Research into a suspended copper channel offers a useful example of how electrical transport and mechanical structure can be combined at the microscale. In a lateral spin valve, researchers measured spin transport through a copper channel suspended above a substrate. The signal was comparable in magnitude to that of a conventional device with a supported channel, while the suspended device showed a modestly shorter spin diffusion length. The results point to both the potential of suspended structures and the importance of controlling surface conditions.
Why suspend a spin-transport channel?
Spintronics uses the spin angular momentum of electrons, as well as their electrical charge. In a lateral spin valve (LSV), a ferromagnetic electrode injects spin-polarized electrons into a nonmagnetic channel, and a second ferromagnetic electrode detects the resulting spin accumulation. The separation between the electrodes allows researchers to study how the spin signal changes as it travels through the channel.
A suspended channel adds a mechanical degree of freedom to this electrical structure. That combination is relevant to spin mechanics: the interaction between spin angular momentum and mechanical motion. One motivating example is the Einstein–de Haas effect, in which a change in magnetization can produce mechanical rotation in accordance with conservation of angular momentum. Detecting such small mechanical effects requires a structure that can move, while electrical conductivity provides a route to injecting a conventional spin current.
Copper is a candidate material because it is conductive, relatively soft, and supports a long spin diffusion length at room temperature. The study therefore focused on whether an LSV could be fabricated with a suspended copper channel and whether the device could retain a clear electrical spin signal.
How the suspended copper channel was fabricated
The device was built on a silicon substrate covered with silicon dioxide. Researchers patterned shallow, square-shaped holes in the oxide and filled them with aluminum, which acted as a sacrificial layer. They then formed the LSV on top of the aluminum, using a copper channel and nickel–iron (Permalloy) electrodes.
After forming the device, the aluminum beneath the channel was removed with an aqueous tetramethylammonium hydroxide solution. Removing the sacrificial material left the copper channel suspended over the patterned opening. For comparison, conventional fixed-channel LSVs were fabricated on the same substrate. Atomic force microscopy was used to check the suspended structure; the measured height from the bottom of the opening to the top of the channel was consistent with the opening depth and channel thickness.
This process illustrates a general fabrication principle: a temporary supporting material can hold a structure in place during device formation, then be selectively removed to create a suspended section. In this case, the resulting geometry made it possible to test spin transport in copper without a substrate directly beneath the channel.
Measuring spin transport in the suspended device
The researchers used a non-local four-terminal measurement at room temperature. Current was injected through one ferromagnetic electrode, while the voltage was measured at the other. This arrangement detects changes associated with the relative magnetization directions of the injector and detector, rather than measuring only the voltage drop along the current path.
The measured magnetoresistance changed between parallel and antiparallel magnetic configurations. That response provided evidence that spin information had travelled through the suspended copper. For devices with electrode separations close to 300 nm, the reported spin signal was 4.0 milliohms in the suspended LSV and 3.2 milliohms in the fixed LSV. These measurements show that suspending the channel did not prevent detection of spin transport.
To estimate the spin diffusion length, the team measured how signal amplitude varied with the distance between the ferromagnetic electrodes. The spin diffusion length describes the characteristic distance over which a spin signal decays in a material. Fitting the distance dependence produced estimates of 340 ± 40 nm for the suspended channel and 390 ± 40 nm for the fixed channel.
Surface oxidation helps explain the difference
The shorter estimated spin diffusion length in the suspended device was attributed not to suspension itself, but to oxidation at the channel’s bottom surface. The suspended copper surface had been exposed to air, where a naturally formed oxide can increase spin scattering. More scattering reduces the distance over which spin information can travel.
Oxidation may also help explain why the suspended devices showed slightly larger signals at some electrode spacings. A resistive oxide can reduce the effective cross-sectional area available for transport, changing the channel’s spin resistance. The researchers also noted that the chemical treatment used to remove the sacrificial aluminum may affect copper oxide. Measurements found that the copper channel became 2–5 nm thinner after treatment, consistent with the proposed influence of surface oxidation and oxide removal.
These details matter when comparing suspended and supported devices. A change in measured transport may arise from interface or surface chemistry, as well as from the mechanical geometry. Separating those effects is essential when interpreting device measurements and refining fabrication processes.
Implications for device engineering
The work demonstrates that a suspended copper channel can support measurable spin transport in a lateral spin valve. It also provides a fabrication route that the researchers indicated could be applied to other spin-current channels. For engineers working with microscale devices, the broader lesson is that creating a mechanically released structure is only part of the challenge: surface condition, interfaces, and post-fabrication processing can influence electrical behavior.
For spin mechanics research, suspended copper offers a conductive platform with mechanical compliance that may support further investigation of interactions between spin currents and motion. The reported results do not establish a practical industrial automation component, but they add an important device-level demonstration for a developing area of materials and microsystems research.
Conclusion
A suspended copper channel can combine a mechanically released structure with detectable spin transport. In the reported lateral spin valve, the spin signal remained comparable to that of a fixed device, while the estimated spin diffusion length was slightly shorter. The proposed explanation—additional scattering associated with naturally oxidized copper—underscores the need to account for surface chemistry alongside geometry. For future spin-mechanics devices, fabrication control and careful interpretation of transport measurements will be central to turning a promising structure into a reliable experimental platform.