TB-6987 · REV E · Technical newsheet

Test & MeasurementDevice profile

Space-drawn ZBLAN averages 0.75 dB/m in 2026 tests

A 2026 study found space-drawn step-index ZBLAN fiber averaged 0.75 dB/m at 1.55 µm, versus 4.50 dB/m for ground fiber; customer qualification still requires matched optical and mechanical testing.

By Olivia Hart5 min read973 words

Features

  • Flawless Space Fibers-1 produced almost 12 km of ZBLAN aboard the ISS in 2024.
  • The longest continuous orbital fiber reached 1,141 m, and researchers analyzed more than 2 km after its return.
  • Space-drawn step-index ZBLAN averaged 0.75 dB/m at 1.55 µm, versus 4.50 dB/m for ground-drawn samples.
  • Single-index ZBLAN averaged about 0.16 dB/m at 1.55 µm in both environments.
  • The U.K. Space Agency announced SkyYield on February 9, 2026, and the feasibility study has since been completed.
Space-made infrared fiber must prove its performance
Device photoSpace-made infrared fiber must prove its performance — cocoparisienne / pixabay

By Olivia Hart

Space-drawn step-index ZBLAN fiber averaged 0.75 dB/m at 1.55 µm in 2026 testing, compared with 4.50 dB/m for corresponding ground-drawn samples.

That result applies to one fiber structure and one wavelength. Step-index fiber has a distinct core and cladding. Single-index ZBLAN averaged about 0.16 dB/m at 1.55 µm in both environments. The 2026 study’s contrast shows how strongly composition and design influence the outcome.

What did the orbital run produce?

Flawless Space Fibers-1 produced almost 12 km of ZBLAN aboard the International Space Station in 2024. The longest continuous length reached 1,141 m; researchers returned the material to Earth and analyzed more than 2 km of selected fiber.

H. A. Moser and colleagues covered the flight work in Acta Astronautica, volume 246, pages 956–967; M. A. Arat and colleagues reported the attenuation comparison in Space Science and Technology, volume 6, article 0572.

ZBLAN comprises zirconium, barium, lanthanum, aluminum, and sodium fluorides. Its transmission reaches into the mid-infrared beyond conventional silica fiber, making it relevant to infrared lasers, spectroscopy, sensing, and specialist light-delivery systems.

Drawing the glass remains difficult because crystals, impurities, bubbles, and core-geometry changes introduce optical loss. Gravity also affects convection and crystallization during heating and cooling, the processes that can separate orbital production from terrestrial processing.

Attenuation measures optical-power loss per unit length, so a customer cannot compare dB/m with dB/km or carry a 1.55 µm result into a 2–5 µm band without new data. The 2026 comparison is a measured design-and-environment result, not a production specification. That distinction matters.

What must SkyYield prove?

The U.K. Space Agency’s February 9, 2026 announcement, “New studies for manufacturing advanced materials in orbit,” named OrbiSky’s SkyYield feasibility study; the company has since completed it. According to OrbiSky, the work established a preliminary payload architecture, process-control requirements, and a verification approach for microgravity ZBLAN manufacturing. The next stage is functioning, instrumented hardware.

OrbiSky founder and CEO Sylvester Kaczmarek, whose company develops autonomous orbital industrial operations, wrote: “At OrbiSky, we’re developing the operational side of this approach: Monitored processing steps, bounded responses to faults, and records that connect commands and sensor readings to each experimental run.”

For an orbital drawing system, OrbiSky says to capture temperature history, draw speed, tension, geometry, and interruptions in the product’s evidence package. Kaczmarek also wrote: “Reliable autonomous operations become valuable when they support consistent manufacturing and usable test results.”

What must a customer qualify before buying?

A short fiber for a mid-infrared laser, a spectroscopic instrument, and a long optical link must meet different loss, beam-quality, geometry, bend-performance, and continuous-length limits. Qualification should begin with a defined use case and agreed acceptance limits, not a headline loss figure.

Measurements at 1.55 µm are informative, but many ZBLAN uses lie farther into the infrared. Candidate programs should measure across the proposed operating band, including 2 to 5 µm when relevant.

  • Measurement basis: fiber structure, launch conditions, specimen length, and uncertainty.
  • Interface losses: connector or coupling losses and how the laboratory treats them.
  • Physical acceptance: core geometry, bend performance, and the longest continuous length meeting specification.

IEC 60793-1-40:2024, Optical fibers—Part 1-40: Attenuation measurement methods, describes cut-back, insertion-loss, and backscattering methods. In a cut-back test, the laboratory compares transmitted power before and after shortening the same specimen while maintaining the optical launch conditions. Any adaptation for fluoride glass or longer wavelengths needs documentation.

The study’s Figure 1 likewise makes clear why fiber type, wavelength, and test method must accompany every reported loss figure.

How can an orbital gain be separated from other variables?

An orbital sample needs a terrestrial draw made from closely matched input material. Investigators must record the variables that can change attenuation:

  • Material: feedstock purity, preform geometry, core and cladding composition.
  • Process: thermal history, draw speed, and atmosphere.
  • Post-draw: coating and storage.

Without those records, a ground-versus-flight difference cannot be assigned cleanly to gravity.

Figure 2’s proposed protocol follows the same material through two processing environments and a common set of independent tests. A matched ground draw helps isolate microgravity; a separate, commercially available reference fiber shows performance customers can buy today.

Material analysis tests the optical result. Microscopy, spectroscopy, crystal-inclusion inspection, diameter checks, and concentricity checks can identify loss mechanisms. The 2026 study linked much of the step-index improvement to reduced crystal-related scattering, while impurities and other defects remained relevant in both environments.

What evidence must a delivery specification contain?

Manufacturers should build the delivery record around several linked requirements:

  • Output: total drawn length and the longest continuous length meeting specifications.
  • Uniformity: loss distribution, with beginning, middle, and end measurements revealing process drift.
  • Yield: the amount rejected during production.
  • Traceability: each coded specimen linked to its preform, operating conditions, processing events, storage, and laboratory results.

Mechanical qualification deserves the same discipline:

  • Mechanical evidence: tensile strength, proof testing where appropriate, bend performance, coating integrity, handling damage, and storage stability.
  • Repeatability: average performance and variability across repeated manufacturing runs.
  • Independence: independent testing of coded specimens to strengthen confidence in the comparison.

What must happen before scale-up?

A flight experiment can identify a manufacturing advantage, but only reproducible production evidence supports customer qualification and investment decisions. The final commercial measure is fiber qualified to a customer’s specification, including inspection, rejected yield, coating, packaging, transport, and full process cost.

The 2026 study of the flight experiment provides a technical baseline, not production proof. Can SkyYield convert the 0.75 dB/m result into repeatable runs that deliver customer-qualified fiber at acceptable cost, with every meter traceable to its preform, process, and test record?

via gov.uk (Original)

Filed under

  • zblan-fiber
  • optical-attenuation
  • fiber-qualification
  • microgravity-manufacturing
  • infrared-optics
Share this article:

More from Olivia Hart

Olivia Hart

Show full bio

Market editor covering technology and product news in Instruments, test and measurement at Testbench Report.

97 articles

Application notes

« Previous article