A 200 kg/h fluid bed reactor derived from the University of Waterloo work was originally constructed in 1986 by Encon for wood and peat pyrolysis. The unit was constructed on the back of a trailer for transportation to test sites for demonstration. This is now used for processing old telegraph poles, These have their outer skin of treated wood removed and which is then pyrolysed in a 50 kg/h fluid bed pyrolyser for recovery of chemicals including creosote and PCP as well as bio-oil. Up to 30,000 poles per year are processed. The skimmed poles are re-treated and recycled (17).

3.5.1.6 ____ Other activities

More fundamental studies are being carried out in several laboratories including the University of Hawaii (18) and the University of Montana (19).

3.5.2 Europe

In Europe the situation is more varied but less developed with only one substantia! pilot plant. Most interest has been stimulated by the EC JOULE and AIR programmes over the last 8 years.

3.5.2.1 CPERl (Greece)

A small fluid bed reactor has been operating for several years to provide pyrolysis liquids (20) from which phenols and related chemicals are recovered for production of oxygenates such as methyl aryl ethers for use as gasoline additives (21).

5.9.1 Summary

This project was initiated in 1980 based on early research which showed that higher oil yields could be obtained from a rotating tube furnace with moving particles (typically 28 wt% liquid yield) rather than a stationary reaction bed (typically 17 wt% liquid yield) (38). From this work, an entrained flow bed reactor was designed for the production of pyrolysis oil. This operated successfully until around 1989. Although apparently successful, the process was never scaled up. The only other entrained flow system for fast biomass pyrolysis was built by Egemin (q. v.). GTRI also developed the tech-Air process which similar to that of Bio- Alternative (q. v.) in that the principal product was charcoal with appreciable quantities of secondary oil derived as a by-product. This material was used extensively in early upgrading work.

5.9.2 Description

A process research unit was built and completed in 1983 (39, 40, 41). In 1985, modifications were made so that optimisation of the oil yield could be further investigated, based on experience gained in the operation of the pilot plant and the results of the commissioning runs. The major changes were the replacement of the 8 in. diameter (20.32 cm) reactor tube with a 6 in. (15.24 cm) tube, the addition of a quench vessel and a second demister (42, 43, 44). The nominal operating feed rate was then 56.8 kg/h of dry biomass. Figure 5.9 is a flowsheet of the process.

The feed is dried, hammer-milled to about 1.5 mm and fed from a loss-in-weight feeder into the reactor via a rotary valve. The reactor used is a 6 in. inner diameter vertical tube made of stainless steel. The initial feed point was the refractory lined mixing section, located below the reactor tube. However, by introducing the feed into feed ports higher up the reactor, the effective length of the reactor could be reduced which in turn reduced the residence time. The wood particles are entrained in a stream of hot combustion gas (927°C) obtained by burning propane gas and air stoichiometrically. Gas and wood flow cocurrently upwards through the reactor tube in which pyrolysis takes place; the resulting mixture consists of non­condensable gases, water vapour (moisture plus pyrolytic reaction water), pyrolysis oil vapours and char.

A cyclone separator is used to remove most of the char particles. The exiting gas stream consists of non-condensable gases, water vapour, pyrolysis oil vapours and some char fines. The hot effluent enters a water-sprayed quench vessel where it is rapidly cooled. Following the quench vessel, the mixture enters an air-cooled condenser in which the pyrolysis vapours are condensed with some water vapour. Early problems were reported with accumulation of tarry material in the first stages of the air-cooled condenser.

The condensed phases are removed via sumps and receivers and the gaseous product is passed through two demisters connected in series. Most of the aerosols present in the gaseous product are removed in the demisters. The remaining effluent, consisting of non-condensable gases, water vapour and remaining
aerosols, enters a flare where it is burnt and the combustion products are exhausted to the atmosphere.

In the scaled up process, it was intended that waste water production would be minimised or eliminated by controlled cooling and condensation of oil to retain the water in the vapour phase with subsequent combustion of the water laden by­product gas for process heat. There was, however, no experience of internal gas recycling in the pilot plant.

5.9.3 Product

Table 5.10 shows some of the last results with liquid yields approaching 60% wt on feed (45). Modelling and optimisation studies produced predictive models which indicated that yields of 70 wt% would be achievable with a well designed reactor and system. The oils are highly oxygenated with no phase separation as shown in Table 5.12. They have an initial boiling point range from 70°C to 90°C. They are heat sensitive and will decompose when heated to temperatures greater than 185°C-195°C, The oils are acidic, have an acrid odour and also exhibit corrosive properties with some metals. A typical bio-oil analysis is shown in Table 5.11. The product was upgraded by hydrotreating at Battelle PNL (46) which is discussed in more detail in that chapter. The liquid was also upgraded in the liquid phase over zeolite cracking catalyst — the only known application of this approach to zeolite cracking which is usually carried out on freshly produced vapours. The results have not been published, but extensive coking is understood to have occurred.

INETI

Fluid bed pyrolysis was investigated by LNETI (now INETI) at two scales of operation using hot gas to effect heat transfer. Initially work was carried out on a 10 cm fluid bed with subsequent research on a 30 cm square fluid bed with top feeding. Several parameters were investigated including pressure, temperature and bed additives or secondary reactants including zeolites, zinc chloride, carbonates and alkali (48). The results were not very satisfactory in that pyrolysis yields were relatively low and the catalysts showed little activity.

More recently a reaction scheme has been proposed by Evans and Milne as shown in Figure 2.5 (16). Similar to the Copper Mountain reaction scheme, the influence of "pyrolysis severity" is considered as the increase in temperature, heating rate and vapour residence time. Under high pressure conditions, the direct formation of a liquid product is due to the wood deforming to a "plastic" state. This has been confirmed by Diebold (49) and Lede et al. (50, 51,52) who have shown that under condition where ablation occurs, the wood may exhibit the properties of a "molten plastic state”. At low pressure, however, it is not clear whether a liquid phase exists, after the primary decomposition of the biomass.

It can be seen that the initial models for the pyrolysis of the main components of wood have developed through to an overall reaction scheme as proposed by Diebold, and Evans and Milne. Evans and Milne do, however, consider a discrete tertiary reaction stage. The process of pyrolysis is complex, but a recent theory is that primary vapours are first produced, the characteristics of which are most influenced by heating rate. These primary vapours then further degrade to secondary tars and gases, the proportions and characteristics of which are a function of temperature and time (53). Yields of liquids from pyrolysis can thus be influenced by the rate of reaction, with fast or flash pyrolysis at lower temperatures of typically 450-650°C giving the highest liquid yields.

Several hundred chemical constituents have been identified to date, and increasing attention is being paid to recovery of individual compounds or families of chemicals. The potentially much higher value of speciality chemicals compared to fuels could make recovery of even small concentrations viable. The opportunities are described in Chapter 5

An integrated approach to chemicals and fuels production offers interesting possibilities for shorter term economic implementation. Chemicals that have been reported as recovered include polyphenols for resins with formaldehyde, calcium and/or magnesium acetate for biodegradable de-icers, levoglucosan, hydroxyacetaldehyde, and a range of flavourings and essences for the food industry. There are substantial problems to be overcome is establishing markets for the less common chemicals and devising low cost and efficient separation and refining techniques. The onfy currently viable market opportunity is for speciality food flavourings with a current market size of around US$ 10 million in North America and which could expand to US $ 40 million in 10 years. A similar market opportunity is believed for Europe (45).

VAPOURS AND LIQUIDS

6.1 INTRODUCTION

In all thermochemical conversion processes, catalysts are being increasingly used to enhance favourable reactions and inhibit unfavourable reactions. In addition, catalytic processes are both available and are being developed to upgrade the primary products of thermal conversion to higher quality and higher value fuels and chemicals as summarised in Table 6.1 for a range of thermally processed products. This study focuses on fast pyrolysis processes and liquid products as a means to increase the value of resultant products at minimal cost. While reference is made to many of the research activities involving the use of catalysts in biomass conversion (see, for example, 1), attention is focused on particular problems requiring short term solution to enable the benefits of these technologies to be realised.

The high heat transfer rate that is necessary to heat the particles sufficiently quickly imposes a major design requirement on achieving the high heat fluxes required to match the high heating rates and endothermic pyrolysis reactions. Reed et al. originally suggested that to achieve true fast pyrolysis conditions, heat fluxes of 50 W/cm2 would be required but to achieve this in a commercial process is not practicable or necessary (6).

Each mode of heat transfer imposes certain limitations on the reactor operation and may increase its complexity. The two dominant modes of heat transfer in fast pyrolysis technologies are conductive and convective. Each one can be maximised or a contribution can be made from both depending on the reactor configuration. The penalties and interactions are summarised in Table 3.2 below with some speculations on heat transfer modes.

For ablative pyrolysis in a vortex reactor, a furnace arrangement equivalent to an ethylene cracking furnace has been proposed by the IEA Bioenergy Agreement pyrolysis and liquefaction group (7, 8). Other possibilities to achieve the pyrolysis temperatures and heat transfer rates necessary have included vapour condensation such as sodium, induction heating of the reactor wall and the use of contact electrical heaters. In a circulating fluid bed, the majority of the heat transfer wilt be from the hot circulating sand which therefore requires an efficient sand re­heating system. In a conventional fluid bed the sand requires an external heat source.

A commercial system would be expected to utilise the by-product char and gas for the process heat requirements as an integrated system as proposed for the NR EL wood to gasoline process evaluated by the IEA (7, 8).

Reactor type Suggested mode Advantages/disadvantaqes/features of heat transfer

5.1.1 Introduction

Alten stands for Alternative Energy Technologies which was a consortium of KTI and Italenergie, but is no longer in partnership. The aim of this activity was to develop a small scale pyrolysis process to convert wood and agricultural wastes into marketable fuel products such as fuel oil, charcoal and char-water slurries. It was envisaged that a network of small pyrolysis plants would produce bio-oil to fire a 27MWe power station in Avezzano. The project is important for being the largest plant in Europe and for providing valuable data on process design, optimisation, and qualitative results. Substantial quantities of oil were produced for testing

The pyrolysis plant was in operation from 1985 to 1989, and was, and still is, the largest biomass pyrolysis unit that has been dedicated to bio-oil production in Europe. The design capacity of the plant was 1 t/h dry biomass, but only up to 500 kg/h has been achieved on a continuous basis. The nature of the process was that relatively slow pyrolysis occurred giving a secondary oil that had a low water tolerance and high viscosity. Although the product quality caused problems, the availability of large quantities of pyrolysis liquids for testing and evaluation caused serious attention to be focused on direct liquefaction for the first time in Europe.

The hydrocarbons produced from both upgrading processes can, in principle, be used directly for some applications including firing in a turbine or an engine but require conventional refining to produce orthodox transport fuels. This would be carried out in a conventional refinery using established catalytic operations such as reforming, alkylation and hydrotreating (110 — 112, 164).

The yields of the various liquid products from both hydrotreating and zeolite synthesis have been estimated from current data and are summarised in Table 6.8 for crude pyrolysis liquid, three varieties of hydrotreated products — partial, complete and refined; and crude and refined aromatics from zeolites (165).

A circulating fluid bed fast pyrolysis system has been designed, built and commissioned. Fast pyrolysis is carried out in the riser of a CFB using reheated sand and hot combustion products from burning char in a bubbling fluid bed at the base of the riser (22).

Egemin (Belgium)

This 200 kg/h capacity entrained downflow system was commissioned in 1991 and operated at up to around 50 kg/h giving liquid yields around 45% wt. (4). it was closed down permanently in 1992 due to unpromising results from the unit and lack of further support from the EC. It is unlikely to be restarted,

3.5.2.4 ENEL (Italy!

A contract has been signed to purchase a 650 kg/h Ensyn RTP-III plant (transported bed reactor) employing recycle gas (23). This will be commissioned around September 1995. The bio-oil product will be tested in a variety of applications including firing in a power station boiler, upgrading by hydrotreating and firing both an engine and turbine.

3.5.2.5 I NET! (Portugal)

A 100 mm diameter fluid bed has been operated for fast pyrolysis at feed rates of 0.18 to 0.6 kg/h. The maximum total liquid yield at 500’C from pine was 51% which increased to 55% at 600’C. The liquid had characteristics similar to those from other fast pyrolysis systems. Tests have also been carried out on a 300 mm square fluid bed (24). Work is continuing on co-pyrolysis of biomass with plastics (25).

Work has recently started on small scale fluid bed fast pyrolysis of contaminated waste wood at around 100 g/h in order to devise a better disposal method for such wastes combined with energy recovery. There are no results as yet.

5.11 INTERCHEM INDUSTRIES INC., USA

5.12.1 Summary

Interchem Industries Inc. was founded in 1985. Initial work to scale up an ablative fast pyrolysis process was carried out by Pyrotech Corporation to solve an energy deficit problem at a pulp mill at Samoa, California (49, 50). In 1989 the company entered into a consortium to develop and exploit the commercial potential of the NREL ablative pyrolysis process (see above) to produce a phenol adhesive substitute and an alternative fuel (51).

Although there were some early attempts to demonstrate the principles of pyrolysis in a vortex type reactor, the first significant plant was built in Springfield, Missouri for the conversion of sawmill wastes. Construction of the 32.7 M6 facility for the production of fuel oil and charcoal, referred to as a Petroleum Synthesis Unit, was completed in September 1990 and this was tested until 1992 (52, 53). The plant was redesigned and relocated and a new reactor was built based on the NREL vortex ablative reactor. This second plant was never completed and the project is understood to have been abandoned.