Out-of-Plane Polymer Coupling Interfaces

• Out-of-plane polymer coupling interfaces are defined by directional transport normal to the polymer plane, affecting symmetry breaking and the flow of electrons, spins, light, or heat.
• These interfaces enable spontaneous geometric distortions, enhanced spin pumping, and engineered optical couplers by exploiting the unique interfacial properties in polymers.
• Tailoring these interfaces—including structural deformation and targeted functionalization—redirects energy from weak out-of-plane channels into more efficient in-plane pathways for improved device performance.

Out-of-plane polymer coupling interfaces comprise a class of interfacial phenomena in which the decisive structural displacement, angular-momentum transfer, optical routing, or heat-transfer bottleneck is oriented normal to a polymer plane, backbone, or substrate-parallel transport axis. Recent work uses the concept in several technically distinct senses: spontaneous out-of-plane backbone distortion in a $\pi$-conjugated polymer, perpendicular spin pumping across a polymer/ferromagnet interface, three-dimensional polymer couplers that rotate guided light by $90^\circ$ relative to the substrate, and polymer–graphene thermal coupling problems in which weak out-of-plane transport limits performance unless the interface is redirected into more favorable channels (Wang et al., 20 Aug 2025, Gaur et al., 28 Mar 2025, Landowski et al., 2017, Muthaiah et al., 2022).

1. Scope and definitions

In structural and electronic chemistry, an out-of-plane polymer coupling interface denotes a situation in which the polymer’s electronic state couples to a backbone displacement normal to its nominal molecular plane. In spintronics, it denotes angular-momentum transfer across a polymer/ferromagnet boundary perpendicular to the interface plane. In integrated photonics, it denotes a coupler that turns a guided optical mode out of the substrate plane. In thermal nanocomposites, the same language arises more indirectly: the central issue is that polymer bonding to a filler may either force transport through weak out-of-plane channels or instead couple into the filler’s strong in-plane pathways (Wang et al., 20 Aug 2025, Gaur et al., 28 Mar 2025, Landowski et al., 2017, Muthaiah et al., 2022).

These usages are not interchangeable, but they share a common interfacial logic. The interface is not merely a geometric boundary; it is the locus at which a symmetry constraint, transport direction, or orbital selection rule determines whether coupling is suppressed, enhanced, or qualitatively redirected. This suggests that “out-of-plane” is best understood as a directional descriptor whose physical meaning depends on the observable under consideration: lattice distortion in conjugated polymers, spin or orbital current in ferromagnetic resonance, optical propagation in waveguide architectures, or phonon injection into anisotropic fillers.

A recurring misconception is that out-of-plane behavior is necessarily extrinsic, arising from roughness, substrate forcing, or fabrication artifacts. The literature considered here repeatedly argues otherwise. In PDFHE, the distortion is interpreted as an intrinsic spontaneous symmetry breaking rather than a substrate-imposed deformation (Wang et al., 20 Aug 2025). In HSQ/NiFe, increasing damping is retained in reversed stacks, which is used to argue against sputter-damage artifacts and in favor of a genuine polymer-mediated coupling channel (Gaur et al., 28 Mar 2025). In polymer waveguides, out-of-plane access is designed into the coupler geometry rather than produced by accidental scattering (Landowski et al., 2017).

2. Spontaneous symmetry breaking in a conjugated polymer backbone

A particularly direct realization of an out-of-plane polymer coupling interface is provided by poly-(difluorenoheptalene-ethynylene) (PDFHE), a one-dimensional, extended-width, $\pi$-conjugated polymer synthesized on noble-metal surfaces and characterized by low-temperature scanning tunneling microscopy. PDFHE is built from difluorenoheptalene (DFH) units linked by ethynylene bridges, and the underlying design motif is a competition between an open-shell, localized diradical form with a triple-bond linker and a closed-shell cumulene-like form with $\mathrm{=C=C=}$-type bonding. In the latter geometry, the linker bonding angle becomes strained by the heptagonal environment, so the system can lower strain by moving atoms out of the molecular plane; the result is a strong coupling between geometry and electronic structure (Wang et al., 20 Aug 2025).

Experimentally, PDFHE does not remain planar. Low-temperature STM at 4.7 K and CO-tip bond-resolved STM revealed long covalently linked one-dimensional polymer chains on Au(111), with BRSTM showing a periodic segmented backbone and high-resolution topographs showing nonplanar DFH cores. Each unit has a trapezoidal appearance, the backbone exhibits periodic corrugation, the units are tilted out of the $xy$-plane, and two internal mirror symmetries of the unit cell are broken. The reported out-of-plane modulation is $\Delta z = 17 \pm 3\ \text{pm}$ with repeat distance $\lambda = 0.95 \pm 0.10\ \text{nm}$, identified experimentally with the all-trans conformation. Rare cis defects appear as localized units with rectangular symmetry flanked by opposite tilts and occur in only about $10\%$ of polymers on both Au(111) and Ag(111) (Wang et al., 20 Aug 2025).

The proposed mechanism is a Jahn-Teller-like instability expressed as spontaneous symmetry breaking of the mirror-symmetric planar geometry. The planar polymer is described as a high-energy saddle point, whereas both distorted forms are local minima. Density functional theory gives relative energies of $-883\ \text{meV/unit\ cell}$ for all-trans PDFHE and $-651\ \text{meV/unit\ cell}$ for all-cis PDFHE, each relative to the planar structure. The same calculations show a systematic bandgap increase as symmetry is lowered: about $90^\circ$0 for the planar form, about $90^\circ$1 for all-trans, and about $90^\circ$2 for all-cis. STS on all-trans PDFHE on Au(111) gives an apparent bandgap of $90^\circ$3, which is attributed to the usual PBE underestimation of gaps and to substrate effects not included in the freestanding calculation (Wang et al., 20 Aug 2025).

The significance of PDFHE lies in its contradiction of a common expectation for mechanically robust extended $90^\circ$4-systems. Such systems are often assumed to be too rigid for strong out-of-plane symmetry breaking, yet PDFHE shows that electron-lattice coupling can still dominate if a non-benzenoid, strained, partially open-shell building block makes the high-symmetry geometry electronically unstable. The existence of two low-energy closed-shell conformations, trans and cis, further suggests bistability with distinct electronic structures. A plausible implication is that polymer interfaces of this type may be engineered not only for static structure control but also for switchable geometry-electronic coupling.

3. Perpendicular angular-momentum transfer at polymer/ferromagnet interfaces

In spintronics, the phrase denotes a different but equally specific geometry: angular momentum is pumped perpendicular to the interface plane from a precessing ferromagnet into an adjacent polymer layer. A representative case is the HSQ/NiFe heterostructure, in which hydrogen silsesquioxane (HSQ, $90^\circ$5) oligomer layers with thicknesses of $90^\circ$6, $90^\circ$7, and $90^\circ$8 are combined with sputtered NiFe films of $90^\circ$9, $\pi$0, $\pi$1, and $\pi$2. HSQ is emphasized as a silicon-based inorganic polymer that is insulating and low-spin-orbit-coupled, yet planarizing and compatible with standard microfabrication, making it both technologically relevant and conceptually nontrivial as a spin sink (Gaur et al., 28 Mar 2025).

The experimental probe is ferromagnetic resonance in a coplanar-waveguide-based NanOsc PhaseFMR-40 system at room temperature over $\pi$3–$\pi$4, with an in-plane static magnetic field and flip-chip sample placement. The main observables are the linewidth $\pi$5 and the resonance field $\pi$6, obtained by fitting the derivative FMR spectra to symmetric and antisymmetric Lorentzian components. The key finding is linewidth broadening and damping enhancement when HSQ is added relative to bare NiFe. $\pi$7 increases linearly with frequency, and the slope is larger for HSQ/NiFe than for reference NiFe, which is interpreted as enhanced Gilbert damping consistent with spin pumping. The inhomogeneous linewidth broadening remains below $\pi$8 for all samples, indicating high film quality (Gaur et al., 28 Mar 2025).

For the strongest case, HSQ$\pi$9/NiFe$\mathrm{=C=C=}$0, the effective damping $\mathrm{=C=C=}$1 reaches $\mathrm{=C=C=}$2, compared with $\mathrm{=C=C=}$3 for the reference sample, corresponding to an $\mathrm{=C=C=}$4 increase. The resonance-field-versus-frequency data fit the Kittel formula well, and the effective magnetization remains similar across HSQ thicknesses, supporting the interpretation that the differences are interfacial and dynamical rather than due to large changes in static magnetic properties. By fitting $\mathrm{=C=C=}$5 versus inverse NiFe thickness, the work extracts effective spin mixing conductances of $\mathrm{=C=C=}$6 for reference NiFe, $\mathrm{=C=C=}$7 for HSQ = $\mathrm{=C=C=}$8, $\mathrm{=C=C=}$9 for HSQ = $xy$0, and $xy$1 for HSQ = $xy$2. The maximum value is described as comparable to heavy metals such as Pt, Ta, W, and Pd (Gaur et al., 28 Mar 2025).

The unusual aspect is that $xy$3, $xy$4, and the effective spin current density all increase with HSQ thickness. This is interpreted as evidence that HSQ behaves as a bulk spin sink with a large effective spin diffusion length, rather than as a thin interfacial perturbation. Reversed stacks, in which NiFe is deposited first and HSQ is spin-coated on top, show the same qualitative trend, and HSQ/CoFeB exhibits a $xy$5 damping increase at HSQ = $xy$6, indicating that the effect is not unique to NiFe (Gaur et al., 28 Mar 2025).

The mechanism is explicitly left open. Conventional spin-pumping expectations would predict weak coupling in an insulating, weak-spin-orbit polymer, so the large response is proposed to arise from a combination of spin pumping and orbital pumping, referred to as possible spin-orbital pumping. This suggests that out-of-plane polymer coupling interfaces in spintronics cannot always be reduced to standard spin-injection arguments; orbital angular momentum transfer may provide an additional interfacial channel, especially in light-element systems.

4. Three-dimensional polymer couplers in integrated photonics

In integrated photonics, out-of-plane polymer coupling interfaces are engineered structures that rotate optical propagation between a planar waveguide and the substrate-normal direction. Direct laser written polymer waveguides fabricated from the negative-tone photoresist EpoClad 50 on silica coverslips provide a canonical example. The platform combines long planar waveguides running parallel to the substrate with three-dimensional couplers that rotate the optical path by $xy$7, allowing light to enter or leave the chip through the glass substrate (Landowski et al., 2017).

The structures are fabricated by two-photon lithography in a Nanoscribe Photonic Professional GT using a $xy$8, $xy$9 objective, a $\Delta z = 17 \pm 3\ \text{pm}$0 laser with $\Delta z = 17 \pm 3\ \text{pm}$1–$\Delta z = 17 \pm 3\ \text{pm}$2 pulses at $\Delta z = 17 \pm 3\ \text{pm}$3, writing speed of $\Delta z = 17 \pm 3\ \text{pm}$4, and power of $\Delta z = 17 \pm 3\ \text{pm}$5–$\Delta z = 17 \pm 3\ \text{pm}$6. The substrate is a $\Delta z = 17 \pm 3\ \text{pm}$7 thick silica coverslip, and each waveguide is composed of four parallel trajectories. The basic structures are bare, air-clad polymer waveguides with approximately $\Delta z = 17 \pm 3\ \text{pm}$8 cross section. Stadium geometries with straight sections connected by bends are used to quantify transmission, propagation loss, and bend loss. Reported dimensions include bend radii down to $\Delta z = 17 \pm 3\ \text{pm}$9 in general operation, fabrication of bends down to $\lambda = 0.95 \pm 0.10\ \text{nm}$0 in some structural descriptions, total waveguide lengths up to $\lambda = 0.95 \pm 0.10\ \text{nm}$1, and writing-field sizes of $\lambda = 0.95 \pm 0.10\ \text{nm}$2 (Landowski et al., 2017).

The coupler itself is a quarter ring with radius $\lambda = 0.95 \pm 0.10\ \text{nm}$3, standing on the substrate and smoothly lowered onto the substrate by a $\lambda = 0.95 \pm 0.10\ \text{nm}$4-shaped transition. The transition section is $\lambda = 0.95 \pm 0.10\ \text{nm}$5 long, and the full coupling structure is about $\lambda = 0.95 \pm 0.10\ \text{nm}$6 high and $\lambda = 0.95 \pm 0.10\ \text{nm}$7 long. A shallow rectangular pad at the end improves adhesion and coupling efficiency, with both parallel and perpendicular pad orientations compared. Because the end facets sit at the interface of resist and substrate, no facet polishing or preparation is necessary (Landowski et al., 2017).

The main significance of the out-of-plane coupler is operational rather than merely geometric. Both input and output ports can be addressed through the substrate, imaged with a single microscope objective, and kept within one microscope field of view. The top side of the sample remains available for other components or manipulation. Reported performance includes insertion losses on the order of $\lambda = 0.95 \pm 0.10\ \text{nm}$8, a best value of $\lambda = 0.95 \pm 0.10\ \text{nm}$9 for a waveguide up to $10\%$0 long, and upper bounds on propagation loss of $10\%$1 for $10\%$2 and $10\%$3 for $10\%$4. The dominant loss mechanism is identified as imperfect mode matching at straight-to-bent transitions and at the in-/out-couplers, indicating that the intrinsic propagation can be comparatively low loss even when total insertion loss remains substantial (Landowski et al., 2017).

The same platform also supports Y-beam splitters with final waveguide separation $10\%$5 and reported splitting angles of $10\%$6, $10\%$7, $10\%$8, and $10\%$9. Longer splitting sections reduce insertion loss and improve balance. Compatibility with nitrogen-vacancy centers in nanodiamonds is emphasized through low enough losses for weak single-photon signals and autofluorescence only modestly above the glass background. Here the out-of-plane interface serves simultaneously as coupler, alignment aid, and observation channel.

5. Thermal coupling geometry in polymer–graphene nanocomposites

In thermal transport, the central issue is not a coupler that sends heat normal to the substrate, but the contrast between undesirable reliance on weak out-of-plane pathways and deliberate redirection into strong in-plane ones. A combined molecular dynamics and atomistic Green’s function study of polyethylene (PE) with graphene nanoplatelets (GnPs) shows that edge functionalization of GnPs outperforms basal-plane functionalization in raising effective thermal conductivity, especially for multilayer platelets (Muthaiah et al., 2022).

The physical basis is graphene’s thermal anisotropy: about $-883\ \text{meV/unit\ cell}$0 in-plane versus about $-883\ \text{meV/unit\ cell}$1 out of plane. In multilayer GnPs, out-of-plane transport is even poorer because adjacent layers are coupled only by weak van der Waals interactions. With edge functionalization, polymer chains bond to the exposed edges of every graphene layer, coupling the polymer directly to the high-conductivity in-plane pathways of essentially all sheets. With basal-plane functionalization, polymer chains bond mainly to the outermost surface layers, so heat entering those sheets must propagate through the platelet thickness to activate inner layers; this route is intrinsically inefficient (Muthaiah et al., 2022).

The molecular dynamics simulations use LAMMPS and the COMPASS force field for PE composites with $-883\ \text{meV/unit\ cell}$2 graphene nanoplatelets of $-883\ \text{meV/unit\ cell}$3 to $-883\ \text{meV/unit\ cell}$4 layers at $-883\ \text{meV/unit\ cell}$5 graphene loading and $-883\ \text{meV/unit\ cell}$6. For a $-883\ \text{meV/unit\ cell}$7-layer nanoplatelet at $-883\ \text{meV/unit\ cell}$8, the edge-functionalized composite has up to $-883\ \text{meV/unit\ cell}$9 higher effective thermal conductivity or heat flux than the basal-plane-functionalized composite. The edge advantage is about $-651\ \text{meV/unit\ cell}$0 for $-651\ \text{meV/unit\ cell}$1 layers and rises to $-651\ \text{meV/unit\ cell}$2 for $-651\ \text{meV/unit\ cell}$3 layers. In single-platelet calculations, the ratio $-651\ \text{meV/unit\ cell}$4 increases from about $-651\ \text{meV/unit\ cell}$5 for $-651\ \text{meV/unit\ cell}$6 layers to greater than $-651\ \text{meV/unit\ cell}$7 for $-651\ \text{meV/unit\ cell}$8 layers (Muthaiah et al., 2022).

The study further reports that, at $-651\ \text{meV/unit\ cell}$9 length and $90^\circ$00, pristine graphene has thermal conductivity of $90^\circ$01, edge-functionalized graphene $90^\circ$02, and basal-plane-functionalized graphene $90^\circ$03. Thus edge functionalization retains much more of graphene’s intrinsic transport than basal-plane functionalization. The difference is attributed to the stronger structural distortion produced when basal-plane covalent bonding converts basal carbon atoms from $90^\circ$04 to $90^\circ$05, which enhances phonon scattering, whereas edge atoms accommodate functional groups with less strain (Muthaiah et al., 2022).

Interfacial analysis supports the same picture. With a $90^\circ$06 temperature difference across an individual polymer–graphene junction, the edge case shows about $90^\circ$07 higher heat flux in molecular dynamics; atomistic Green’s function calculations based on DFT-derived interatomic force constants give about $90^\circ$08 higher interfacial thermal conductance for the edge case at $90^\circ$09. The key microscopic result is that edge functionalization strongly enhances transmission into graphene’s in-plane phonon modes, while out-of-plane transmission is similar for edge and basal-plane cases. The result is conceptually important for out-of-plane polymer coupling interfaces because it shows that simply lowering interfacial resistance is not sufficient; the interface must couple the polymer into the filler’s favorable transport manifold rather than trap it in a weak out-of-plane bottleneck (Muthaiah et al., 2022).

6. Orbital symmetry, mediation layers, and cross-domain design principles

A related organic spin-interface study, though not itself a polymer system, clarifies a principle that generalizes to polymer interfaces: coupling strength and even coupling sign can be governed by orbital symmetry, especially by the presence or absence of out-of-plane active orbitals. In metal phthalocyanine/graphene/Co/Ir(111) structures, FePc and CuPc adsorb with essentially the same geometry on graphene/Co, yet FePc couples antiferromagnetically to Co whereas CuPc couples ferromagnetically and more weakly. The decisive difference is electronic: FePc has active $90^\circ$10 and $90^\circ$11 orbitals with out-of-plane character, while CuPc has a moment mainly in the in-plane $90^\circ$12 orbital (Avvisati et al., 2018).

Graphene functions simultaneously as an electronic decoupling layer and an exchange mediator. Because the molecules retain largely molecular spectral character, the coupling can be interpreted primarily in terms of symmetry-selected superexchange rather than uncontrolled chemisorption. The proposed pathways are Fe–N–Gr–Co for FePc, favoring antiferromagnetic coupling through a $90^\circ$13-type superexchange geometry, and a weaker $90^\circ$14-type superexchange for CuPc. Temperature-dependent XMCD yields approximate coupling energies of $90^\circ$15 for FePc on $90^\circ$16 Co, $90^\circ$17 for FePc on $90^\circ$18 Co, and $90^\circ$19 for CuPc on $90^\circ$20 Co; FePc remains antiferromagnetically coupled up to about $90^\circ$21 in the out-of-plane-Co case and retains residual XMCD at room temperature in the in-plane-Co case, whereas CuPc loses remanence above about $90^\circ$22 in the out-of-plane-Co case (Avvisati et al., 2018).

Across the polymer-centered cases, several design principles recur. First, mechanical rigidity does not preclude strong out-of-plane response: PDFHE shows that electronically driven symmetry lowering can overcome lattice bending costs (Wang et al., 20 Aug 2025). Second, large dynamical coupling does not require a conventional heavy-metal receiver: HSQ demonstrates giant effective spin mixing conductance even though it is insulating and low-spin-orbit-coupled (Gaur et al., 28 Mar 2025). Third, out-of-plane access can be an enabling systems-level feature rather than a parasitic one: the polymer waveguide coupler uses the vertical channel to simplify addressing and microscopy (Landowski et al., 2017). Fourth, the most effective interface is often the one that avoids weak out-of-plane transport and instead injects energy into strong in-plane channels, as in edge-functionalized graphene nanoplatelets (Muthaiah et al., 2022).

Taken together, these studies delimit the present meaning of out-of-plane polymer coupling interfaces. They are interfaces in which normal-direction geometry or transport is not a secondary perturbation but the determining variable for symmetry breaking, damping, optical access, or thermal conductance. The literature therefore treats out-of-plane coupling not as a single mechanism but as a family of interface problems unified by directional selectivity, interfacial mediation, and the competition between geometric constraint and electronic, magnetic, optical, or phononic stabilization.