The final format of the database contribution on thermal processing was agreed as in Table 7.1 with staff at Hendyplan and the EC project officer at the time and 20 sets of data provided in hard copy and disc copy for incorporation into the master database. Copies of this data is included as an annex to this chapter.

Table 7.1 BIS Thermochemical Database Methodology & Preliminary Worksheet

NO. PRIMARY FIELD CONTENTS

Secondary field

Tertiary field

1 ORGANISATION

Name

Contact person Address Telephone Fax

2 KEYWORDS

3 OBJECTIVES

4 CONVERSION PROCESS TYPE

5 REACTOR TYPE

6 PERFORMANCE

Efficiency Capital cost Problems

7 PROCESS

Description Objective

Conversion efficiency Capital cost Scalability

8 FEEDSTOCK

Main feedstock used Analysis Characteristics Cost

Other feedstocks used

Analysis Characteristics Costs

9 PRODUCT GAS

Yield, wt %

Analysis Quality Clean-up

Methods Efficiency Main use

continued

In spite of repeated requests directly to Hendyplan and representations via the EC a review copy of the final database was never supplied.

7.3 CONCLUSIONS

The Biomass Information System as proposed was an ambitious project that required considerably greater resources and time to be successful. The collation and management of extensive data such as was requested requires skilled and knowledgeable scientists and computer experts as well as solid experience at managing extensive and complex information, crystal clear objectives and a degree of ruthlessness in managing the project. Unfortunately none of the requisite criteria was available. Although the final product has not been seen, it seems likely that it has not met many of its original objectives.

A good database is an invaluable tool for planners and researchers, but it requires an ongoing commitment to first establish it with sufficient resources under experience and sound management, then needs critical review to ensure it meets the objects of the users, and finally needs ongoing support to maintain it and update it as such databases quickly become outdated.

There are four thermochemical methods of converting biomass: pyrolysis, gasification, liquefaction and direct combustion. Each gives a different range of products and employs different equipment configurations operating in different modes. These are summarised below in Table 1.1.

The basis of a fuel or chemical production system is that the feedstock is converted to a useful primary energy product and either used as such, or further converted, upgraded or refined in subsequent processes to give a higher quality and higher value secondary product as shown in Figure 1.1.

When organic materials are heated in the absence of air, they degrade to a gas, a liquid, and a solid as summarised in Figure 1.1. It is possible to influence the proportions of the main products by controlling the main reaction parameters of temperature, rate of heating, and vapour residence time. For example fast or flash pyrolysis is used to maximise either the gas or liquid products, depending on temperature as summarised below:

• Slow pyrolysis at low temperatures of around 400°C and long reaction times (which can range from 15 minutes to days in traditional beehive kilns) maximises charcoal yields at about 30% wt.

* Flash pyrolysis at temperatures of typically 500°C; at very high heating rates and short vapour residence times of typically less than 1 second or 500 ms; maximises liquid yields at up to 85% wt (wet basis) or up to 70% dry basis.

• Similar flash pyrolysis at relatively high temperatures of above 700°C; at very high heating rates and similarly short residence times maximises gas yields at up to 80% wt. with minimum liquid and char production.

* "Conventional" pyrolysis at moderate temperatures of less than about 500°C and low heating rates (with vapour residence times of 0.5 to 5 minutes) gives approximately equal proportions of gas liquid and solid products.

This study is focussed on pyrolysis for the production of liquid fuels by the relatively novel process of fast pyrolysis, as these are currently viewed as a promising process and a promising product both in Europe and North America. The pyrolysis liquids are variously referred to as "bio-oil", "bio crude oil" or even as "oil" although they share few similarities with any oil products.

In order to appreciate the scientific and technical complexities of flash pyrolysis, the development of understanding of reaction mechanisms and pathways is first reviewed which will explain some of the unusual product properties that have been reported and some of the reasons for the way the technology has developed. Modeling of the complex and interactive physical and chemical processes that occur in pyrolysis has also attracted considerable attention and this area is also reviewed to at least partly explain why technology and scale up is still largely empirical. Developments of some of the process technologies that have been promoted and scaled up are subsequently described to show the underlying principles have been applied.

Water is variously claimed to be completely miscible at up to 30% or 55% by weight of total liquid, above which an aqueous layer separates. Any water that does separate must be carefully managed as it will be heavily contaminated with dissolved organics and require extensive treatment before disposal. Utilisation and consideration of oil on a "wet" basis, therefore, seems to be more sensible. An alternative approach is to condense the liquids at a temperature above the dew point of water i. e. above about 110°C. This has been used successfully by Bio — Alternative in their continuous carbonisation unit (128) but produces a very dirty gas which has to be burned or incinerated immediately. Roy has also used this approach to preferentially separate fractions from his vacuum multiple hearth pyrolyser (129), and early work at Waterloo also employed hot and cold water condensers (18). Indirect cooling has the disadvantage of increasing the residence of the vapour at temperatures where further reaction can still occur, thus impairing the product quality and yield. As processes increase in scale, this effect would become more pronounced. The current approach is for rapid quenching of the total product stream for maximum recovery of the liquid fraction

Water content is important as it has several effects: it reduces the heating value, affects the pH, reduces the viscosity, influences both chemical and physical stability, reduces potential pollution problems from waste water disposal and could affect subsequent upgrading processes (130). The interactions are poorly understood. The water is difficult to measure and remove, since evaporation or distillation at normal temperatures of around 100°C can cause significant and potentially deleterious physical and chemical changes in the liquid. Lower temperature drying is not successful due to the nature of the relationship between

water and the organic component in which the water seems to be chemically combined, analogous to water of hydration.

A key feature of flash pyrolysis processes for liquids is that no discrete aqueous phase is produced as all the water of reaction and feed water is incorporated or dissolved in the product bio-oil. This water is thus incinerated when the oil is combusted, and there are no environmental or pollution implications.

From slow pyrolysis processes, however, water is produced in significant quantities of typically between 20 and 40% wt on the feed, depending on feed moisture content. If a liquid product is collected from slow or conventional pyrolysis units, then the maximum water load of the liquid product is around 20 % wt. Above this level a discrete aqueous phase separates. This water phase is highly contaminated with dissolved and suspended organics, with a COD of typically 150000 (125, 126). This therefore represents a major problem of disposal or utilisation. In the selection of the primary pyrolysis products this waste water must be considered. If biological treatment is not appropriate or too expensive, part of the heat of combustion of the products will be required for incineration of this heavily contaminated water fraction. A potentially more attractive alternative route than incineration is oil condensation above the dew point of water, i. e. about 110- 120°C. The water then stays in the vapour phase and can be burned with the product gas (128). The pyrolysis gas should primarily be used for this purpose but this may not be enough in cases where the primary feedstock has a high water content and the gas is required for feed drying. Since slow and conventional pyrolysis is now only considered for charcoal production or for disposal of difficult wastes, this waste water problem is not significant in the context of bio-energy.

5.17.1 Summary

The objective of this project was to investigate the potential of low pressure pyrolysis of biomass to produce high yields of condensable vapours and selectively condense fractions from different sections of the reactor to examine the potential for fractionation and recovery of chemicals. The reactor is a 30 kg/h multiple hearth furnace operated under vacuum. A range of feedstocks have been tested of which tyre pyrolysis was subsequently successfully scaled up to 250 kg/h. A notable feature of the vacuum type pyrolysis plant was handling whole tyres. Recent work has concentrated on operating the system for waste disposal. A system is believed to have been sold in Switzerland in 1994.

5.17.2 Description

A 30 kg/h vacuum pilot plant multiple hearth reaction system with liquid

condensation and collection was designed, constructed and tested (74). Figure

5.15 shows the arrangement of equipment. The optimum temperature range was found to be between 350-400°C and a yield of 60 wt % (on a dry ash free wood basis) of pyrolytic oil was obtained at an average heating rate of 10°C/min and at a total system pressure between 0.3 and 2.3 mm Hg (40-307 Pa). The feedstock used was aspen poplar. There is an extensive literature of which some key references are quoted (75, 76, 77, 78, 79). See also (3) for a more comprehensive overview and extensive referencing. Wood chips with a particle size from 1/4" to 1/2" Tyler mesh (6 mm to 12.7 mm) are fed via a hopper on the top of the reactor, which is hermetically sealed. This is equipped with a variable rate feeding device that feeds the chips into the preheated reactor at a constant feed rate of between 0.8 to 4 kg/h.

The reactor is a multiple hearth furnace 2 m high and 0.7 m diameter, with six hearths. Electric heating elements are used to heat the reactor. The temperatures of the heating plates increase from top to bottom of the reactor. A typical temperature profile is 200°C to 400°C. At steady state conditions, the absolute system pressure of the system is less than 80 mm Hg (10.7 kPa). The organic vapours and gaseous products are removed from the reactor by a mechanical vacuum pump via six outlet manifolds which correspond to the six heating plates. The char falls to the bottom of the reactor where it is collected in a metallic jar on a load cell. The process unit is connected to a central microprocessor which simultaneously gathers data and controls about 75 operating parameters.

The clean-up system is a series of shell and tube heat exchangers (primary condensing unit) and a train of receivers (secondary condensing unit). Each outlet manifold is connected to a heat exchanger where the vapours are condensed and recovered as organic liquid in individual receivers. Cool to warm tap water is used as the cooling medium on the shell side. The vapours from the heat exchangers are then collected in the secondary condensing unit where the aqueous phase is primarily recovered. The first receiver is immersed in a bath of water-ethylene glycol mixture. The next two are immersed in baths of dry ice-acetone while the final receiver is filled with glass wool at room temperature. The non-condensable gases are pumped into a 500 litre vacuum vessel.

The low pressure removes the primary products quickly and avoids secondary reactions. Fractionation provides some separation of liquids evolved at different temperatures in a continuously operating system, but this has not proved as effective as had been hoped.

Recent testing of this unit at throughputs of 30 kg/h showed that the primary condensing unit composed of six shell and tube heat exchangers were inefficient due to clogging problems. The six individual exchangers were substituted by a single spray type condensing unit, similar to the secondary condensing unit. The new system proved to work very satisfactorily.

A major design consideration is the large volume of equipment due to operation at low pressure and the high cost of maintaining vacuum operation. Scale-up will have to consider the optimum operating pressure and the problems of heat transfer in the multiple hearth furnace.

5.17.3 Products

Some results are shown in Table 5.15. The highest yields of bio-oil are obtained at the lowest pressure and the higher temperature conditions. The optimum temperature range for maximum oil yield from wood was found to be between 425- 450°C at the bottom of the reactor.

One potential advantage of using a multiple-hearth reactor configuration is the capacity to fractionate the pyrolysis products by use of multiple outlets at different levels (see Table 5.16). The separation of the aqueous and the oil phases is important at the industrial level because the recovery of chemicals during distillation of large amount of water is less economical.

A typical analysis of the oif is shown in Table 5.17. This oil is highly oxygenated and consists of phenols, sugars and both aliphatic and aromatic hydrocarbons. The gases are mainly CO and CO2.

For small particles the kinetics are sufficient to predict the reaction rate. However, for large particles, both the physical and the chemical changes are essential for obtaining a global pyrolysis rate. To formulation an analytical pyrolysis model, the known parameters that can influence the pyrolysis process must be considered. These affect the energy and mass flows into and out of a pyrolysing particle given by the following methods:

* heat transfer from the reactor environment to the particle surface by convection, and/or radiation and and/or conduction;

* heat transfer from the outer surface of the particle into the interior of the particle by conduction and in a few situations to a lesser degree by convection;

* convective heat transfer between the volatile reaction products leaving the reaction zone and the solid matrix ;

* primary pyrolysis leads to conversion of the biomass to gas, char and a primary liquid product;

* secondary and tertiary pyrolysis leads to conversion of the primary product to a gas, char and a secondary liquid product which then forms primary and secondary products;

* changes in physical properties, enthalpy and heats of reaction of the biomass

* changes in the enthalpy of the pyrolysis products;

* diffusion of volatiles out of the solid and away from the particle surface. Pressure gradients may also occur due to vapour formation in larger particles.

Process Parameters that can influence pyrolysis are given below in Table 2.3 with effects.

These processes are all temperature dependent and, since temperature changes with time and space, they will also be time and spatially dependent. Furthermore, they will also be dependent on the physical structure of the particle along with its properties such as density, thermal properties, size and the orientation of the particle with respect to grain.

Fast pyrolysis requires high heating rates. As biomass has a low thermal conductivity, high heating rates throughout the particle either result from small particle sizes or erosion of the char product from around the particle as it pyrolyses as occurs in ablative pyrolysis (see Chapter 2 and 3). The current view of the maximum size feasible for fluid bed or circulating fluid bed or entrained flow reactors for giving high yields of liquids is 6 mm. The exception is ablative pyrolysis which is a surface area controlled process and has no upper size limit as the char layer is constantly eroded or abraded away. This upper particle size limitation for most processes thus carries a cost and energy penalty to grind the feedstock. Some biomass forms exist in a stringy form which cannot be ground in conventional milling equipment. Chopping or slicing mechanisms may work, but extensive testing is needed to evaluate alternative comminution processes. Low bulk density feeds give handling problems and fluidisation problems in bubbling fluid bed reactor systems. Feed that is pretreated such as by acid washing has reduced mechanical strength and gives high fines level which can result in higher solids levels in the product.

As all the feed moisture appears in the product together with all the reaction water, the feed has to be dried to as low a water content as possible. Practically this limit is around 10% moisture. At normal high yields of liquids of 60 to 65 % on a dry liquid basis, the reaction water contributes about 15% moisture to the wet product from a bone dry feed. Each 1% moisture in the feed adds approximately a further 1.25% moisture to the product oil on a wet basis. Drying to much below 10% is generally considered too expensive as well as significantly increasing the fire and safety hazards from handling very dry materials. Forced drying is probably the most expensive pretreatment operation for the feedstock from most primary energy crop sources.

High lignin feeds give higher viscosity products that are also more "sticky" due to the presence of higher levels of partially depolymerised lignin. Optimisation of the time-temperature processing window may improve the product quality and a small sacrifice in yield may be more than compensated for a significant improvement in quality or properties. Resinous feeds give analogous product handling problems from tar products resulting from resin degradation.

All biomass contains some ash, usually appreciable quantities of alkali metals which report to both the char and the liquid. This is compounded by the presence of char in the liquid from incomplete char removal in orthodox cyclones. Although developments are in hand to filter char out in the vapour phase, analogous to hot gas filtration in gasification processes, a proven system is not yet available and preliminary results have suggested that product yields may be adversely affected from the higher vapour residence times.

The crude primary liquid product from flash pyrolysis contains very high levels of oxygen which give the product some unusual characteristics including water miscibility and temperature sensitivity as described above. While the oil can be used directly in a number of thermal and power generation applications, this is considered to be unsuitable in some demanding applications such as turbines and engines and substitution for transport fuels, so upgrading to more conventional liquid hydrocarbon fuels has been examined.

There are currently two basic processes: modified conventional hydrotreating to a naphtha-like product can be characterised as CH2; or zeolite cracking and synthesis to a highly aromatic product which can be characterised as CH1.2 . Recent developments have been reviewed by Elliott et al. (3), Sharma and Bakshi (60) and Milne and Soltes (61) and future needs identified (62,63).

3.4.1 Product quality

The complex interaction of time and temperature on liquid product quality has not

been explored, at least partly because the characteristics of pyrolysis oil for different applications has not been defined and there is no "standard" pyrolysis liquid. It is this definition of oil quality that is a major uncertainty and requires to be defined by potential users and may differ by application. While there are set standards and methods of measurement for conventional fuels, analogous standards and methods have not yet been defined for biomass pyrolysis liquids. Density, viscosity, surface tension and heating value are known to be typical key properties for combustion applications in boilers, furnaces and engines; but other characteristics such as char level and particle size and ash content may have a major effect.

3.4.2 Unusual characteristics

Bio-fuel-oil has a number of special features and characteristics which require consideration in any application including production, storage, transport, upgrading and utilisation. These are summarised in Table 3.3 below. The conclusion is that some problems are soluble and some are more intractable but none are insoluble. Recognition of problems and awareness of potential difficulties is a major contribution to dealing with any of the above difficulties.

CPERI have been investigating the extraction and upgrading of phenolic components in fast pyrolysis liquids for several years. A small fluid bed pyrolyser has been operated to produce liquids (16, 17, 18) from which phenols are extracted then converted into methyl aryl ethers as possible fuel additives (19, 20, 21). The focus of the work has been in chemicals recovery and production and this is discussed in more detail in Chapter 6

Production of primary liquids by fast pyrolysis of biomass has been commercialised at a small scale in Canada and has been actively and extensively developed throughout Europe. Some of the early problems with liquid quality are being successfully addressed and rapid progress is being made in producing a liquid that will satisfy most applications. There are many aspects of the pyrolysis system that require addressing as well as in the development of more innovative, more efficient and more lower cost reactors. Of particular importance is minimisation of solid char and ash in the product through vapour or liquid phase processing and improvements in liquid collection. In addition the quality of the product is important for different applications and work is needed to both measure these properties against application requirements and provide appropriate norms and standards. There is a need for better co-ordination between the research institutions and companies involved in this area to ensure that problems, solutions and developments are fully and sensibly exploited to more quickly bring the technologies to commercial fruition. This will be partly addressed in the new IEA programmes and the PyNE Concerted Action project.

Upgrading of pyrolysis oils to transport fuel and higher quality energy products by hydrotreating is very expensive and unlikely to be viable in the short to medium term, as well as having a number of technical problems relating to catalyst stability that require resolution. Work should however continue to ensure that the science is sufficiently well understood so that the technology can be developed when required. Zeolite upgrading and analogous in-situ or close-coupled catalytic upgrading is less well developed but offers more short term promise, as a separate process is not required, there is no hydrogen requirement and costs would probably be little different from fast pyrolysis. This is an area that deserves more consideration and support. Other innovative possibilities for upgrading should also be encouraged.

Chemicals recovery is potentially important from the novel chemicals that might be recovered the high added value of speciality chemicals and the contribution that these would make to an integrated biomass based conversion system. There is a need for continued evaluation and exploration of products and applications. Development of catalytic pyrolysis systems will add to the interest of this approach.