Exploratory studies on wire mesh fast pyrolysis of biomass has been carried out to derive fundamental kinetic data and compare this approach to that of ablative pyrolysis at the University of Aston (q. v.) (86). The advantage of this approach is that very small quantities of biomass can be processed under very high and controllable heating rates which makes derivation of basic reaction kinetics more reliable. There is little data available.

5.20.1 Summary

The objective of this research is to convert biomass such as sewage sludge, agricultural wastes and refuse/MSW into fuels or raw materials for the organic chemicals industry as an alternative solution to landfill and incineration of sewage sludge (87,88). Several batch and continuous laboratory scale plants up to 5 kg/h have been built and tested. The principle is to use low temperatures of less than 350°C and long reaction times of up to an hour to achieve low oxygen content oils and high yield of fine chemicals. The concept has been licensed to several organisations in Europe (Stenau q. v.), North America (Wastewater Technology Centre q. v.) and Australia (Wastewater Technology Centre q. v.) and several plants of up to 2 t/h are planned or operating based on sewage sludge.

5.20.2 Description

Figure 5.17 shows one configuration of the auger reactor system based on dried sewage sludge.

At a laboratory scale, both batch and continuous screw reactors have been used. Sludge dewatered to about 20% wt solids, or other biomass, is heated slowly to 300-350°C in an oxygen free environment for about 20 minutes, and the liquid product collected in an ice-cooled bath. No additives are needed as the silica, silicates and heavy metals present in the sludge are claimed to act as catalysts. The vapour is condensed and collected. Feedstocks tested include sewage sludge, rape, lupine and Euphorbia.

Four types of equipment have been used on a larger scale for low temperature conversion. There are: a rotary furnace (up to 80 kg/h), a fluidised bed, (up to 400 kg/h), a reactor with a transport belt for MSW conversion (up to 200 kg/h) and a cone screw converter for the conversion of agricultural wastes.

5.20.3 Products

Oil yields ranging from 18-27 wt% (feed basis) and char yields from 50-60 wt % (feed basis) have been achieved. Table 5.19 shows the elemental analysis of the products.

Liquids with very low oxygen content (less than 5% wt oxygen) have been reported from a sewage sludge feed. The low oxygen level and chemical intermediates are claimed to be due to low reaction temperature, natural catalysts in the feed and slow reaction times. The oils contain aliphatic hydrocarbons and fatty acids as the main components (89).

The sewage sludge derived oil has been reported as being used as fuel for a diesel engine.

modelling have been summarised in the previous section. The main benefits of the

studies carried out over the past 25 years have been:

1 optimisation of the process parameters to allow the prediction of yields and specification for the production of chemical intermediates.

2 kinetic data has been obtained which aids in the prediction of reaction times and ultimate yields.

3 reaction pathways allow the fast/flash pyrolysis process parameters to be specified for the production of speciality chemicals in significant yields.

4 reactor design and specification is now much simpler due the more data on relevant physical properties becoming available, allowing better design methods.

The known problems with the liquid product that may require attention are suspended char, high viscosity, alkali metals, and potential inhomogeneity of the oil. The more significant are discussed below while a more comprehensive summary is given in Table 4.4.

4.4.3.1 Char

Direct combustion of the bio-oil is probably the least demanding application and a variety of tests have established that combustion does not present any significant problems. Some problems to date include char particles which cause filtration and pumping problems, atomisation problems and "sparklers" in combustion giving poor burn-out rates and high viscosity which contribute to the handling and pumping problems. More severe problems can be anticipated with atomisation for injection into turbines and engines where precise control of flow rates is essential.

The optimum solution to char suspensions interfering with bio-oil injection and atomisation is to reduce it at source in the vapour phase before condensation. Some success has been achieved with this approach by Ensyn, but only for limited periods of time. There is clearly considerable potential for development in this area.

Char filtration from the liquid bio-oil is known to be difficult due to the nature of the suspension, the high viscosity and the unusual relationship between the char and heavier components of the bio-oil. Dilution with water adversely affects heating value and may cause phase separation. Addition of diluents such as methanol or ethanol reduces viscosity and improves filterability but at a financial and energetic penalty. Heating is not advisable due to the temperature sensitivity of the bio-oil

with some changes occurring at around 50°C and significant modification of the oil above 100°C. Pressure filtration is effective for short periods but costly due to the unusual relationship between the char and heavier components of the bio-oil, and may result in unacceptably high bio-oil losses.

Table 4.4 Characteristics of Bio-oil and Methods for Modification (її)

Characteristic Effect

There are complementary measures that may be taken such as reduction of the injection pressure into the boiler or engine/turbine, thus creating a less arduous environment and greater tolerances for the injector system. This will require careful matching with the atomiser and application characteristics. Also as application size increases, this problem reduces as more bio-oil is required to be injected and nozzles are increased in diameter.

For some applications, such as gas turbine fuel, complete de-oxygenation may not be necessary to give a suitable fuel with acceptable chemical and physical properties. This concept of partial hydrotreating has not yet been properly explored. For example, raising the LHSV (liquid hourly space velocity) from around 0.1 to 0.2 or 0.3 will give an oxygen level of around 4 or 15% respectively compared to 0.5% oxygen at 0.1 LHSV (116). The typical oxygen content of a flash pyrolysis oil is about 35-40%. The uncertainty lies in the selectivity of hydrotreating and hydrogenation of different classes of compounds in the crude bio-oil and the properties of the resultant partially upgraded oil. Some exploratory work has, however, been carried out (96) including studies on model compounds (75, 101).

A 15% oxygen content product would have a similar oxygen content as pressure liquefaction products which have very high viscosity and limited water miscibility under ambient conditions. This is, therefore, relatively difficult to use in most conventional applications. A partially hydrotreated pyrolysis oil has not been produced, however, so it’s characteristics cannot be defined. If there is a separate water phase, this would be highly contaminated with soluble organics and would require extensive and costly treatment. However, if this approach were successful in producing a stable and directly usable product, the hydrogen consumption could be reduced by 60% and the capital cost reduced by about 30%. In addition, the product may either retain the water in a solubilised form, or a relatively clean water could be separated. The concept of stabilisation by partial hydrotreatment deserves further consideration.

NREL have been developing their 36 kg/h vortex ablative pyrolyser since 1980. Designs for scale up have been produced and support has been provided to Interchem for the reactor in their 31 te/d demonstration plant (see below). A second R&D unit has now been constructed and a small fluid bed fast pyrolysis unit has also been commissioned. Zeolite upgrading of the pyrolysis vapours to aromatics has been accomplished at the fundamental level on the MBMS system, on a slip­stream of the vortex reactor and on the full reactor output. Novel multi-functional zeolite catalysts have been produced and tested. A circulating fluid bed zeolite cracker with regenerator is under construction. Only limited results have been published on any of this work including recent work by Czernik et al. (13) on vapour phase thermal treatment to lower the oxygen content of the liquid.

3.5.1J3____ Interchem

The 31 te/d demonstration plant was moved from Mountain View Missouri to Kansas City and the reaction system completely redesigned with the support and assistance of NREL (14). The construction of the new reactor was completed in late summer 1993 and work has now stopped on the project. Future plans are unclear.

3.5.1.4 BBC

A 10-25 kg/h unit was constructed to investigate the reactor parameters of surface temperature and gas/vapour product residence time and particle size. Although the unit was built to pyrolyse tyres, it is capable of using wood, MSW and other similar materials as a feedstock. Limited details of the continuous ablation reactor are available (15). Particles up to 6 mm have been used with liquid recovery in a two stage direct liquid quench of the product vapours with recycled liquids after char removal in a high temperature cyclone. Liquids are collected in a sump tank and are returned to quench the hot vapours via a water cooled heat exchanger. Liquid yields of 54% have been obtained from tyres at 470-540*C at 0.88 s residence time and 1.3 mm size particles (16). A 35-50 t/d plant has been built in Nova Scotia where it is operating successfully with combustion of the raw product gas and no liquids collection.

5.7.1 Summary

The aim of this process was the production of bio-oil by the flash pyrolysis of fine wood particles in a 200 kg/h vertical, down flowing entrained flow pilot plant shown in Figure 5.6 (24, 25). The Egemin plant was commissioned in October, 1991, part funded by the EEC from the Joule programme. The research was finished in 1993 and the project was abandoned due to tack of funding and less promising results. A basic problem of this plant was obtaining a sufficiently high rate of heat transfer in the short residence time needed for high liquid yields. This was one of two entrained flow reactors that have been researched (with GTRI), and abandoned.

Other work has been carried out on the pyrolysis of hemicellulose but this has received less attention due to its lower abundance, variety of constituents, high reactivity and rapid degradation at low temperatures (150-350°C). It is believed that the intermediate levoglucosan is replaced by a furan derivative (16, 19). This may be due its lack of crystallinity (22). Soltes and Elder (23) suggest a two step degradation process where the first step is depolymerisation to water soluble fragments subsequently followed by decomposition to volatile components.

2.2.2 Lignin Pyrolysis

The complex structure of lignin has led to a lack of understanding of the pyrolysis of this component. Lignin is the most thermally stable component but its structure varies according to its source and the method of isolation. To date therefore, most detailed work in lignin pyrolysis has been obtained from model compounds. Minor decomposition begins at 250°C but most significant lignin pyrolysis occurs at higher temperatures (13,18, 22). High molecular weight compounds such as

2-6

coniferyl alcohol and sinaptyl alcohol are formed during the initial stage of pyrolysis by the formation of double bonds in the alkyl side chain of the lignin structure.

Low temperature pyrolysis of lignin (< 600°C) has been carried out by a large number of researchers (16,18, 19, 24, 25, 26, 27, 28, 29, 30). Detailed work using Kraft lignin has also been carried out by Jegers and Klein (31, 32) who identified and quantified 33 products (12 gases, water, methanol, and 19 aromatic compounds such as phenol, creso! and guaiacol) at a range of temperatures from 300 to 500°C. latridis and Gavalas (33) studied the pyrolysis of kraft lignin at 400- 700°C using a captive sample reactor, obtaining a total volatiles yield of 60 wt%. Nunn et al. (34) have also carried out work in this area obtaining a maximum of 53 wt% liquid at 625°C, again in a captive sample reactor.

High temperature pyrolysis of lignin (> 600°C) leads to complex cracking, dehydrogenation, condensation, polymerisation and cyclisation reactions resulting in the formation of products such as СО, CH4, other gaseous hydrocarbon, acetic

acids, hydroxyacetaidehyde and methanol. Polyaromatics, benzene, phenylphenols, benzofurans and naphthalenes are formed by other secondary reactions (9, 16, 18, 19, 25, 34, 35).

Other work has been carried out with model compounds and mathematical models to obtain reaction mechanisms and reaction kinetics for lignin pyrolysis. Klein and Virk have proposed a reaction mechanism derived from the pyrolysis of the model compound phenethylphenyl ether (18, 32, 33).

Mathematical modelling of lignin pyrolysis has been attempted using the Monte Carlo technique (36, 37). The overall simulation is comprised of two Monte Carlo simulations: one comprising the lignin structure and the second the degradation of its oligomers. The simulation contained model compound reaction pathways and kinetics in a Markov-chain based simulation of the reaction of lignin polymers which subsequently produced yields of various hydrocarbons and oxygenated compounds. Other mathematical models have been developed by Solomon (25, 38) to predict the molecular weight distribution of the tars and Anvi (24, 39) who predicted the rate of evolution of lignin pyrolysis gases.

3.7.1 Introduction

A summary of the opportunities for using pyrolysis liquids is included as Figure 1. Within Europe the most promising application is seen as electricity production due to the anticipated ability to use raw bio-oil as produced in an engine or turbine without the need for extensive upgrading as well as the ability to de-couple fuel production from electricity generation with storage and/or transport of the liquid fuel which is not possible for gasifier products and IGCC systems. There are believed to be substantial longer term opportunities for chemicals from either catalytic pyrolysis or extraction and recovery of chemicals from the liquid products which would enhance the overall economics of conversion.

This liquid product may be readily burned (3) and has been employed for this purpose (13), provided the viscosity is not too high (Huffman Interlaken). Preheating to reduce viscosity is not usually favourable due to thermal degradation of the bio-oil, although short term exposure to temperatures up to 90°C have not had any adverse effects. The water content can be considered an advantage both for the combustion process and because it reduces the viscosity of the liquid. Therefore the oil can be considered an outlet or disposal route for the pyrolysis water. Phase separation is only likely to occur at water concentrations greater than 50% in the case of flash pyrolysis oils which is unlikely to arise. Some precautions may be needed in handling, storage and combustion due to the water and high oxygen content as described above.

Pyrolysis liquids are immiscible with any form of hydrocarbon liquid, and cannot, therefore, be expected to be assimilated into a conventional fuel marketing infrastructure without some conversion or upgrading to give a product that is compatible with conventional fuels. One alternative is to feed to crude pyrolysis liquid into a refinery for upgrading in orthodox refinery operations, utilising the hydrogen availability and blending opportunities (23). The basic problem with this approach is that bio-oil is immiscible with crude oil and creates handling problems. The alternative is to create a discrete pyrolysis liquids storage, distribution and utilisation system, that is managed by experts who understand the special problems of this fuel.

3.7.2 Combustion

Liquid products are easier to handle and transport in combustion applications and this is important in retrofitting existing equipment. Existing oil fired burners cannot be fuelled directly with solid biomass without major reconstruction of the unit, which may not be attractive in uncertain fuel markets, however bio-oils are likely to require only relatively minor modifications of the equipment or even none in some cases. Problems have been reported in handling the fuel with high viscosity and suspended char causing atomisation difficulties and incomplete combustion with some fuels. With recognition of the problems, solutions are unlikely to prove onerous.

5.21.1 Summary

This work was instigated under the JOULE programme to develop a new reactor technology for the pyrolysis of biomass by sliding and pressing the particles on a heated surface. Ablative pyrolysis is achieved in a rotating cone where the particles "slide" across a heated metal surface. The process as developed and described has been licensed to BTG and a larger plant of 50 kg/h is being built by Schelde for delivery to China in 1995.

5.21.2 Description

The concept is that biomass particles are fed onto an impeller which is mounted at

the base of the heated rotating cone and are then flung on to the heated surface with the final char and ash residue flung out of the top of the cone. The concept is depicted in Figure 5.18.

An initial cold flow experimental cone was built to investigate the dynamic behaviour of nearly spherical mono-sized PVC particles, diameter 140-780|im (91). This work was carried out in a cone of angle 60° [vertical height 0.31 m and cone top width of 0.5 m] and 90°. Rotational speeds up to 1800 rpm were used to determine the influence of gas flow on the residence time and motion of the particles. Particle motion was recorded using an endoscopic camera. Particles greater than 400pm appeared to be unaffected by gas viscous forces and the residence time of the particles is almost independent of the particle diameter. For particles smaller than 200pm, viscous forces become dominant and the residence time of the particles is strongly dependent on the particle diameter. Derived particle residence times were 0.01-0.3 s (92).

A heat transfer rig was then tested with the same cone dimensions to investigate the heat transfer to particles ablatively "sliding" on the cone surface (93, 94). Measurement of particle temperature as it travels along the wall of the cone has bee achieved by fluoroptic methods. The experimental reactor is depicted in Figure 5.19. Initial experiments showed that the fine biomass particles [-250 pm] adhered to the cone surface, thereby reducing the reactor throughput quickly.

Rice husks with a high ash content [-20 wt%] have flowed successfully through the reactor zone with no surface adhesion. Modifications have been carried out to the cone to add sand to promote the flow of solids in September 1992 (95). Initial experiments were carried out mainly with a cone temperature of 600°C and a cone rotational speed of 900 rpm.

The reactor interior has been modified as shown in shown in Figure 5.20 to reduce

the operational volume from 0.25 m3 to 0.003 m3, otherwise the gas/vapour residence time would be around 80 s giving significant vapour cracking (96). To remove the problem of particle adhesion to the reactor wall, sand is added in a mass ratio of 10 or 20:1 sand to biomass fed. Preheated sand is used on a once through basis. The reactor is to be modified to permit internal sand recycle. The reactor outside the cone quickly becomes filled with sand and char, restricting experimental runs to 10 minutes. The liquids are collected in a series condenser system.

The reactor is to be modified so that the sand is removed from the reactor with the char, the char combusted and the hot sand re-fed to the reactor, i. e. an internal sand recycle. A cold model has been constructed at the beginning of 1995 and this will subsequently be tested as a hot system.

5.21.3 Products

Isothermal reactor operation leads to significant cracking of the product vapours. Typical yields at 1 s residence time and a heated surface temperature of 600°C are: 50 wt% liquids, 30 wt% gases and 15 wt% char. Reactor throughputs are claimed to be 7.2 kg/h of biomass.

3.1 INTRODUCTION

This chapter shows how the underlying scientific principles and requirements can be applied to the design and operation of pyrolysis processes. Included are the key product liquid characteristics that result from practical application of pyrolysis.

3.2 PYROLYSIS CONCEPTS

It has already been explained that biomass is a mixture of hemiceflulose, cellulose, lignin and minor amounts of other organics which each pyrolyse or degrade at different rates and by different mechanisms and pathways. Lignin decomposes over a wider temperature range compared to cellulose and hemicellulose which rapidly degrade over narrower temperature ranges, hence the apparent thermal stability of lignin during pyrolysis. The rate and extent of decomposition of each of these components depends on the process parameters of reactor (pyrolysis) temperature, biomass heating rate and pressure. The degree of secondary reaction (and hence the product yields) of the gas/vapour products depends on the time-temperature history to which they are subjected to before collection which includes the influence of the reactor configuration. Although some research has been carried out on the individual components of biomass, most applied and larger scale work has focused on whole biomass as the cost of pre-separation is considered too high. In addition, the separation and recovery of pure forms of lignin and hemicelfufose are difficult due to structural changes in their processing, although pure cellulose is relatively easy to produce.

Research has shown that maximum liquid yields are obtained with high heating rates, at reaction temperatures around 500’C and with short vapour residence times to minimise secondary reactions. Fast pyrolysis processes have been developed for production of food flavours (to replace traditional slow pyrolysis processes which had much lower yields) and speciality chemicals which utilise very short vapour residence times of typically 100-300 ms and reactor temperatures around 500*C. Both residence time and temperature control is important to "freeze" the intermediates of most chemical interest in conjunction with moderate gas/vapour phase temperatures of 400-500’C before recovery of the product to maximise organic liquid yields.

Liquids for use as fuels can be produced with longer vapour residence times [up to ~6 s] and over a wider temperature range although yields might be affected in two ways: secondary gas decomposition at temperatures above 500’C and condensation reactions at gas/vapour product temperatures below 400’C. Most woods give maximum liquid yields of up to 80% wt% dry feed basis [64 wt% organics and 16 wt% water] at 500-520*C with vapour residence times not more than 1 second. Very short residence times result in incomplete depolymerisation of the lignin due to random bond cleavage and inter-reaction of the lignin macromolecule resulting in a less homogenous liquid product, while longer residence times can cause secondary cracking of the primary products, reducing yield and adversely affecting bio-oil properties. Evidence from SEC analysis of the

liquids would suggest that the reactor configuration and the dominant mode of heat transfer strongly influences the average molecular weight of the products (1). This is discussed further below.