FOA Guide

Hollow Core Fiber

anti-resonant hollow core fiberOFS hollow core fiber preform OFS Hollow Core Fiber Preform

Hollow core fiber (HCF) is exactly that - rather than a core formed of soliid glass, the core of hollow core fiber is empty except for an inert gas. The reason it exists is that a gas has a lower index of refraction than glass so light travels about 50% faster and can have much less attenuation. Applications include high-frequency trading, high-performance computing and data center interconnection.

News during 2025 about hollow core fiber (HCF) indicated that it is becoming more widely accepted, not just by users but by major fiber manufacturers who are adding it to their product lines. In 2025, Prysmian announced a partnership with Relativity Networks, a company started by researchers at the University of Central Florida to manufacture HCF at their facilities in The Netherlands. In September 2025, Corning announced a partnership with Microsoft Azure who acquired Luminisity started by researchers at the University of Southampton, UK. OFS/Furakawa (now named Lightera) has been producing HCF for 5 years already.

In addition, equipment companies are announcing support for HCF. Test equipment companies EXFO and VIAVI announced support for HCF testing. Furakawa and Lightera announced a splicer for HCF and multicore fiber. Recent studies have been published on special conversion fibers used to splice HCF to regular singlemode fiber. When installation equipment manufacturers begin offering products for a new technology, you can bet it's becoming real.

What's the reason for all this interest? Latency can be reduced by the speed of HCF fiber. When people talk about speed in relation to optical fiber, they usually mean bandwidth, not the speed a pulse of light is transmitted down an optical fiber. But typical optical fibers where the light is traveling in pure glass, has a "speed of light" about 2/3 that of light in the vacuum of space or air. So in HCF, the pulse of light travels at about 1.5 times as fast as in a glass fiber. That makes a big difference in latency - about 1.5 microseconds per kilometer. For applications that want minimal latency, high speed stock trading for instance, that is a lot of time to save.

Besides high speed latency, data center operators are looking at the lower latency of HCF that can allow them to place data centers in more remote areas - especially those with more available electrical power or water for cooling - without incurring high latency.

Not that HCF doesn't also have the bandwidth for today's communications speeds. HCF is compatible with coherent transmission used for data rates above 100 Gb/s. Data rates above 1 Tb/s have been achieved.

A problem until recently has been the attenuation of HCF. But recent technical advantages in anti-resonant HCF have not only matched typical SM fiber but bettered it, with losses below 0.1 dB/km being achieved, much better than current glass SM fibers.

Cabling is still an issue. HCF fibers are larger than regular SM fibers (up to 300 microns in diameter) and more sensitive to stress, so loose tube cable designs are common with fewer fibers per tube.

Another obstacle to adoption of HCF is cost. Estimates seem to indicate that HCF is about ten times more expensive than regular singlemode fiber. And there are the practical considerations of installations and testing discussed below.

At the current time, HCF is a very small specialty use in fiber optics, but as data speeds increase and more users want lower latency, it may become more common and fiber techs should keep up to date on what they may be working with in the future.

For installation of HCF cable plants, the technology is about where optical fiber was in 1976-77. Installation is being done primarily by Ph.Ds from the manufacturers, but as applications grow in number, more qualified fiber installation technicians will be needed.


Practical Considerations - Installation and Testing
No new tech can succeed unless the support for installation and testing becomes available. That's beginning to happen for HCF. Fibers are cabled like regular fibers but the splicing and testing are quite different according to the sources FOA contacted.

Furakawa/FITEL is the only company currently offering a splicer for HCF that's also useful for multicore fiber, the FITEL S185EVROF. This machine has some very different features. It has an optical system that not only provides X-Y alignment, but it views the ends of the fibers to inspect cleaves and can rotate them to align the complex inner structure of HCF fibers. It has three instead of 2 electrodes that  create a uniform heat profile around the fiber
to heat the fiber evenly and provide precise temperature control, necessary for the complex structure of HCF. As we understand it, getting low loss requires only fusing the cladding, not the HCF core which must remain undisturbed at the splice. Datasheet for the splicer.

FITEL splicer for HCF

Having the proper tools is only part of the solution. As we have been told, splicing needs to be done quickly before too much outside air gets into the fiber. HCF is really hollow, but filled with a gas as part of the manufacturing process. As the fiber cools down the gas pressure in the fiber decreases. To keep outside air out of the fiber, the ends are sealed. Moist outside air is a problem; in regular fiber the fiber is coated with a plastic coating for protection. When HCF is cleaved, the low pressure sucks outside air into the fiber causing potential long term problems. Splicing seals the two spliced HCF fibers to prevent further moisture ingress. It almost sounds like a technique similar to TIG welding is needed, where an inert gas is used during the welding process to protect the materials being welded.

And while you can splice HCF to HCF to create long lengths, at the ends you need to splice HCF to regular SM fiber to connect to equipment or patch panels. HCF and SM fiber have very different mode field diameters (MFD), creating a problem for loss and reflectance at splices than can cause multipath interference. One solution today is to use an "adapter fiber," a graded index fiber that provides matching MFDs, and cleaving at an angle to reduce reflectance.

Once the HCF cable plant is installed, it needs testing. EXFO recently announced support for HCF testing in their products and shared their knowledge with FOA.

Insertion loss testing with a light source and power meter or OLTS is basically the same as any other fiber optic cable plant, except where they measure length. An OLTS that measures the length of the fiber is calibrated for an index of refraction of glass fiber of 1.46 - 1.47, while HCF has an index of refraction of around the index of refraction of air, say 1.003 - 1.005. If you measure the length of a HCF fiber link with a regular OLTS, the length will indicate about 2/3 of the actual length. So an instrument measuring length needs to be programmed for the index of refraction of HCF.

OTDR testing however is quite different and requires considerable differences in analysis software. Just like an OLTS, the length measurement requires reprogramming to the index of refraction of the HCF to get accurate length measurements. But the biggest difference is HCF fiber is essentially hollow, so the OTDR does not see the backscatter created in glass fiber, making the trace look very different. If anything the OTDR trace looks like the trace of a copper time domain reflectometer (TDR) without the characteristic backscatter of the OTDR trace of a glass optical fiber.

OTDR trace for HCF

HCF also needs testing for attenuation at typical wavelengths (1310 and 1550 nm), chromatic dispersion (CD) and polarization mode dispersion (PMD) just like glass fibers. Test equipment is becoming available for all these tests from companies like EXFO.

As we say above, installation today is being done primarily by Ph. D.s from the manufacturers, but as applications grow in number, more qualified fiber installation technicians will be needed. And they will need some specialized training.




 

Table of Contents: The FOA Reference Guide To Fiber Optics




 


(C)2025, The Fiber Optic Association, Inc.