When a polymer melt crystallizes under strong flow (extrusion, injection molding, fiber spinning), the classic 'shish-kebab' morphology arises: a central fiber-like crystal core (the shish, from the Japanese word for skewer) nucleated by flow, onto which chain-folded lamellae (kebabs) grow radially. Kimata et al. (Science 2007) used deuterium-labeled isotactic polypropylene and small-angle neutron scattering to overturn the textbook view that the longest chains physically constitute the shish: long chains are not overrepresented in it, but act catalytically, recruiting shorter neighbors into shish formation. The post-2007 literature leaves the mechanistic core open: the molecular-weight composition of the shish is condition-dependent and contested between experiments; the catalytic recruitment picture has support only from indirect scattering, kinetics, and molecular-dynamics simulation, never from direct molecular-scale experimental observation; deformation-versus-concentration interpretations of the labeling-dependent scattering anisotropy remain unresolved; and no predictive quantitative model couples long-chain stretching kinetics and flow-induced orientation to shish propagation velocity.

Background

In semicrystalline polymers, crystallization usually proceeds by chain-folded lamellae. Under sufficiently strong flow (shear or extension), the melt develops oriented, fiber-like crystal precursors that act as nuclei for radially growing lamellae, producing the 'shish-kebab' morphology — 'shish' (skewer) for the extended-chain fibrillar core and 'kebab' for the lamellae threaded on it — which underlies the highly oriented, self-reinforced 'skin' layers of injection-molded and spun polyolefin products. The central mechanistic question of flow-induced crystallization (FIC) has long been how the melt's molecular-weight distribution (MWD) maps onto the shish: textbook accounts based on the coil-stretch transition (the unfolding of a chain from a random coil to a nearly fully extended conformation when flow overcomes chain relaxation) held that the longest, most easily extended chains dominate the shish. Testing this requires distinguishing chain lengths inside a solidified sample, which is done by small-angle neutron scattering (SANS, scattering at small angles from nanometer-scale structure) on blends in which a chosen fraction of chains is deuterated: deuteration changes the neutron scattering length without materially changing the chemistry, so the scattering anisotropy of the labeled subpopulation reports how that chain-length class is distributed and deformed within the shish-kebab morphology. Kimata et al. (Science 316, 1014, 2007) labeled successive thirds of the MWD of isotactic polypropylene (iPP; the stereoregular form of polypropylene) and showed that long chains are not enriched in the shish relative to their melt concentration; instead, the longest chains play a catalytic role, apparently nucleating and recruiting neighboring chains of any length into shish formation, consistent with the strong acceleration of oriented crystallization seen when even ~2 wt% ultrahigh-molecular-weight chains are added to a short-chain matrix (Fernandez-Ballester et al., 2012, who distinguish thresholds for shish initiation versus propagation). However, a same-month neutron-scattering study of polyethylene (PE) blends found the opposite composition — ultrahigh-molecular-weight components dominantly included in the shish (Kanaya et al., 2007) — and later wide-spatial-scale X-ray/neutron work found that shish composition and core-shell structure vary with molecular weight, flow type, flow rate, temperature, and polymer (Kanaya et al., 2017; Sakurai et al., 2019). In situ studies added further constraints without settling the mechanism: real-space optical microscopy of sheared iPP shows shish forming through discrete multi-step pathways via a transition state (Shen et al., 2013), and in situ SANS with small-angle X-ray scattering (SAXS, the X-ray analogue probing crystalline-electron-density contrast) on lightly cross-linked deuterated/hydrogenated PE shows shish forming at chain deformation ratios as small as about 1.3, contradicting both the coil-stretch picture and simple stretched-network models, while revealing periodic concentration modulation of the deuterated chains (Yang et al., 2016). On the theoretical side, stretch-based and thermodynamic nucleation models culminate in the polySTRAND model (Read et al., 2020), which predicts overrepresentation of long chains inside pre-critical flow-induced nuclei and superexponential nucleation-rate growth with shear rate, and which explicitly reconciles that enrichment with Kimata's non-enriched shish ('our model predicts enrichment only in pre-critical nuclei. Post-nucleation growth is thermodynamically favourable so recruits all chains equally, giving the melt distribution in well-developed crystals') while stating the open direction as 'To quantitatively understand shish formation, we require a detailed understanding of polydispersity in point-like nucleation'; united-atom molecular-dynamics (MD) simulations (in which each chemical group such as CH2 is treated as a single interaction site) of bimodal PE blends directly observe this long-chain enrichment in flow-induced nuclei (Anwar & Graham, 2021) and reproduce the 'catalytic effect' — oriented long-chain segments promoting the deformation, orientation, and co-nucleation of nearby short chains through bond-bond interaction energetics and the entanglement network (Wen et al., 2024; reviewed by Nie et al., 2024 and Graham, 2019). Yet all molecular-scale pictures of the recruitment process remain computational: flow-induced nucleation occurs on very short time and length scales that, in the words of the simulation literature (Anwar & Graham, 2021), 'cannot be directly observed with current experimental techniques.' The unresolved scientific objective is therefore a molecularly resolved, experimentally grounded account of what the shish is made of and how flow and long chains create it, together with the quantitative coupling of that mechanism to shish nucleation and growth — the open issues left by the 2007 Science paper and explicitly framed as still-unsettled in the 2017 follow-up ('there are various mechanisms for shish-kebab formation that depend on the factors such as the type of flow fields, flow rate, type of polymer, temperature, molecular weight and molecular weight distribution') and in current reviews (Li & Fu, 2026).

Problem Statement

The open problem left by Kimata et al. (2007) is to establish how flow and the longest polymer chains jointly create the shish in flow-induced crystallization of polymer melts: a molecularly resolved, experimentally grounded mechanism of shish formation. Here 'shish' means the mature extended-chain fibrillar core of the solidified shish-kebab structure — the object whose molecular-weight composition was measured by scattering in Kimata et al. (2007) and Kanaya et al. (2007, 2017) — as distinct from the pre-critical flow-induced nuclei in which polySTRAND and MD find long-chain enrichment (Read et al., 2020, reconcile the two: enrichment is predicted only in pre-critical nuclei, while post-nucleation growth is thermodynamically favorable and recruits all chains equally, giving the melt distribution in well-developed crystals), and as distinct from earlier oriented precursors ('strings', 'sausages', densified oriented domains), whose operational separation from the shish is itself part of the difficulty. Open facets of this single mechanistic question are: (1) Molecular mechanism — resolve at molecular scale how — or whether — the longest chains promote and recruit neighboring chains of any length into shish formation, and the local segmental-orientation field around a growing shish; no direct molecular-scale experimental observation of the recruitment event exists, the present evidence being scattering-inferred, kinetic, or simulation-based. (2) Shish composition and its condition dependence — determine which chain-length classes are incorporated into the mature shish, and account for the standing contradiction among the shish-composition experiments: long-chain-dominant shish in drawn deuterated/hydrogenated PE (Kanaya et al., 2007), composition-neutral shish in sheared iPP (Kimata et al., 2007), and low-molecular-weight-dominant shish in drawn d/h PE (Kanaya et al., 2017); within this, establish whether long chains act catalytically, constitutively, or both depending on conditions. The measured dependence on flow field, flow rate, polymer type, temperature, molecular weight, and molecular-weight distribution is a feature the mechanism must explain, and the existing evidence base is system- and condition-specific (sheared iPP; drawn d/h PE blends at particular drawing rates). (3) Deformation versus concentration — separate enhanced molecular deformation of labeled chains from genuine concentration heterogeneity (segregation) as the origin of the labeling-dependent SANS anisotropy in deuterated blends, the ambiguity made explicit by Yang et al. (2016) on drawn, lightly cross-linked d/h PE, where shish forms at chain deformation ratios of only about 1.3 while a periodic concentration modulation of the deuterated component appears. (4) Quantitative coupling — connect the nucleation-level molecular understanding to shish formation quantitatively, so that chain stretching kinetics, flow-induced orientation, and crystallization probability together account for measured shish nucleation and propagation velocities from molecular parameters (long-chain length and concentration, MWD, flow strength, temperature); as Read et al. (2020) put it, 'To quantitatively understand shish formation, we require a detailed understanding of polydispersity in point-like nucleation.' These facets are aspects of one mechanistic question rather than separately mandatory deliverables: a resolution must be experimentally grounded — meaning consistent with the cited experiments, i.e. the existing scattering and kinetic data, where a proposed new experiment, simulation, or model may support but not substitute for that consistency — and must address all four facets within that single account, covering both the molecular mechanism — what the shish is made of and how flow and long chains create it (facets (1)–(3)) — and the quantitative coupling of that mechanism to shish nucleation and growth (facet (4)); no single facet alone resolves it. None of the facets has been closed by later literature: how flow and long chains create the shish remains open.

References

  1. Shuichi Kimata, Takashi Sakurai, Yoshinobu Nozue, Tatsuya Kasahara, Noboru Yamaguchi, Takeshi Karino, Mitsuhiro Shibayama, Julia A. Kornfield, 'Molecular Basis of the Shish-Kebab Morphology in Polymer Crystallization', Science 316(5827):1014-1017 (2007). DOI: 10.1126/science.1140132, https://doi.org/10.1126/science.1140132
  2. Toshiji Kanaya, Go Matsuba, Yoshiko Ogino, Koji Nishida, Hirohiko M. Shimizu, Takenao Shinohara, Takayuki Oku, Junichi Suzuki, Toshiya Otomo, 'Hierarchic Structure of Shish-Kebab by Neutron Scattering in a Wide Q Range', Macromolecules 40(10):3650-3654 (2007). DOI: 10.1021/ma062606z, https://doi.org/10.1021/ma062606z
  3. Hiroki Murase, Yasuo Ohta, Takeji Hashimoto, 'A New Scenario of Shish-Kebab Formation from Homogeneous Solutions of Entangled Polymers: Visualization of Structure Evolution along the Fiber Spinning Line', Macromolecules 44(18):7335-7350 (2011). DOI: 10.1021/ma2008817, https://doi.org/10.1021/ma2008817
  4. Lucia Fernandez-Ballester, Derek W. Thurman, Weijun Zhou, Julia A. Kornfield, 'Effect of Long Chains on the Threshold Stresses for Flow-Induced Crystallization in iPP: Shish Kebabs vs Sausages', Macromolecules 45(16):6557-6570 (2012). DOI: 10.1021/ma3000384, https://doi.org/10.1021/ma3000384
  5. Bo Shen, Yongri Liang, Julia A. Kornfield, Charles C. Han, 'Mechanism for Shish Formation under Shear Flow: An Interpretation from an in Situ Morphological Study', Macromolecules 46(4):1528-1542 (2013). DOI: 10.1021/ma3023958, https://doi.org/10.1021/ma3023958
  6. Haoran Yang, Dong Liu, Jianzhu Ju, Jing Li, Zhen Wang, Guanyun Yan, Youxin Ji, Wenhua Zhang, Guangai Sun, Liangbin Li, 'Chain Deformation on the Formation of Shish Nuclei under Extension Flow: An in Situ SANS and SAXS Study', Macromolecules 49(23):9080-9088 (2016). DOI: 10.1021/acs.macromol.6b01945, https://doi.org/10.1021/acs.macromol.6b01945
  7. Zhen Wang, Zhe Ma, Liangbin Li, 'Flow-Induced Crystallization of Polymers: Molecular and Thermodynamic Considerations', Macromolecules 49(5):1505-1517 (2016). DOI: 10.1021/acs.macromol.5b02688, https://doi.org/10.1021/acs.macromol.5b02688
  8. Toshiji Kanaya, Momoko Murakami, Tadahiko Maede, Hiroki Ogawa, Rintaro Inoue, Koji Nishida, Go Matsuba, Noboru Ohta, Shin-ichi Takata, Taiki Tominaga, Jun-ichi Suzuki, Young-Soo Han, Tae-Hwan Kim, 'Role of molecular weight in shish-kebab formation during drawing by small-angle neutron and X-ray scattering', Polymer Journal 49(12):831-837 (2017). DOI: 10.1038/pj.2017.65, https://doi.org/10.1038/pj.2017.65
  9. Richard S. Graham, 'Understanding flow-induced crystallization in polymers: A perspective on the role of molecular simulations', Journal of Rheology 63(1):203-214 (2019). DOI: 10.1122/1.5056170, https://doi.org/10.1122/1.5056170
  10. Masayuki Sakurai, Go Matsuba, Keisuke Sotoyama, Koji Nishida, Toshiji Kanaya, Shin-ichi Takata, 'Molecular weight component dependence of shish-kebab structure of polyethylene blends with X-ray and neutron scattering measurements covering a wide spatial scale', Polymer Crystallization 2(1) (2019). DOI: 10.1002/pcr2.10034, https://doi.org/10.1002/pcr2.10034
  11. Daniel J. Read, Claire McIlroy, Chinmay Das, Oliver G. Harlen, Richard S. Graham,, 'PolySTRAND Model of Flow-Induced Nucleation in Polymers', Physical Review Letters 124(14):147802 (2020). DOI: 10.1103/PhysRevLett.124.147802, https://doi.org/10.1103/PhysRevLett.124.147802
  12. Muhammad Anwar, Richard S. Graham, 'Direct observation of long chain enrichment in flow-induced nuclei from molecular dynamics simulations of bimodal blends', Soft Matter 17(10):2872-2882 (2021). DOI: 10.1039/D0SM01361G, https://doi.org/10.1039/D0SM01361G
  13. Jianlong Wen, Jun Yang, Martin Kroeger, Mingyang Ma, Tongfan Hao, Zhiping Zhou, Yijing Nie, 'Molecular Details of the "Catalytic Effect" of Long Chains on Short Chains in Stretch-Induced Polymer Crystallization', Macromolecules 57(4):1612-1624 (2024). DOI: 10.1021/acs.macromol.3c01541, https://doi.org/10.1021/acs.macromol.3c01541
  14. Yijing Nie, Jianlong Wen, Yongqiang Ming, Yuhang Zhou, 'Mechanisms of Flow-Induced Multistep Structure Evolution in Semicrystalline Polymers Revealed by Molecular Simulations', Crystal Growth & Design 24(15):6531-6547 (2024). DOI: 10.1021/acs.cgd.4c00592, https://doi.org/10.1021/acs.cgd.4c00592
  15. Runlai Li, Qiang Fu, 'Shish-kebab as a self-composite in polyethylene: From Keller's crystallization framework to processing-enabled hierarchy engineering', Polymer 358:130213 (2026). DOI: 10.1016/j.polymer.2026.130213, https://doi.org/10.1016/j.polymer.2026.130213