The situation is quite different with the key material used to make optical cable – “bare fibre.” There is a global shortage. It has come about because the investment in fibre-making capacity did not match the growth in the amount of fibre needed. The shortage is not in silicon tetrachloride or other chemicals that may be considered the raw materials of fibre. Rather, the shortage is in the preforms – the cylindrical rods that are fabricated from the starting chemicals and then drawn into fibre.
The preform is the interim step between silicon chemicals and drawn fibre, somewhat similar to the way that wire-rod is the interim step between copper raw materials and copper wire for cable making. In the copper cable business, wire-rod capacity is not a limiting factor. In the optical cable business, preforms are the limiting factor. The other step in making fibre is the draw process. Currently, there is excess draw capacity, and this has been the case for many years.
The time to plan, build, and commission a new wire-rod facility may be 18 to 24 months. This also is about the time needed for a new preform factory. A new preform factory requires a minimum investment in the tens of millions of dollars – probably near US$100 million – to produce enough material to be cost effective. The investment in a new wire-rod facility may be slightly lower.
The current shortfall in preform capacity, however, has resulted more from the time element than the amount of the investment. For more than 20 years, there have been many years when global fibre demand has jumped up more than 15% or 20% in one year. Recently, this has been due to the massive network construction projects of China’s telecom operators. With growth at these rates and the 18-24 months to bring up new capacity, the industry hasn’t really achieved a good long-term balance. With metallic cable, demand in most places is growing at rates of 1.5 to 3.0 % per year, and maybe 5 to 7% in higher-growth markets such as China. So there is more time to plan for wire-rod capacity requirements.
In 2016, production of communications grade fibre used about 14,250 tonnes of preforms. This figure is about 0.1% of the 14.9 million tonnes of copper conductor used in wire and cable in 2016. Note there are fibres made for non-communication applications, such as laser power delivery, illumination, imaging bundles, and sensors. Some of these non-communication fibres are based on different optical materials, not silica glasses. The communications fibre, however represent the vast bulk of all fibres in terms of material tonnage, market value, and fibre kilometres.
The 14,250 tonnes of preforms made in 2016 were drawn in to 456 million kilometres of fibre. With most processes, one kg of preform can be converted to 32 km of fibre. There is a yield factor of about 94% in making cable. That is, some of the fibre is “lost” due to line start-up and shut-down processing, the handling of fibre reels and slight excesses on the reels, and other yield factors. The 2016 cable market therefore consumed 425 million km of optical fibre.
Large-scale fibre operations are making and drawing preforms that are about 20 cm in diameter. The drawn glass fibre has an outside diameter of 125 micrometres (µm), less than one hundredth of the preform’s diameter. An acrylic coating on the glass fibre brings the “finished” fibre’s outside diameter up to 200 µm or 250 µm for most telecom applications. This acrylic coating is applied on the draw tower, before the finished fibre is proof tested and reeled up at bottom of the draw tower.
In 2014, the Hengtong group of China published a paper describing its development of six-metre preforms, with a diameter 20 cm. This preform weighs more than 400 kg, takes the better part of a week to draw, and yields 14,000 km of fibre. Other factories are turning out preforms typically between 1.5 and 3.0 metres in length, with diameters of 15 to 20 cm.
There are several processes for making preforms, but most use a chemical vapour deposition process to achieve the high-purity needed in the fibre core. The starting material for this process in most cases is SiCl4, which is sourced at grades of at least 99.99% purity. Some key suppliers offer six “nines” or 99.9999% purity. The material quality includes metal ion impurities characterized as on the order of one part per billion or less. The chemical vapour deposition process has some similarities to part of the process for making semiconductor “chips.” The chip process, however, has other similarities to printing – laying down chemicals in a pattern, rather than bulk deposition on a cylinder.
The SiCl4, supplied to fibre makers in liquid form, is bubbled into the preform machine with O2. A heat source, either a flame or an RF plasma, drives the reaction in which SiO2 is deposited onto the substrate and Cl2 is collected. The preform machine can be set up horizontally, like a lathe, to deposit the silica materials while the preform is spinning on the centre line of the cylinder, or in a vertical chamber, to “grow” the silica materials downwards at the bottom of the cylinder.
Over the years, fibre makers have pursued improvements in the fibre-making process in three main areas: deposition rate, preform size, and draw speed. Other enhancements that have been critical in recent years include the coating materials and processes, the fibre geometries, the mix of dopants, and the chemical processes in the reactor chamber. But the three areas of deposition rate, preform size, and draw speed have a big effect on the costs and profitability of making fibre.
With the worldwide fibre shortage that began in 2016, telecom fibre prices have gone up 10% to 20%, depending on the country. Delivery times have stretched from months into quarters. Some large “integrated” fibre and cable companies have enough internal fibre capacity to supply their cable-making operations. Other “independent” cable companies must buy fibre, and the shortage has caused some independent cable makers to find new sources.
As of early 2017, several companies have announced projects to build new preform factories or expand existing factories. There are projects underway in China, India, and the US. These projects will add 2,000 to 3,000 tonnes of preform capacity to the world’s total starting in 2017. Some of the new capacity may not be on-line until late 2018 or early 2019. In any case, the additional capacity could bring 60 million to 100 million km of new fibre capacity into the market.
Will this be enough to end the shortage? This depends on the trend in optical cable demand. Installations in 2016 consumed 425 million fibre-km. If this market continues to grow at 10% per year, then 90 million km of new capacity will be used up in two years. Currently China is more than half the world market. China’s demand in 2017 is not expected to grow 10% compared with 2016, and the extent of growth in 2018 is further in doubt. Thus, all eyes are watching China’s carriers to see what their demand trend will be.
If China’s market is flat after 2017, then there will be a good chance that the new preform capacity will push the world into an oversupply situation in 2018. This could result in lower prices and also more difficulty in achieving payback on some of the new preform capacity. In the longer term, there are signs that 5G mobile network construction and fibre-to-the-home projects can keep the market at high levels – above 425 million fibre km well into the next decade. In some of the 5G mobile networks, fibre will be used to deliver the “traffic” or signals to the mobile base stations and the radio heads.