Spin-Polarisation measurement using NbN-Insulator-Ferromagnet Tunnel Junction with oxidized barrier

The Gist

Imagine you are trying to sort a mixed bag of red and blue marbles. In the world of electronics, these "marbles" are electrons, and they come in two flavors: "spin-up" and "spin-down." For many modern technologies (like faster computers), we need to know exactly how many of these electrons are red versus blue. This mix is called spin-polarization.

To count them, scientists use a clever trick involving a special kind of "magnetic sieve." This paper describes a new, easier way to build that sieve.

The Old Way: A Picky, Cold Filter

For decades, scientists used a material called Aluminum to build this sieve. Think of Aluminum as a very sensitive, high-precision filter. It works great, but it has a major flaw: it only works when it is freezing cold (colder than 1 Kelvin, or -272°C). To get it that cold, you need expensive, complex equipment (like a 3He cryostat), which is like needing a specialized industrial freezer just to keep a popsicle frozen.

Also, building these Aluminum filters was like assembling a complex Lego set with four different layers, requiring precise masks and many steps.

The New Way: A Robust, Simple Filter

The researchers in this paper found a better material: Niobium Nitride (NbN). Think of NbN as a tougher, more robust filter.

• It stays cold longer: NbN can handle temperatures up to 1.6 Kelvin (still very cold, but much warmer than Aluminum). This means you can use a standard, cheaper "home freezer" (a 4He cryostat) instead of the industrial one.
• It's easier to build: Instead of a complex 4-step assembly, they used a simple two-step process.

How They Made It: The "Rust" Trick

Here is the clever part of their invention. Usually, to make a tunnel junction (the filter), you need to sandwich a superconductor, an insulator (a barrier), and a metal.

• The Old Method: You had to deposit a separate insulating layer (like MgO) between the layers.
• The New Method: They took the NbN film and simply let it rust (oxidize) in the air or pure oxygen. This created a thin, uniform layer of "rust" (oxide) right on the surface of the NbN. They then placed a metal strip (Cobalt) on top of this rust.
• The Result: The rust acts as the perfect insulating barrier. It's like turning the surface of a metal sheet into a natural, self-made wall that electrons have to tunnel through.

How It Works: The Magnetic Split

To measure the spin, they put the device in a strong magnetic field.

1. The Split: In a superconductor, electrons usually pair up. But when you apply a strong magnetic field parallel to the film, these pairs get pulled apart. The "spin-up" electrons and "spin-down" electrons get pushed into different energy lanes. It's like a highway where the magnetic field forces red cars into the left lane and blue cars into the right lane.
2. The Tunnel: When they push electricity through the device, the electrons try to tunnel through the rust barrier.
3. The Asymmetry: If the metal on the other side (Cobalt) has more "red" electrons than "blue" ones, the current will flow more easily into the matching lane. This creates a lopsided (asymmetric) signal. By measuring this lopsidedness, they can calculate exactly how many red vs. blue electrons are in the Cobalt.

What They Found

• Thickness Matters: They found that the NbN film had to be very thin (less than 10 nanometers, which is about 100,000 times thinner than a human hair) for the magnetic "splitting" to work clearly. At 5 nanometers, the effect was very strong.
• Reliable Results: They tested this with Cobalt and found they could measure its spin-polarization reliably at temperatures up to 1.6 K.
• Air vs. Pure Oxygen: They tried making the "rust" in regular air and in pure oxygen. The pure oxygen version made a better, more consistent barrier with higher resistance, which is easier to measure without heating up the sample.

The Bottom Line

This paper shows that you don't need the ultra-expensive, ultra-cold equipment or the complex manufacturing steps to measure electron spin anymore. By using a simple "rust" barrier on a tougher material (NbN), scientists can now measure spin-polarization in standard, cheaper lab equipment. This makes the technique much more accessible for testing new materials for future electronics.

Technical Summary

Technical Summary: Spin-Polarisation Measurement using NbN-Insulator-Ferromagnet Tunnel Junctions

Problem Statement
Determining the spin-polarization of ferromagnets is critical for spintronic applications. While the Meservey-Tedrow (MT) technique using superconducting tunneling spectroscopy (STS) is a standard method for measuring both the magnitude and sign of spin-polarization, it faces significant limitations. The technique requires splitting the superconducting density of states (DOS) into spin-up and spin-down sub-bands via a parallel magnetic field. However, this splitting is often smeared by diamagnetic orbital currents and spin-orbit coupling, limiting sensitivity.

The most widely used superconductor for MT measurements, Aluminum (Al), has a low critical temperature ($T_c \approx 1.5 - 2.5$ K), necessitating measurements in $^3$He cryostats (below 1 K). While Niobium Nitride (NbN) offers a much higher $T_c$ ($\sim 16$ K), allowing for measurements in standard $^4$He cryostats, previous attempts to fabricate NbN-based tunnel junctions for MT measurements involved complex processes requiring four separate shadow masks and distinct deposition steps for the bottom electrode, tunnel barrier (MgO), isolation pads, and top electrode.

Methodology
The authors propose a simplified, two-step fabrication process for Superconductor-Insulator-Normal metal (SIN) and Superconductor-Insulator-Ferromagnet (SIF) tunnel junctions.

1. Substrate and Growth: Thin NbN films are grown on (100) MgO substrates via reactive DC magnetron sputtering at 600°C, achieving a bulk $T_c$ of $\sim 16$ K.
2. Fabrication: A single shadow mask defines a 300 $\mu$m wide NbN strip. The insulating tunnel barrier is formed by thermally oxidizing the NbN surface in air or high-purity oxygen at 250°C for 1.5–2 hours, creating a uniform amorphous oxide layer ($\sim 2$ nm thick).
3. Electrode Deposition: A cross-strip of a normal metal (Ag or Au) or a ferromagnet (Co, 10–20 nm) is deposited at room temperature. For Co junctions, an Au overlayer is added to provide a parallel low-resistance path, eliminating artifacts caused by the series resistance of the Co arms.
4. Measurement: Tunnelling conductance ($dI/dV$) is measured using a 6-terminal configuration in a 3He cryostat with a 110 kOe magnetic field aligned parallel to the film plane. The data is analyzed using Maki's theory, which accounts for orbital depairing, Zeeman splitting, and spin-orbit scattering.

Key Results

• Thickness Dependence: The study systematically varied the NbN film thickness ($t$). Clear signatures of Zeeman splitting in the tunnelling spectra were observed for films with $t \le 10$ nm, becoming pronounced at $t \approx 5$ nm. Thicker films (20 nm, 40 nm) showed broadening but no distinct splitting even at 105 kOe.
• Parameter Extraction: Fitting the spectra for 5 nm NbN/oxide/Ag junctions yielded consistent parameters. The orbital depairing parameter ($\zeta$) increased quadratically with the magnetic field, while the spin-orbit coupling parameter ($b$) showed a weak increasing trend with thickness.
• Spin-Polarization of Cobalt: Using NbN/oxide/Co junctions (5 nm NbN), the authors measured the spin-polarization ($P$) of Cobalt.
  • At 0.78 K, a spin-polarization of $P \approx 0.21 \pm 0.005$ was obtained.
  • This value is consistent with spin-polarized photoemission measurements but lower than early MT reports ($P \approx 0.35$). The authors attribute the discrepancy in earlier reports to the use of simplified analysis schemes that ignored spin-orbit coupling.
  • Measurements remained successful up to 1.57 K, demonstrating the viability of the higher $T_c$ material.
• Oxidation Environment: While air oxidation produced functional junctions, the resulting low resistance ($< 0.5 \Omega$) caused contact heating issues. Oxidation in 99.99% pure oxygen yielded junctions with higher resistance (20–30 $\Omega$) and significantly improved reproducibility, without altering the extracted spin-polarization value ($P \approx 0.21$).

Significance and Claims
The paper claims that the proposed two-step fabrication process using NbN with an oxidized surface barrier is a viable, simple alternative to the complex multi-mask processes previously required for NbN devices. By utilizing the natural oxide of NbN, the authors eliminate the need for depositing separate insulating layers like MgO.

The primary significance lies in the ability to perform Meservey-Tedrow spin-polarization measurements at temperatures up to 1.6 K (and potentially higher) using conventional $^4$He cryostats, rather than the sub-1 K $^3$He systems required for Aluminum-based junctions. Combined with the ease of fabrication without sophisticated lithography, the authors suggest this approach makes the MT technique more accessible for measuring spin-polarization in new materials.