^17,18 This point is further supported by the alternating crystalline poly(aryl ether ketone) ridges and amorphous poly(aryl ether ether ketone) valleys that appeared in spherulites of their blends, and such rhythmic growth stems from a crystallization-caused phase separation process in which the amorphous component is rejected periodically from the growing crystal front.^19,20 Nonbirefringent bands of homopolymer can also be produced in polystyrene and poly(bisphenol A hexane ether) thin films.^10–12 It has revealed that such a periodic pattern comprises discrete stacks of flat-on lamellae, and its formation derives from a depletion-induced rhythmic crystallization. Similar ringed structures were also found in poly(lactic acid) (PLA) films, and the lamellae in the ridges changed from uniform flat-on to random orientations with an increase of thickness, leading to a transition of the nonbirefringent feature to a birefringent one.²¹ Furthermore, the birefringent banding showing a periodic change of thickness appeared in thin films of a six-arm star-shaped poly(ε-caprolactone) under subcritical CO₂.²² Recently, both birefringent and nonbirefringent ringed spherulites were produced in PLA amorphous films after solvent annealing.^23,24 It is obvious that rhythmic crystallization can take place in various polymer systems under quite different conditions, but most studies focus on analysing their microstructures. It is, therefore, of particular interest and importance to investigate the evolutionary regularity and environment dependences of these periodic structures in different systems and then to uncover the underlying origins for the occurrence of rhythmic crystallization behaviour.

In our previous research, we have reported the emergence of rhythmic-crystallization-caused concentric ringed spherulites in the evaporating solution films of poly(ε-caprolactone) (PCL), poly(ethylene adipate), and poly(ε-caprolactone-block-ethylene oxide) in a controllable confined environment, and proved that the birefringent properties of such spherulites can be tuned by modulating the radial lamellar organization.^25–27 We have also illustrated by in situ optical observations that the rhythmic crystallization in such evaporating polymer solution films originates from a periodic dimple generation and rupture driven by an evaporative-convection ahead of the growing front, and the spherulite growth follows a nonlinear feature.²⁸ These previous works focused on the crystal microstructures, growth dynamics and crystallization mechanism. However, several questions, such as the effects of the confined environment and pure solvent, the thickness dependence of growth rate and ring periodicity, and the key structural evolution of rhythmic crystallization, etc., still should be investigated to deepen our understanding of the rhythmic crystallization behaviour in such confined evaporating polymer solution films.

Herein, by following an analysis of the physical principles of evaporative crystallization, we first demonstrate the role of a confined environment on the generation of rhythmic crystallization by uncovering the special effects of some tuneable parameters. By slanting the substrate, the unique growth detail and thickness effect on both ring periodicity and growth rate in the same film are then illustrated to elucidate the development regularity of such rhythmic crystal growth. Finally, the structural evolution and micromorphology of the valley section, that is the key step of rhythmic crystallization, are also explored.

Experimental

Materials

The poly(ε-caprolactone) (PCL_10.0k, Aldrich Inc), with a number average molecular weight M_n = 10.0 kg mol⁻¹ and polydispersity index PDI = 1.4, and the poly(ε-caprolactone-b-ethylene oxide) diblock copolymer (PCL_24.5k-b-PEO_5.0k, Polymer Source Inc) with M^PCL_n = 24.5 kg mol⁻¹, M^PEO_n = 5.0 kg mol⁻¹ and PDI = 1.3, were purchased and used here as received. Both polymers were dissolved in toluene to prepare solutions for observations.

Confined evaporation environment and its manipulation

As depicted in Fig. 1, the confined evaporative crystallization of polymer solution films was conducted on silicon wafers inside the weighing bottle with the height and radius of 2.0 and 2.5 cm, respectively. The controllable environment can be achieved by adding extra solvent and then covering with a piece of glass. For further control, vaseline was also coated on the edge of the bottle to shrink the gap between the bottle and the covered glass. All the experiments were performed by casting a droplet of 10 μL solutions at room conditions (temperature 18–22 °C and relative humidity 30–50%). Except for detecting the effect of its amount, the added pure solvent is held at an invariant volume of 200 μL, and unless otherwise noted, the solution concentration is fixed at 10 mg mL⁻¹. Also, once comparison becomes involved, the crystallization conditions are strictly identical.

Fig. 1 Schematic representation of the controllable experiment apparatus for confined evaporative crystallization.

Observation and characterization

Optical microscopy (OM) observations of the crystal morphologies and the evolution process were conducted upon the Carl Zeiss A2m microscope equipped with a CCD camera under the reflectance mode, and the time interval for the in situ capture of each picture is within a range of 2 to 10 s. Atomic force microscope (AFM) (Molecular Imaging Inc., now Agilent 5500AFM/SPM System) was employed to analyse surface microstructures with a tapping mode using silicon cantilevers (Nanosensors, PPP-NCL).

Results and discussion

The physical principles for confined evaporative crystallization

It has been shown that rhythmic-crystallization-induced concentric ringed spherulites only appear upon slow evaporation.^25–28 To further understand such a unique crystallization behaviour, it is necessary to realize the physical principles of the evaporative crystallization in polymer solution films. It is known that two phase transitions, solvent evaporation and solute crystallization, take place during the whole process. For the situation at a fixed temperature, humidity, and atmospheric pressure, the driving force of solvent evaporation from solution can be expressed by ΔP = P_s − P, the pressure difference between the solvent saturation vapour pressure P_s of the solution and the solvent partial pressure P of the environment, and that the solute crystallization corresponds to the degree of supersaturation ΔC = C_ss − C_s, where C_ss and C_s are the concentrations of supersaturated and saturated solutions, respectively. It is evident that P_s decreases with the increase of solution concentration, and P relies on the confined conditions, while C_s holds constant, and C_ss is controlled by P. Therefore, P becomes the key controllable parameter. For free evaporation, P approaches zero because of the infinite atmosphere, while the case in the confined conditions becomes quite complex.

As illustrated in Fig. 1, the overall course can be described as that the solvent gradually escapes outside the bottle through the gap, giving rise to the successive concentration of solution and eventually crystallization of solute. It is evident that the process can be divided into two different stages. In the first stage, only solvent extraction and evaporation happen, and once crystal nucleation occurs, the system enters the second stage, in which solvent evaporation and solute crystallization develop simultaneously. Considering the existence of extra solvent, the situation becomes rather complex. On one hand, due to the finite and confined space and sufficient extra solvent, it is reasonable that, as the process proceeds, P inside the environment increases gradually until the balance is established between the solvent evaporated from the liquids and that escaped outside, leading to a nearly constant P within the system. One the other hand, since the saturation vapour pressure of solution P_s is smaller than that of pure solvent, only when P is also lower than the P_s can the extra solvent and solution extract simultaneously, so a P less than the saturation vapour pressure of the supersaturated solution P_ss is the prerequisite for the extra solvent to do work during the second stages.

The nature of the confined environment: unusual effects of several controllable parameters on rhythmic crystallization

Based on the above analysis, we now discuss the effects of several tuneable parameters to reveal the role of the confined environment upon rhythmic crystallization in evaporating polymer solution films. Herein, the results of PCL_24.5k-b-PEO_5.0k are employed to demonstrate these issues. We first pay attention to the whole confined environment. Fig. 2 exhibits POM photographs of the spherulite patterns developed under different confined times under the conditions, which can be achieved by opening the covered glass at a fixed time after casting. It is evident from Fig. 2a that the rhythmic crystallization cannot occur when completely exposed to air, i.e., P = 0, and a comparison of the spherulites in Fig. 2a with those in Fig. 2b implies that the nucleation does not occur even when the time elapsed is 25 hours, which can be attributed to the fact that the solution concentration is still below the extreme value. After keeping the confined condition for 28 hours and subsequently exposure to air, a twofold morphology in which compact spherulites surround concentric ringed spherulites is encountered (Fig. 2c), revealing that the rhythmic crystallization just happens under the confined evaporation situation. In other words, the covering lid ensures a suitable P for rhythmic crystallization. This point is again supported by the fact that the completely confined crystallization results in a single concentric ringed morphology (Fig. 2d). Meanwhile, the uniform structures further denote a rather stable situation during the whole process of evaporative crystallization during the second stage. Furthermore, a comparison of Fig. 2a and c indicates that the solvent withdrawing in the first stage mainly determines the nucleation density and thus the spherulite size, and that the second stage dominates the crystal evolution process and then the resulting morphological features.

Fig. 2 POM pictures of PCL_24.5k-b-PEO_5.0k spherulites obtained at varied confined times of (a) 0, (b) 25, (c) 28, and (d) 32 h. The average solvent evaporation rates (R_e) of the two situations, confined environment and free evaporation, are 2.50 × 10⁻⁴ and 6.92 × 10⁻² mL h⁻¹ respectively.

It is clear from Fig. 2 that the coupling of adding 200 μL pure solvent and covering with a lid provides a suitable and stable P for the whole rhythmic crystallization. We are thus interested in examining the influence of the amount of extra solvent. This point is unveiled in Fig. 3, where the solvent volume varies, but the other conditions are held unchanged. The major features of rhythmic crystallization no longer appear with the absence of extra solvent (Fig. 3a). However, the periodic ringed structures are found in the centre of spherulites by adding 50 μL solvent (Fig. 3b), and are spread along radial directions with increasing solvent volume (Fig. 3c and 2d). It has been shown that rhythmic crystallization in such evaporating polymer solution films stems from the fact that an evaporative convection that carries liquid to the growing front induces a periodic generation and rupture of the dimple.²⁸ Without extra pure solvent, a quite small amount of solvent in solution is apparently insufficient to construct an environment possessing a pressure large enough for the formation of evaporative convection and consequently rhythmic crystallization, so compact spherulites are evolved. Meanwhile, owing to the faster evaporation of the extra solvent than solution, the insufficient pure solvent volume cannot maintain the entire crystallization occurring under the appropriate conditions. Once the pure solvent is consumed completely, the subsequent growth again leads to continuous structures (Fig. 3b and c). It is now fully convincing that the amount of extra solvent determines the time for keeping a stable environment and so rhythmic crystal growth into a concentric ringed morphology. It is noted that the minimum amount of pure solvent for the occurrence of rhythmic crystallization of PCL_24.5k-b-PEO_5.0k in this confined environment is about 25 μL. We also mention that this amount depends on the crystalline polymer and the volume of container. For a PCL_10.0k solution confined in a smaller bottle, without pure solvent, the rhythmic growth can take place in the early stage of crystallization, resulting in a similar morphology with that in Fig. 3b.²⁹ The better mobility of PCL_10.0k leads to the easy occurrence of evaporative convection and the smaller size of bottle needs less solvent vapour to obtain a suitable pressure.

Fig. 3 POM images of PCL_24.5k-b-PEO_5.0k spherulites formed by adding different amounts of pure solvent of (a) 0, (b) 50, and (c) 100 μL. The R_es of the three cases are 2.8 × 10⁻³, 9.5 × 10⁻⁴ and 4.92 × 10⁻⁴ mL h⁻¹ respectively.

To proceed further in the analysis, it is expected that if the solvent partial pressure P in the confined environment is higher than the saturation vapour pressure of the supersaturated solution P_ss, the solute cannot crystallize until the extra solvent withdraws completely and then P gradually falls below P_ss. It is reasonable that P is controlled by the size and number of gaps, i.e., the smaller and the fewer gaps, the harder for the solvent to escape outside, so that a larger P results within the condition. An example is given to illustrate this point. Fig. 4a and b presents the OM pictures for the spherulite morphology developed under the same confined container but, before covering with the glass, the half-sides of the bottle are coated with vaseline to shrink the gap. We find that the complete evaporative crystallization in this case usually takes 4–5 days, dramatically slower than that in the common situation (ca. 32 hours, see Fig. 2d), supporting the enlargement of P. At first glance, the resulting spherulites also display a particular pattern that comprises a centric ringed region surrounded by continued textures, resembling that given in Fig. 3b and c. On closer inspection the differences are discernible. First, excepting the last several periods, the centre ringed zone shown in Fig. 3b and c is quite uniform, but those in Fig. 4 become poor rapidly. Clearly, the features of the last rings of Fig. 3b and c are the same as those of the all ringed area of Fig. 4b. Not unexpectedly, the uniform periodic structures can be attributed to the stable environment that is achieved and controlled by a simultaneous withdrawing of the extra solvent and solution. While without pure solvent, as the process proceeds, a decrease in the solvent amount in solution reduces the ability to supplement the vapour that escaped outside and thus a reduction of P occurs in this condition, which again accelerates the subsequent solvent extraction. The vicious cycle should be responsible for a rapidly poor and disappearing ring pattern. Second, the periodicity of the centric rings in Fig. 4a greatly enlarges in relation to those in Fig. 3b, c and 2d. The larger P is, the slower the evaporation occurs, so that the time for chain motion is longer, which leads to a decrease in crystallization rate and an increase of ring periodicity. It is evident that before the complete evaporation of pure solvent, P is unquestionably larger than the lower limit value for rhythmic crystallization, so the periodic morphology can be developed at the very beginning of the absence of pure solvent. Once P decreases rapidly to its lower bound, the centric ringed structures vanish, leading to a compact pattern. As further demonstrated in Fig. 4c, wider rings that appear in the centre of the spherulites observed in the case with the whole coating of the bottle perimeter again supports the conclusion that a larger P leads to an increase of ring periodicity.

Fig. 4 (a) POM and (b) OM photos of spherulites produced from a film of 10 mg mL⁻¹ PCL_24.5k-b-PEO_5.0k solution under the same confined environment, but the half-edge of the container was coated with vaseline to shrink the gap for solvent vapour escape. A quite slow R_e of 9.25 × 10⁻⁵ mL h⁻¹ was obtained in this case. (c) A POM image of crystals formed in the case that the whole perimeter of bottle was coated with vaseline, which is expected to further enlarge P and thus the ring periodicity.

It is now evident that the nature of the confined environment offers a suitable and steady solvent partial pressure for the occurrence of rhythmic crystallization in polymer solution films. Therein, the covered lid, and consequently the gap, determines an equilibrium pressure, and the extra solvent volume controls the duration time of the pressure inside the environment. In essence, such evaporative crystallization is similar to the dehydration of aqueous ascorbic acid solutions at varied humidity, which also causes the formation of ringed structures and the corresponding morphological transitions with altering humidity.^30,31

The morphological changes in an oblique solution film: special thickness impacts upon ring periodicity and growth rate

It is now recognized that such a confined environment provides the appropriate conditions for the generation of rhythmic crystal growth and thus the emergence of periodic ringed morphology. We are then interested in the impacts of the solution film itself on structural development and evolution. For a solution, the only variable is concentration, but for a film the key parameter is thickness. Clearly, the thickness of the resulting dried film thickens with the initial solution concentration for which enlarges the solute weight spreading per unit area. Undoubtedly, it is impossible to monitor simultaneously the structural evolution processes in different films, and fluctuations are unavoidable. To obtain a more accurate insight into these issues, we employ a slightly slanted substrate that is expected to induce a flow of solution and thus to result in a gradient variation of thickness. A slanted angle of about 1° is selected to do the experiment for larger angles drive the solution flow outside the substrate.

For the purpose of comparison and representativeness, an area including three different initial regions and associated growing crystals is selected, as shown in Fig. 5. It is evident from Fig. 5a that ringed structures that evolve within the regions A, B, and C exhibit the smallest, intermediate and largest ring periodicity, respectively, demonstrating that the ring periodicity is greatly dependent on the initial state prior to crystallization. This point is further supported by the fact that once crystal A grows into region B its periodicity enlarges immediately to the value approximately equal to that of crystal B (Fig. 5b). The average ring periodicities of crystals A, B, and C are ca. 7.1, 14.0, and 24.1 μm, respectively, reflecting a big span of rhythmic length in the same film. Within an identical timescale and in the same film, the crystals certainly progress at completely unvaried conditions, so such a regional dependence of periodicity is clearly indicative of a thickness effect. This point is illustrated by AFM observations. Representative AFM height images and corresponding height profiles for three ringed crystals that have similar periodicity with that in Fig. 5 from the resulting film are exhibited in Fig. 6. The average periodicities are about 7.9, 13.6, and 21.8 μm, and the corresponding height differences between the ridges and valleys are about 200, 470, 800 nm respectively, implicating a positive correlation between the two parameters. The data for nine periodicities in the same film and two in other films are displayed as Fig. 7. Considering the deviation, the histogram appears to support that there is a direct proportion between ring periodicity and height difference, especially in the same film. Owing to quite similar morphology, it is reasonable that the height difference is a reflection of the film thickness. It is, therefore, concluded that the ring periodicity stemming from rhythmic crystal growth is proportional to the film thickness in the evaporating polymer solution film under unvaried conditions. This conclusion can be further supported by the observation of the spherulites that show a gradual increase of ring width along the radial direction in this slanted substrate, as displayed in Fig. 8.

Fig. 5 OM pictures exhibiting the in situ evolution process (a–b) and the resulting morphology (c) of three ringed crystals in different regions of a film of 10 mg mL⁻¹ PCL_10.0k solution upon a slightly slanted substrate (a degree of slope about 1° for a larger slant angle causes the solution to flow outside the substrate). The inset in (c) is the corresponding POM photograph. The plots of radius with time for the indicated distance within an identical timescale of the three crystals are demonstrated in Fig. 9.

Fig. 6 Representative AFM height images and the corresponding height profiles along the indicated radial distances implicating the interaction between the height difference and ring periodicity. The three patterns are all collected from the resulting film of Fig. 5 with the average periodicities of about (a) 7.9, (b) 13.6, and (c) 21.8 μm respectively.

Fig. 7 The variation of average height difference between ridges and valleys against the average ring periodicity of concentric ringed spherulites. The first nine values are obtained from the same film as Fig. 6, and the latter two are gained from other films.

Fig. 8 The gradual increase of ring width of a spherulite formed in the slant substrate further indicates a proportional relationship between the ring periodicity and the film thickness.

It is then of particular interest to detect the radial rates of the crystals developed in different regions to uncover the influence of thickness on crystal growth. This point is illustrated in Fig. 9, where the radial growth distance is plotted against time for the three crystals within an identical timescale, as denoted in Fig. 5. Taking account of a nonlinear growth within one period,²⁸ here the step length is selected as each ring periodicity. The major feature of linear growth is again observed for the three crystals. Meanwhile, the average radial growth rates for the periodicities of 7.1, 14.0, and 24.1 μm are 199, 182, 173 nm s⁻¹, respectively, indicating a slight decrease in the growth rate with increasing film thickness. This point can be further supported by the decrease in the slope along the growth curve of crystal A with enlarging periodicity in region C (Fig. 9). This effect is not only different from the data collected from solution films formed from different initial solution concentrations, but also different from the conclusions gained in the melt crystallization. It has been shown that the average ring periodicities of solution films with the initial concentrations of 5 and 10 mg mL⁻¹ are ca. 6 and 17 μm and the corresponding average growth rates are 266 and 181 nm s⁻¹, respectively.²⁸ Apparently, the ringed crystals that have an approximate periodicity in the films resulting from different initial concentrations display a larger difference in growth rate, but those exhibiting very different periodicities in the same film present a smaller divergence of growth rate. The result indicates that it is the initial concentration rather than the thickness that is more important for determining the crystal radial growth rate. This point is quite understandable when concerning evaporative crystallization behaviour. Under the same confined evaporation environment, it can be assumed that solute nucleation occurs under the same supersaturation, and that the overall evaporation rate is nearly constant. Hence, the larger the initial concentration is, the more solvent remains, consequently the longer the time for complete crystallization to take and place, leading to a slower crystal growth rate. For crystals developing in the same film but within different regions, the overall crystallization rate is determined by the amount of residual solvent, but the thinner zone has less solvent and thus the larger growth rate. Although the solvent evaporation and solute crystallization are mutually reinforced processes in evaporative crystallization, the above behaviour further indicates that the solute crystallization in such circumstance is controlled by solvent extraction. Moreover, due to the presence of solvent, the mobility of polymer chains to a growing face is easy and thus cannot become the determining step for crystal growth, even in quite thin films. Meanwhile, a thicker film means that the more solvent needs to escape. A coupling of the two aspects induces an opposite dependence of the growth rate upon film thickness in such an evaporative crystallization compared with that in the common melt crystallization.

Fig. 9 Temporal changes of radius against time for the denoted distance (Fig. 5b) within the same timescale of the three crystals reflecting the unique dependence of radial growth rate upon ring periodicity in this film. Taking account of the nonlinear growth within one period, here the length of each step is selected as the corresponding ring periodicity, and for convenient comparison, the later time axis is right shift 100 s relative to the former one.

Finally, it is should be noted that a comparison of OM and POM pictures demonstrates that the birefringence is enforced with increasing thickness (Fig. 5c and its inset), which can be attributed to the transition of lamellae from uniform flat-on to tilted and edge-on orientations. This point is similar to those in melt crystallization, and has been explained previously.²¹

The structural development within one period: the melt-like growth during the evolution process of valley section

Thus far, the impacts of both external environment and internal solution film upon overall structural development and transition are elucidated. We now emphasize the details for the structural evolution within a period to further deepen the understanding of rhythmic crystallization in evaporating polymer solution films.

It has verified that possessing an evaporative convection that carries liquid towards growing front drives the occurrence and rupture of a dimple that gives rise to the occurrence of rhythmic crystallization.²⁸ In other words, the key step for such rhythmic growth is the generation of valleys. The periodic evolution of growing front is again depicted in Fig. 10 by in situ OM photos, and a comparison of Fig. 10c and d also displays that before the emergence of next period, there is a period of time, ca. 160 s, in which no obvious radial growth is discernible. Actually, this induction time is a common existence is such rhythmic crystal growth process.²⁸ So what causes this phenomenon and what happens during this period of time?

Fig. 10 In situ POM photos revealing the evolution of growing front within one period and the induction time before the growth of next period in a solution film casting from 40 mg mL⁻¹ solution, (a) the evolution of the ridge, (b) the beginning of the valley, the coupling of (c) and (d) indicating the induction time.

As unveiled previously,²⁸ the rhythmic crystallization in such solution evaporation can be described by the schematic given in Fig. 11. The material depletion near the crystallization face induces the occurrence of a dimple ahead of the growing crystal front. That is, the solvent evaporates from a curved surface. As can be seen in Fig. 11, in the evolution of each valley, the combination of an enhanced withdrawal on the convex surface and a slowed extraction on the concave one leads to a drainage flow that carries material to the two opposite sides, which causes a rapid rupture of the dimple. Hence, it can be expected that just a few polymer chains are left in the valley section, which may be absorbed by substrate. This point can be confirmed by AFM results, as represented in Fig. 12. The combination of height, phase and amplitude images (Fig. 12a–c) reveals that only a layer of lamellar crystal appears in the thinnest section. Moreover, the existence of the impurity induced dewetting results in the direct observation of substrate, as denoted by white arrows. A similar observation in another case (Fig. 12d) again supports the common nature of the phenomenon. Meanwhile, due to the fast evaporation of two sides on convex surfaces and the better mobility of solvent, it is reasonable that the solvent is absent in these few chains when they crystallize. In other words, for these absorbed chains in the valley zone, a melt-like ultrathin film growth takes place, which cannot be spotted by an OM microscope, resulting in the emergence of an induction time. Furthermore, the dislocation generation kinetics that control the slight thickening process at the beginning of upward section also play a small role in the disability of the observation of an obvious radial extension during the induction time.¹⁰

Fig. 11 Schematic illustration unveiling the evaporation and flow mechanism during the formation process of valley. The length of dash arrows denotes the relative rate of evaporating vapour flux that leaves at the convex surface with different curvatures.

Fig. 12 AFM (a) height, (b) amplitude, and (c) phase micrographs depicting the microstructures of a valley in a concentric ringed spherulite evolved from a 40 mg mL⁻¹ solution film, which reveals the structural evolution in the induction time. (d) An image of a spherulite formed from a 20 mg mL⁻¹ solution film exhibiting similar features supports the generality of such phenomenon.

Conclusions

In summary, we have investigated rhythmic crystallization behaviour in confined evaporating polymer solution films from three aspects, the confined environment, the solution film, and the structural development of one period, to further enhance our understanding of such a unique crystallization phenomenon. It is demonstrated that the coupling of a confined condition and extra pure solvent sustains a suitable and stable solvent partial pressure that ensures the occurrence and continuation of such rhythmic crystal growth, leading to the emergence of concentric ringed spherulites with the nearly unvaried periodicity. Slanting the substrate resulted in a flow of solution and thus a gradual change of film thickness, which provides a more accurate way to detect the effects of the solution film in a quite unchanged situation. It is shown that the ring periodicity is approximately proportional to the film thickness in the same film, but it is the initial solution concentration rather than film thickness that is more important for determining the radial growth rate, again supporting the fact of drying-driven crystallization behaviour. In the key step, the valley formation of rhythmic crystallization, there is an induction time, which can be attributed to a coupling of a drainage flow under the concave surface and subsequently a melt-like growth of the quite few lamellae close to substrate in the valley section. These interesting findings provide several significant insights into the rhythmic crystallization behaviour in confined evaporating polymer solution films.

Acknowledgements

The authors acknowledge financial support from the National Natural Science Foundation of China (21404113, 21274148), China Postdoctoral Science Foundation (2013M541801), and Ningbo Natural Science Foundation (2014A610134).

Notes and references

1. B. Lotz and S. Z. D. Cheng, Polymer, 2005, 46, 577–610 , and its references.
2. H. M. Ye, J. Xu, B. H. Guo and T. Iwata, Macromolecules, 2009, 42, 694–701.
3. H. M. Ye, J. S. Wang, S. Tang, J. Xu, X. Q. Feng, B. H. Guo, X. M. Xie, J. J. Zhou, L. Li, Q. Wu and G. Q. Chen, Macromolecules, 2010, 43, 5762–5770.
4. Z. B. Wang, Y. Li, J. Yang, Q. T. Gou, Y. Wu, X. D. Wu, P. B. Liu and Q. Gu, Macromolecules, 2010, 43, 4441–4444.
5. J. Li, Y. Li, J. Zhou, J. Yang, Z. Q. Jiang, P. Chen, Y. Wang, Q. Gu and Z. B. Wang, Macromolecules, 2011, 44, 2918–2925.
6. M. Rosenthal, G. Bar, M. Burghammer and D. A. Ivanov, Angew. Chem., Int. Ed., 2011, 50, 8881–8885.
7. M. Rosenthal, G. Portale, M. Burghammer, G. Bar, E. T. Samulski and D. A. Ivanov, Macromolecules, 2012, 45, 7454–7460.
8. M. Rosenthal, M. Burghammer, G. Bar, E. T. Samulski and D. A. Ivanov, Macromolecules, 2014, 47, 8295–8304.
9. Y. Q. Zhang, H. G. Fang, Z. K. Wang, M. Tang and Z. G. Wang, CrystEngComm, 2014, 16, 1026–1037.
10. Y. X. Duan, Y. Jiang, S. D. Jiang, L. Li, S. K. Yan and J. M. Schultz, Macromolecules, 2004, 37, 9283–9286.
11. Y. X. Duan, Y. Zhang, S. K. Yan and J. A. Schultz, Polymer, 2005, 46, 9015–9021.
12. Y. Wang, C. M. Chan, L. Li and K. M. Ng, Langmuir, 2006, 22, 7384–7390.
13. K. Kawashima, R. Kawano, T. Miyagi, S. Umemoto and N. Okui, J. Macromol. Sci., Part B: Phys., 2003, 42, 889–899.
14. H. K. Henisch, Crystals in Gels and Liesegang Rings, Cambridge University Press, Cambridge, 1988.
15. H. D. Keith and F. J. Padden, J. Polym. Sci., 1958, 31, 415–421.
16. H. Brumberg, Nature, 1970, 227, 490–491.
17. Y. Okabe, T. Kyu, H. Saito and T. Inoue, Macromolecules, 1998, 31, 5823–5829.
18. T. Kyu, H. W. Chiu, A. J. Guenthner, Y. Okabe, H. Saito and T. Inoue, Phys. Rev. Lett., 1999, 83, 2749–2752.
19. J. Chen and D. C. Yang, Macromol. Rapid Commun., 2004, 25, 1425–1428.
20. J. Chen and D. C. Yang, Macromolecules, 2005, 38, 3371–3379.
21. S. Nurkhamidah and E. M. Woo, Macromol. Chem. Phys., 2013, 214, 673–680.
22. Y. F. Zhang, L. Liao, X. L. Luo, S. L. Liu, Q. Yang and G. X. Li, RSC Adv., 2014, 4, 10144–10150.
23. S. Y. Huang, H. F. Li, Y. R. Shang, D. H. Yu, G. Li, S. C. Jiang, X. S. Chen and L. J. An, RSC Adv., 2013, 4, 13705–13711.
24. S. Y. Huang, H. F. Li, H. Y. Wen, D. H. Yu, S. C. Jiang, G. Li, X. S. Chen and L. J. An, CrystEngComm, 2014, 16, 94–101.
25. Z. B. Wang, G. C. Alfonso, Z. J. Hu, J. D. Zhang and T. B. He, Macromolecules, 2008, 41, 7584–7595.
26. Y. G. Li, H. Y. Huang, T. B. He and Z. B. Wang, ACS Macro Lett., 2012, 1, 154–158.
27. Y. G. Li, H. Y. Huang, Z. B. Wang and T. B. He, Macromolecules, 2014, 47, 1783–1792.
28. Y. G. Li, H. Y. Huang, T. B. He and Z. B. Wang, Polymer, 2013, 54, 6628–6635.
29. Y. G. Li, H. Y. Huang, T. B. He and Z. B. Wang, ACS Macro Lett., 2012, 1, 718–722.
30. A. S. Paranjpe, Phys. Rev. Lett., 2002, 89, 075504.
31. H. Uesaka and R. Kpbayashi, J. Cryst. Growth, 2002, 237–239, 132–137.

---