Torrefactation is one of the technologies that promises a true commodity fuel
Topell Energy

Co-firing biomass in power plants that run primarily on coal is one way of trying to reduce the greenhouse effect. Consequently, turning biomass into ‘bio-coal’ is something of a holy grail for both engineers and business managers. And torrefaction could be the way to go, argue Mark Beekes and Marcel Cremers.

Utilities are facing major challenges in the coming decades. Energy policy envisions a transition to a sustainable energy supply, while ensuring security of supply. Therefore, current energy policy spurs utilities to improve the sustainability of their coal fired power plants.

Co-firing biomass is one of the major measures widely applied to reduce CO2 emissions. Since the mid-nineties, power plants designed to burn pulverised coal have additionally been firing organic materials, such as wood and agricultural waste.

However, coal fired power plants are originally not designed to process biomass, which limits the co-firing percentage to 5 or 10 per cent. With investments in dedicated supply chains and biomass pre-treatment equipment co-firing percentages of 25 to 50 per cent (thermal) have already been achieved.

From the fuel perspective, the ideal situation is to process the biomass so that its properties resemble those of coal. The main form of processed biomass currently in use is wood pellets – pelletised dry sawdust – because they provide a relatively clean fuel that is internationally available, easy to handle (free flowing capabilities, less dust emission) and relatively inexpensive to transport. Wood pellets work well in coal fired plants and are now regarded as a well-proven technology.

Nevertheless, wood pellets do have their drawbacks. Wood pellets need dedicated silo storage to avoid degradation. Co-firing wood pellets has consequences for the milling and combustion of the wood pellets. At >5 per cent co-firing, the pellets need to be hammer milled to a typical particle size of no more than to 1 mm, whereas the coal mills grind the coal to a pulverised coal particle size of about 50 microns on average.

Co-firing may also influence primary air requirements, combustion behaviour, heat transfer pattern in the boiler, boiler efficiency, by-products and emissions. The various problems mean that wood pellets are not really a commodity fuel that you can blend with coal in whatever proportions you like.

A true commodity fuel

Utilities are now seeking innovations that could increase the co-firing percentage further. Torrefaction is one technology that, by creating a biomass product with superior handling and co-firing capabilities than wood pellets, holds promise for realising the dream of a true commodity fuel.

Torrefaction is essentially a biomass cracking technique. It is an additional pre-treatment step that heats the biomass to 260-320 ºC for up to one hour in an atmosphere with no or little oxygen. After torrefaction the biomass is brittle, due to the disintegration of hemicelluloses and to a lesser extent lignin and celluloses, responsible for the tough fibre structure. In other words, the fibrous structure of the biomass is partially broken down. The weakened structure improves milling properties and enables the biomass to be processed with coal at the power plant.

Turning biomass
Turning biomass into a product suitable for co-firing can cut coal’s environmental impact
Topell Energy

Furthermore, the calorific value of the biomass increases typically from 12-16 MJ/kg to 18-22 MJ/kg, due to the loss of volatiles and moisture. The product has lost its ability to absorb moisture. The features of torrefied biomass enable co-firing rates of more than 50 per cent of generating output, while keeping investments for co-firing biomass at a minimum. Depending on the distance from biomass source to the co-firing site, it is economically attractive to pelletise the torrefied biomass. Torrefaction pellets have a volumetric energy density of 14.5 – 17.5 GJ/m3 (bulk density of 800 kg/m3), which is about 70-80 per cent higher than conventional wood pellets (8.5 – 10 GJ/m3). In order to pelletise, the torrefaction temperature must stay below 300 ºC to keep intact a large part of the lignin, which serves as a natural binding agent for making pellets. Biomass that has been torrefied at higher temperatures might need additives to produce good quality pellets. Once their hydrophobic nature is proven, pellets can be stored in the open air – doing away with the need for silos. We also consider the feasibility of using particles that are larger than 8 mm pellets.

Torrefaction of biomass had already developed in the seventies and eighties. After a quiet period, the market of biomass then started to grow more rapidly at the beginning of this century. A number of small equipment suppliers with different technical processes started to torrefy biomass in pilot plants, producing small quantities that proved woody biomass can be torrefied.

Indicative fuel properties

Torrefaction is now attracting more and more attention. Enormous amounts of energy and entrepreneurship are being devoted to this processing technique. Biomass suppliers, investors and end users are all starting up projects. Numerous consortiums featuring investors have been formed. About 30 projects are now running, mainly in Europe and North America. Although most projects are pretty small scale, some larger ones are also getting off the ground. The best known torrefaction unit is Topell in Duiven, the Netherlands, which is designed to output 60 000 tonnes of product each year.

Several torrefaction reactors are being developed in parallel. It is too soon to say which approach is going to prevail. The suppliers of torrefaction technology are in different stages of developing a commercial scale torrefaction installation. KEMA has made an inventory of the existing concepts for torrefaction and evaluated them on their technological performance. Almost all suppliers have developed an integrated concept, in which the energy efficiency is optimised by combusting the volatile-rich torrefaction gases and by using the heat of the flue gases to dry and torrefy the biomass. No single technique is fundamentally superior to the others and several approaches will ultimately prove successful. The idea is to have a process that can be managed easily – cracking is an extremely complex business and not just one step on from drying.

An integrated approach is essential. You have to think not only about the reactor itself, but also about the drying, the milling and the heat recovery. If the material is not pre-processed properly, that has implications for how the reactor works. For example, the pre-drying step is crucial for good torrefaction conditions. Higher moisture contents of the biomass will result in ‘wet’ torrefaction gas, which requires energy to combust and lowers the overall energy efficiency. KEMA has also seen seasonal aspects play a role.

It is a mistake to look at everything from a purely technical viewpoint: the most economical solution must also be found. For instance, where the biomass comes from makes a big difference to the viability of a scheme. As does whether one need to create something from scratch, or if a torrefaction unit can be added to an existing plant.


Various torrefaction concepts exist. All concepts have been tested on at least a pilot scale size. Some concepts are currently being implemented or have already been implemented in a torrefaction plant. The typical size of realised plants or plants under construction is about 20-60 kt/year in product output. Apart from the upscaling challenges, all suppliers of torrefaction technology struggle to find feasible solutions for several issues.

Overview of reactor technologies and associated companies

The co-firing rate will still be limited by the chemical composition of the biomass, because components such as alkalis (K, Na, Ca, Li, Mg), phosphor and chlorine will still be present in the biomass after torrefaction and affect boiler integrity (corrosion, fouling), by-products and emissions.

Site-specific bottlenecks will usually be present, and may include dust emissions, health and safety issues, operational limits of primary air fans, operational limits of the coal mills, and shifting of the heat balance in the boiler. Using the KEMA CoFiring Control model, and extensive experience in co-firing at specific sites, KEMA is well positioned to calculate and predict site-specific bottlenecks, including:

  • Flue gas cleaning: In order to avoid permit problems, additional flue gas cleaning is needed after combustion of the torrefaction gas. An alternative would be to inject the torrefaction gas into a coal fired boiler to completely oxidise all organic compounds.
  • Process control: The challenge is to control the biomass feed, torrefaction temperature and retention time so that all biomass is completely torrefied without being carbonised.
  • Fuel flexibility: European and national legislation is restricting biomass available for co-firing. A different type of biomass will change the process conditions significantly and thereby also the choice of optimal reactor technology and integrated concept.
  • Sustainability: Concepts with relatively low efficiencies and relatively high emissions will fall off.

To conclude, we foresee that torrefaction will play an important role in co-firing biomass at coal fired power plants. At the moment, torrefaction technology is making its first careful steps towards commercialisation, while the technology and product quality are still surrounded by uncertainties.

The performance of torrefaction is highly dependent on the pre-treatment of biomass. Moreover, a large part of the added value of torrefaction will be allocated before the power plant gate, and can be calculated with the KEMA BioCase software.

Mark Beekes and Marcel Cremers are consultants at DNV KEMA.

Torrefaction reactors

Torrefaction concepts differ in reactor technology, torrefaction conditions and heat exchange methods.

Multiple Hearth Furnace (MHF)

The MHF consists of six hearths, which are each about 1 metre in height. The biomass is fed at the top of the reactor, after which it moves down through the different levels in the reactor. The hearths can be identified as either ‘IN hearths’ or ‘OUT hearths’. An IN hearth passes the biomass to the next hearth by moving the biomass to a centralised passage. An OUT hearth processes the biomass to the next hearth by moving it to drop holes which are at the periphery of the reactor. To process the biomass through the different hearths, a centralised shaft is used, which drives rabble arms at each hearth of the furnace. In torrefaction, the reactor is operated down draft, which means the flue gas flow follows the same direction as the product flow.

Rotary kiln reactor

The rotary kiln process resembles the concept for commercially successful pyrolysis units. When the rotary kiln reactor is applied for torrefaction, the biomass needs to be dried, preferably to 10-15 per cent moisture by weight. Several different concepts are based on rotary kiln technology. In one concept the rotary kiln is indirectly heated by thermal oils. In another concept the rotary kiln is directly heated by super heated steam.

In torrefaction, the rotational speed of a rotary kiln is a crucial process parameter for product quality. When the rotational speed is too slow, the biomass will be carbonised instead of torrefied. When the rotational speed is too high, the biomass is not fully torrefied and has low product quality.

Moreover, the rotational speed of the rotary kiln reactor has a wearing effect on the biomass, leading to a reduction in particle size over the reactor’s length. Variations in particle size should be avoided in a rotary kiln. The basic reactor technology has no capacity to differentiate in particle size, which means that particle size variations are critical for product quality.

Torbed reactor

The principle of a Torbed reactor is the toroidal flow of the bed, which is created by injecting air with high velocity (50-80 metres/second) through stationary angled ‘blades’. The injection angle results in a flow with a horizontal and vertical velocity vector, which lifts and moves the fuel bed in a horizontal motion at the same time. This creates a shallow solid material bed, which circulates around a vertical axis in the centre of the reactor and around a horizontal axis in the freeboard of the reactor. The toroidal motion allows a higher air speed, which reduces the boundary layer between solid particles and gases. As a result, heat and mass transfer between gases and solids improves, which allows lower retention times and a more homogeneous product.

Compact moving bed reactor

In a moving bed reactor the biomass is fed in at the top and moves slowly down to the bottom where the product is discharged. The length of the reactor is largely determined by the retention time needed to produce the desired product. When applied to torrefaction, the retention time is 25-30 minutes. In a moving bed reactor the biomass is directly heated to 250-300 ºC by partial recycling of torrefaction gases. The remaining gases are combusted in an afterburner and the heat of flue gases is directly fed into the torrefaction gas recycle stream. The recycling consists of repressurisation of the torrefaction gas to compensate for the pressure drop in the recycle-loop and of the heating of the recycle gas to deliver the required heat demand. A potential issue in moving bed reactors is the unequal heating of the fuel bed, due to limited mixing possibilities.

Screw conveyor reactor

The screw reactor is heated by the flue gases after combustion of the torrefaction gases, just as in the other concepts. However, heat transfer in a screw reactor is less efficient than fluidisation technology and, due to the transport capabilities of the screw, the biomass feed is limited to particles with a size smaller than 10 mm. Moreover, biomass with a very low bulk density and high moisture content needs to be pre-treated before feeding it to the screw reactor. In order to have a good product quality, the screw diameter is limited, which limits the upscaling potential.