HELIOS-3D: Alternative Materials and Methods Decision Memo

This memo evaluates lower-risk alternatives that preserve the HELIOS-3D architectural thesis while reducing exposure to the two weakest assumptions in the current plan:

1. Conformal deposition of a vdW ferromagnet such as $Fe_3GaTe_2$ onto a complex 3D polymer scaffold.
2. Microwave-first readout as the primary sensing path for early prototypes.

The conclusion is not that the current hypothesis is invalid. The conclusion is narrower and more precise: the current hypothesis is stronger as a phased program than as a single integrated fabrication jump.

1. Decision Criteria

Alternatives are evaluated against the following criteria:

2. Bottom-Line Recommendation

For the first demonstrator, HELIOS-3D should pivot to a planar-first, electrically read, multilayer spintronic stack. The most defensible path is:

1. Material system: Ir/Fe/Co/Pt or Ir/Co/Pt for initial MCA experiments.
2. Transport-optimized branch: GdFeCo or related ferrimagnetic multilayers if skyrmion Hall suppression becomes dominant.
3. Fabrication path: planar lithography plus sputter/etch, then stacked or transferred integration for 3D density.
4. Readout/control: Hall, TMR, AHE/THE, SOT, or VCMA before any microwave-first architecture is attempted.

This preserves the logic of HELIOS-3D while replacing the least-supported fabrication step with an experimentally credible progression.

Among later-stage branch candidates, the strongest family is now the compensated ferrimagnet/altermagnet route: the compensated-ferrimagnet paper gives a natural spin-pumping readout mechanism, while the altermagnet paper shows current-driven nonlinear skyrmion dynamics and THz-scale helicity motion. That makes this path a stronger long-range candidate than freeform 3D magnetic coating, but still a branch beyond the planar demonstrator.

3. Candidate Materials

3.1 Ir/Fe/Co/Pt Multilayers

Supporting studies:

• Moreau-Luchaire et al., Nature Nanotechnology (2016), DOI: 10.1038/nnano.2015.313
• Soumyanarayanan et al., Nature Materials (2017), DOI: 10.1038/nmat4934

Advantages:

• Strongest empirical support for tunable, room-temperature skyrmions in a device-relevant multilayer stack.
• Entire stack can be deposited by established thin-film methods.
• Parameter tuning is well-documented through layer-thickness control.

Disadvantages:

• Still fundamentally a planar platform.
• Heavy-metal interfaces may impose power and damping penalties.

Compatibility with HELIOS-3D:

• MCA: Excellent alignment with deterministic track-based transport and switching experiments.
• BRC: Usable for confined stochastic reservoirs, though not yet a proven reservoir substrate.

Assessment:

• Best near-term replacement for $Fe_3GaTe_2$ in a first hardware demonstrator.

3.2 Ir/Co/Pt or Pt/Co/Ir Multilayers

Supporting studies:

• Woo et al., Nature Materials (2016), DOI: 10.1038/nmat4593
• Boulle et al., Nature Nanotechnology (2016), DOI: 10.1038/nnano.2015.315

Advantages:

• Simpler than Ir/Fe/Co/Pt.
• Extensively studied on standard substrates.
• Compatible with Hall-style electrical characterization.

Disadvantages:

• Smaller optimization space than the full Ir/Fe/Co/Pt family.
• Some operating regimes remain field-history sensitive.

Compatibility with HELIOS-3D:

• Excellent for a minimum viable MCA demonstrator.

Assessment:

• Best low-complexity first-build option.

3.3 GdFeCo Ferrimagnetic Films

Supporting study:

• Woo et al., Nature Communications (2018), DOI: 10.1038/s41467-018-03378-7

Advantages:

• Strong empirical evidence for suppressing the skyrmion Hall effect.
• Better candidate if rectilinear transport is more important than ultimate material novelty.
• Entirely compatible with sputtered-film processing.

Disadvantages:

• Compensation-point engineering adds process complexity.
• Rare-earth composition control is not trivial.

Compatibility with HELIOS-3D:

• Especially attractive for MCA routing and drift control.

Assessment:

• Best transport-focused pivot.

3.4 Pt/Co/Gd and Related Ferrimagnetic Multilayers

Supporting basis:

• Recent ferrimagnetic skyrmion reports in Pt/Co/Gd-like stacks.

Advantages:

• Retains ferrimagnetic transport benefits with conventional fabrication flows.
• PMA and effective DMI remain tunable through interface engineering.

Disadvantages:

• Less mature than Ir/Fe/Co/Pt as a benchmark system.

Compatibility with HELIOS-3D:

• Strong option if the project wants ferrimagnetic transport but a different engineering tradeoff from GdFeCo.

Assessment:

• Strong secondary option.

3.5 Co-Doped Fe5GeTe2

Supporting study:

• Room-temperature skyrmion lattice in Co-doped Fe5GeTe2, Science Advances (2022), DOI: 10.1126/sciadv.abm7103

Advantages:

• Preserves the layered-metal spirit of the current repository.
• Stronger direct skyrmion evidence than the present Fe3GaTe2 record.

Disadvantages:

• Does not eliminate the fabrication blocker that affects vdW materials on 3D polymer hosts.

Compatibility with HELIOS-3D:

• Scientifically closer to the current hypothesis than metallic multilayers.

Assessment:

• Better science-risk profile than Fe3GaTe2, but only modestly better manufacturing risk.

3.6 Mn-Based Heuslers such as Mn3Ga

Supporting review:

• Hirohata et al., Materials (2018), DOI: 10.3390/ma11010105

Advantages:

• High PMA, strong spintronic device ecosystem, MTJ compatibility.
• More fab-compatible than vdW magnets.

Disadvantages:

• The skyrmion evidence base is weaker than for multilayer DMI systems.
• May force the project toward domain-wall or bubble carriers rather than strict skyrmion logic.

Compatibility with HELIOS-3D:

• Preserves the broad architecture but may alter the exact carrier model.

Assessment:

• Robust fallback if manufacturability becomes dominant over carrier novelty.

3.7 Compensated Ferrimagnets and Altermagnets

Supporting studies:

• Lee et al., Spin Current Generation Controlled by the Néel State in a Compensated Ferrimagnet (arXiv:2507.05618, 2025/2026)
• Liu et al., Current-driven nonlinear skyrmion dynamics in altermagnets (arXiv:2601.13499, 2026)

Advantages:

• Compensated ferrimagnets provide a magnetization-free readout route via Néel-state-controlled spin pumping.
• Altermagnets preserve strong spin dynamics without the stray-field burden of conventional ferromagnets.
• Together they form the strongest late-stage candidate family for transport and readout once planar demonstrators are validated.

Disadvantages:

• Compensation-point tuning and current-density windows still need process validation.
• These materials do not remove the need for planar fabrication discipline in early prototypes.

Compatibility with HELIOS-3D:

• Strong late-stage candidate for the MCA transport layer and any spin-pumping-based sensing path.

Assessment:

• Best long-range material family currently visible for a magnetization-free transport/readout branch.

4. Candidate Fabrication Methods

4.1 Planar-First Deposition, Then Transfer or Stacking

Concept:

Deposit magnetic stacks on flat substrates or membranes where the literature support is strongest, then transfer them onto intermediate structures or vertically stack them to recover density.

Advantages:

• Removes the unsupported assumption that vdW or multilayer magnetism must be established directly on a freeform polymer host.
• Allows magnetic, transport, and readout physics to be validated independently before volumetric integration.

Disadvantages:

• Introduces transfer alignment and yield as new engineering risks.

Assessment:

• Best fabrication pivot because it directly attacks the largest current uncertainty.

4.2 Nanoimprint Lithography Plus Blanket Sputter/Etch

Supporting studies:

• Chou et al., Applied Physics Letters (1995), DOI: 10.1063/1.114851
• Chou et al., Science (1996), DOI: 10.1126/science.272.5258.85

Concept:

Use planar nanopatterning to define tracks, reservoirs, Hall crosses, and sensor regions, then deposit and pattern the magnetic stack conventionally.

Advantages:

• Much more realistic than the current DISH -> TPP -> ALD pipeline for a first prototype.
• High throughput and wafer-scale patterning are plausible.

Disadvantages:

• Gives up arbitrary freeform 3D geometries in early phases.

Assessment:

• Best route to a credible first demonstrator.

4.3 Direct Laser Writing and Volumetric Printing (DISH/TPP) for Passive Structures

Concept:

Retain TPP or related writing methods only for passive microfluidic, confinement, packaging, or photonic structures while keeping the magnetic layers planar. The demonstration of in-situ DISH printing (Wang et al., 2026) on fixed surfaces enables these 3D structures to be added directly to the planar spintronic substrate after magnetic thin-film processing.

Advantages:

• Preserves part of the current HELIOS fabrication identity without forcing magnetic deposition onto a poorly supported substrate.
• In-situ printing avoids the risks of sample rotation or mechanical disturbance of sensitive magnetic layers.
• High-speed (0.6s) and high-resolution (19 µm) packaging/scaffolding becomes a standard post-processing step.

Disadvantages:

• Reduces the rhetorical simplicity of a single monolithic 3D fabrication story.

Assessment:

• Strong compromise path.

4.4 Reconfigurable Scaffolds via Flexible-Rigid Transitions (Y-zipper)

Supporting study:

• Li et al., Y-zipper: 3D Printing Flexible–Rigid Transition Mechanism for Rapid and Reversible Assembly (CHI ‘26).

Concept:

Utilize 3D-printed interlocking zipper structures that can transition from flexible strips to rigid rods. This allows for the fabrication of complex 3D scaffolds that can be printed flat or in a flexible state, then “zipped” into a final rigid geometry for spintronic operation.

Advantages:

• Enables deployable or reconfigurable 3D spintronic architectures.
• Simplifies the fabrication of complex spatial curves by decomposing them into flexible components.
• Reversible assembly allows for modular testing and reconfiguration of functional blocks.

Disadvantages:

• Alignment precision of the zipping mechanism is likely in the tens-to-hundreds of microns range, far above the sub-micron requirements for spintronic interconnects.
• Introduces mechanical complexity and potential failure points at the zipper interfaces.

Compatibility with HELIOS-3D:

• Best suited for the macroscale scaffold (DISH/TPP) where it provides structural support and modular assembly.
• Could support the assembly of large-scale 3D spintronic networks from smaller, high-resolution TPP-printed blocks.

Assessment:

• Useful for structural reconfigurability and assembly at the macro-to-meso scale, but requires significant refinement to meet spintronic alignment tolerances.

4.5 Twist-Driven Deterministic Nucleation

Supporting study:

• Shi et al., Multistimuli-Controlled Topological Nucleation of Skyrmion Loops and Monopoles in Liquid Crystals (PRL 2026).

Concept:

Adapt the liquid crystal “twist reservoir” mechanism for magnetic hopfion synthesis. Instead of relying on stochastic thermal nucleation, the system uses the DISH optical periscope to rotate the incident polarization or apply localized laser heating. This drives a continuous accumulation of elastic twist in the van der Waals magnetic lattice. Once a critical threshold is reached, the system undergoes a topological phase slip, “zipping” the twisted magnetization into a stable 3D hopfion.

Advantages:

• Deterministic: Replaces stochastic nucleation with a controlled physical process.
• Low Energy: Uses elastic energy accumulation to overcome high topological barriers.
• Multistimuli: Can be triggered by light, electric fields, or heat, providing architectural flexibility.

Disadvantages:

• Requires precise calibration of the “twist-accumulation” threshold for different vdW stacks.
• Scaling to sub-100nm geometries may increase the required elastic stress beyond the material’s structural limits.

Assessment:

• Highly promising for the Topological Compiler’s “write” phase. Moves the nucleation problem from a bottleneck to a controllable engineering parameter.

5. Candidate Readout and Control Methods

5.1 Hall-Based Electrical Readout

Methods: anomalous Hall effect (AHE), topological Hall effect (THE), standard Hall crosses.

Advantages:

• Low-overhead compared with microwave spectroscopy.
• Already common in skyrmion and multilayer characterization.
• Readily multiplexed and integrated with electronics.

Disadvantages:

• Signal amplitude can be small and geometry-dependent.

Assessment:

• Best early sensing path for MCA experiments.

5.2 TMR / MTJ Sensors

Advantages:

• Stronger electrical contrast than Hall-only sensing in many device settings.
• Compatible with memory-style readout infrastructure.

Disadvantages:

• Requires careful sensor placement and multilayer integration.

Assessment:

• Best route if the project needs more robust binary or multistate readout.

5.3 SOT and VCMA Control

Advantages:

• Both are already embedded in modern spintronic device literature.
• Strong alignment with low-latency electrical control.

Disadvantages:

• Stack engineering and gate integration can become nontrivial.

Assessment:

• Strong control-layer candidates for both MCA and later BRC experiments.

5.4 Graphene Electron-Optic Spin Interfaces

Supporting study:

• Burrow et al., Ballistic Spin Valve in Graphene Realized via Electron Optics (2026), DOI: 10.1103/nz6m-kb4l

Concept:

Use high-mobility graphene spin valves and transverse magnetic focusing to route, amplify, or invert spin signals through gate-tunable electron optics.

Advantages:

• Demonstrates spin-coherent ballistic transport in a planar device at low temperature, with quasiballistic spin-dependent focusing still observable at room temperature.
• Provides electrostatic control over spin-signal amplitude and polarity without requiring spin-orbit coupling in the transport channel.
• Suggests a low-voltage planar interface branch for spin routing or CMOS-adjacent control layers.

Disadvantages:

• It is not a magnetic-texture carrier path and does not validate skyrmion, hopfion, or Brownian reservoir operation.
• The demonstrated device depends on high-mobility graphene, ferromagnetic edge contacts, and transverse magnetic focusing geometry.

Assessment:

• Useful secondary branch for planar spintronic interfaces and control, not a replacement for the magnetic multilayer demonstrator path.

5.5 Microwave Spectroscopy and Hopfion Magnonics

Advantages:

• Strong physics basis for probing skyrmion and hopfion resonance modes.
• Breathing Mode Detection: Recent identifies distinct sub-GHz breathing modes in 3D hopfions. By using Ferromagnetic Resonance (FMR) or Brillouin Light Scattering (BLS), these frequencies can be detected as a fast, frequency-encoded spectral fingerprint.
• Electrical Integration: Offers a path for electrical readout that bypasses the energy-intensive processing of optical point clouds (NV-centers).

Disadvantages:

• RF shielding, conversion overhead, and signal-chain complexity may negate some energy benefits.
• Distinguishing between individual hopfion states in a dense array remains a spatial-resolution challenge for global RF probes.

Assessment:

• Valid as a parallel, electrical readout modality. Stronger as a hybrid sensing path than the purely optical baseline for high-density 3D nodes.

6. Comparative Decision Matrix

6.1 Readout Integration: hBN $V_B^-$ and Microwave Spectroscopy vs. Diamond NV Centers

HELIOS-3D formally shifts the primary readout strategy from Diamond NV centers to negatively charged boron vacancies ($V_B^-$) in hBN or direct Microwave Spectroscopy. This solves the $1/r^3$ distance-decay problem associated with NV centers and ensures compatibility with silicon bus architectures.

A critical architectural decision for the HELIOS-3D readout chain involves the choice of quantum sensor for detecting magnetic hopfion and skyrmion textures. The original hypothesis proposed diamond NV centers; the following analysis recommends a pivot to monolayer hexagonal boron nitride (hBN) with negatively charged boron vacancies ($V_B^-$).

6.1.1 Distance-Decay Problem with NV Centers

Diamond NV centers offer excellent spin coherence at room temperature, but suffer from a fundamental sensing geometry constraint:

• $1/r^3$ Falloff: The magnetic dipole-dipole interaction between the NV center and a magnetic texture scales as $1/r^3$, where $r$ is the separation distance.
• Optimal Distance: For reliable detection of skyrmion core fields ($B_z \sim 10-100$ mT), the NV center must be within $\sim 10-20$ nm of the magnetic texture surface.
• Integration Challenge: Achieving this proximity in a 3D stacked geometry while maintaining fabrication yield is non-trivial.

6.1.2 hBN $V_B^-$ as Alternative

Monolayer hBN with boron vacancies ($V_B^-$) offers a van der Waals-compatible alternative:

6.1.3 Recommended Integration Path

1. Primary: Monolayer hBN transferred onto the surface of the magnetic layer stack (either $Ir/Fe/Co/Pt$ for planar-first or $Fe_3GaTe_2$ for advanced nodes).
2. Sensing: Optically detected magnetic resonance (ODMR) via confocal microscopy or wide-field imaging.
3. Alternative: If optical access is limited, consider electrical readout via tunneling magnetoresistance (TMR) as a fallback per Section 5.2.

6.1.4 hBN-$Fe_3GaTe_2$ Interface Considerations

• Fermi Level Alignment: The $V_B^-$ charge state is stabilized by nearby charge traps; care must be taken during hBN transfer to avoid contamination that shifts the defect levels.
• Magnetic Proximity Effect: $Fe_3GaTe_2$ has high spin polarization; the hBN layer may experience magnetic proximity-induced splitting. This can be either a feature (enhancing sensitivity) or a noise source (introducing field variability).
• Temperature Operating Range: $V_B^-$ in hBN retains spin coherence up to $\sim 400$ K, comfortably covering data center operating temperatures (T < 85^\circ C).

7. Recommended Minimum Viable Demonstrator Revision

The current MVD sequence in OPEN_QUESTIONS.md should be interpreted more conservatively.

Revised Phase I

• Fabricate planar skyrmion or ferrimagnetic tracks on a conventional substrate.
• Validate carrier stability and electrical readout.

Revised Phase II

• Demonstrate deterministic MCA-style transport and state switching under current control.
• Use Hall or MTJ readout rather than a microwave-first chain.

Revised Phase III

• Implement confined stochastic dynamics in a planar or membrane-based reservoir.
• Establish whether the ensemble actually performs useful nonlinear transformations rather than merely exhibiting noise.

Revised Phase IV

• Add vertical integration via transfer, membrane stacking, or wafer-bonded assembly.
• Delay any freeform 3D magnetic coating claim until the flat-stack physics is secure.

8. Final Position

The strongest defensible version of HELIOS-3D is not a direct 3D-vdW fabrication jump. It is a staged program:

1. Prove the spin-texture logic on planar or membrane-compatible multilayers.
2. Replace microwave-first assumptions with electrical sensing and control.
3. Earn 3D density through stacking, transfer, or bonded integration.
4. Keep $Fe_3GaTe_2$ as a long-range candidate, not as the immediate fabrication anchor.

That revision preserves the project’s scientific ambition while removing the least-supported integration claims from the critical path.