Effect of Thermal Treatment on Combustion Process of Spruce Wood ( Picea abies )

This paper deals with the effect of thermal treatment of spruce wood on its burning process. Samples of 100 mm x 100 mm were dried out and then heat-treated at 150 °C, 200 °C, 250 °C and 300 °C. Thus prepared samples were tested on a cone calorimeter at a heat fl ow of 50 kW·m-2. The obtained results clearly show that with the increasing temperature of the treatment, the released heat from the surface unit decreases and the unit weight increases. The carbon dioxide concentration in the fl ue gas increases as well, however, the optical density of the smoke decreases substantially. It can be stated that the burning of spruce wood is considerably affected mainly by the thermal treatment at temperatures above 200 °C. From the point of view of the use of spruce wood as fuel, the most optimal treatment temperature is 250 °C.


INTRODUCTION 1. UVOD
The treatment of wood by thermal load can be divided into three basic areas on the basis of the required output raw material.Drying is intended to reduce wood moisture.It is usually performed at a temperature of up to 115 °C (Edvardsen and Sandland, 1999; Hansson and Antti, 2006;Zarea Hosseinabadi et al., 2012;Sehlstedt-Persson, 1995).The second area is thermal modifi cation.
Modern thermal modifi cation processes are limited to temperatures no higher than 260 °C.As a result of thermally induced chemical changes to the macromolecular constituents, the physical and biological properties of wood are altered (Hill, 2007).Finally, wood is used for the preparation of torrefi ed wood.This semi refi ned material is an intermediate between
Thermal decomposition of wood can be described using a fi ve-reaction mechanism, consisting of three devolatilization reactions for the pseudo-components -hemicellulose, cellulose and lignin, and of two additional reactions in the air for char devolatilization and combustion (Broström et al., 2012).
Heat-treatment affects the mechanical properties of wood, whereby hardwood species are more susceptible to this process than softwood.The decay resistance of heat-treatment may be achieved but at a high cost and with reduced mechanical properties (Kamdem et al., 2002).Heat treatment mainly resulted in a darkening of wood tissues, improvement of the dimensional stability of wood and reduction of its mechanical properties (Bekhta and Niemz, 2003).Increased heat treatment and resulting weight loss reduce the modulus of rupture (MOR) and modulus of elasticity (MOE) (Mburu et al., 2008).Živković et al. (2008) state that the results of laboratory tests show that the heat-treated wood, when compared to genuine wood, except of improvements in dimensional stability, also exhibits a lower equilibrium moisture content in room conditions.Results of Mburu et al. (2007) showed that resistance to fungi and termites was greatly improved by the treatment.The acidity and wood-water contact angle are higher and polar component of surface free energy is lower after thermal modifi cation (Miklečić and Jirouš-Rajković, 2016).
A mild thermal treatment (modifying at 160 -180 °C in oxygen-poor atmosphere) leads to clear changes of the measurable acoustic characteristics, such as Young's modulus, damping and sound velocity, so thermally modifi ed wood is a material with favourable characteristics for making musical instruments (Pfriem, 2015).
As a solid energy carrier, biomass generally has a few disadvantages, which limits its use for coal replacement and as a feedstock for entrained fl ow gasification.The hydrophilic and fi brous nature, low calorifi c value and low bulk energy content imply high accumulated costs in the whole supply chain and severe challenges in more advanced conversion systems.By thermally pre-treating the biomass by torrefaction, these properties may be signifi cantly improved (Strandberg et al., 2015), so torrefaction is a promising technique for improving the biomass performance for energy utilization (Tapasvi et al., 2012).
In contrast to fossil fuels, biomass has a unique potential for making positive environmental impact.In the plan of the sustainable biomass production and use, the carbon dioxide emitted would be absorbed by newly grown biomass.It can be burnt without emitting large amounts of nitrogen oxides, and with low emissions of sulphur dioxide (Quaak et al., 1999).
During torrefaction, the main thermal decomposition reactions involve the hemicellulose polymers, resulting in improved fuel properties exhibited by the torrefi ed samples.Compared to raw samples, the com-position of the torrefi ed samples is closer to that of coal, with higher carbon content and a lower volatile matter content, and with much higher hydrophobicity (Tapasvi et al., 2012).The carbon contents of torrefi ed wood are greater than those of wood, but lower than those of charcoal.Carbon increases at the expense of oxygen and hydrogen, thus leading to decreases in both H/C and O/C ratios (Pentananunt et al., 1990).
In comparison to non-thermally treated wood, torrefi ed biomass fuels contain a lower amount of volatile matter and a higher amount of fi xed carbon (Ndibe et al., 2015).
Both torrefaction temperature and reaction time have strong effects on the torrefaction process, but temperature effects are stronger than effects of reaction time (Tran et al., 2013;Felfl i et al., 2005;Pimchuai et al., 2010).
It was demonstrated that the weight loss of thermally modifi ed wood is mostly the result of the reduction of the polysaccharide fraction, while hemicelluloses degraded faster than cellulose during heat exposure.The acid-insoluble lignin content increased with the severity of the treatment at the expense of the carbohydrate component.Upon thermal decomposition of spruce wood with increasing temperature, the amount of lignin and the amount of extracts increased, but the abrasive layer had a decrease in extracts.At the same time, the polysaccharide fraction degraded signifi cantly, the average polymerization degree of cellulose decreased where the crosslinking reactions occurred.The results of Calonego et al. (2016) show that the thermal modifi cation of wood causes signifi cant increases in the net calorifi c value, and the extractive content, and significant decreases in the holocellulose, galactose, xylose and glucose contents.
At higher temperatures, wood shrinks in the transversal plane, due to volatilization of wood constituents and due to a slight densifi cation of the cell wall substance (González-Peña et al., 2009).
The torrefaction process can also be applied to wood-based fuels.In terms of torrefi ed product properties, the torrefi ed samples absorb approximately onethird of the moisture compared to the raw fuels, and the total grinding energy decreases up to 40 -88 % (Tapasvi et al., 2012).
Degradation of hemicelluloses, cellulose and lignin and removing of moisture from the material, have a strong effect on the pelletizing properties of biomass.The friction in the press channel of a pellet mill increases, resulting in high pelletizing pressures that increase the energy uptake of the mill and might result in a decrease of capacity and in worst case in overheating of a blockage of the mill press channels (Stelte et al., 2011).The common practice in pelletizing of thermally untreated biomass, using water to decrease energy consumption and to improve bonding properties, is not applicable in pelletizing of torrefi ed materials (Larsson et al., 2013).The composition of torrefi ed briquettes at 220 °C does not undergo many changes.
However, at higher temperatures, changes in the composition are perceptible, with the briquette hemicellulose practically degraded and cellulose depolymerisation process initiated (Felfl i et al., 2005).
According to the report of the Ministry of Agriculture and Rural Development of the Slovak Republic and the National Forest Centre, in the year 2016, 9.3 mil.m 3 of wood was produced, of which 55.2 % was wood of coniferous trees.During this period, the production of spruce was 61.6 % of coniferous trees.A large amount of spruce timber was signifi cantly affected by the activation of harmful biotic agents (Moravčík et al., 2017).Usually such a contaminated wood is not suitable for construction or furniture purposes and it can be used for energy purposes.Therefore, this paper deals with the possibilities of thermal treatment of spruce and its infl uence on the burning of the resulting material.

MATERIJALI I METODE
For the measurements, 5 sample pieces of tangentially split spruce wood (Picea abies (L.) Karst.) with dimensions of 100 mm x 100 mm and 20 mm thickness were used.The samples were fi rst dried to zero humidity at 102 °C for 48 hours.Subsequently, four of them were further thermally treated at 150 °C, 200 °C, 250 °C and 300 °C in a preheated furnace.The heat treatment was carried out for 4 hours in a muffl e furnace under an inert atmosphere (nitrogen fl ow rate of 200 ml•min -1 ).Finally, the samples were transferred to an exsicator, where they cooled down to ambient temperature.The total time of procedure was approximately 55 hours (48 hours for drying, 4 hours of thermal treatment at appropriate temperature, approximately 3 hours for cooling).The percentage mass loss of samples due to their thermal treatment is shown in Figure 1.With the increase of treatment temperature, the weight loss of the sample also increases.Ramos-Carmona et al. (2017) attribute this phenomenon to hemicellulose and cellulose decomposition and to oxidation reactions.Barta-Rajnai et al. (2017) also describe strong decrease of the volatile extractive content.Since the spruce samples were thermally treated in a nitrogen atmosphere, the oxidation reactions did not occur and, therefore, it was possible to predict the effect of decomposition of the mentioned wood components.
The measurements were made on cone calorimeters meeting ISO standards 5660-1 (2015).Although a higher number of samples is required for the measurement of wood materials, due to the heterogeneous nature of wood, for cone calorimeter measurements the use of one sample for one heat fl ux is common, as evidenced by other authors.The samples were placed in the horizontal position under the emitter and initiation of the combustion was provided by an electric spark initiator.The measurement conditions are shown in Table 1.The time to ignition of the samples, the amount of heat released, the concentrations of carbon oxides in the fl ue gas and the amount of smoke released were observed.

REZULTATI I RASPRAVA
The rate of heat release (Figure 2) is very similar for most samples.There is almost no oxidation of the sample material before initiation of fl ame burning.Volatile fl ammable substances are released, while the concentration of these substances in the mixture with air does not reach the value required for their ignition.This fi rst phase takes a relatively short time, which can be attributed in particular to high external heat fl ow.As soon as the rate of release of the fl ammable fuel is suffi cient, initiation occurs.Due to rapid homogeneous burning, the rate of heat release increases sharply.Once the sharp peak has been reached, a hardened layer begins to form on the surface of the samples.This acts as an insulator and partially restricts the heating of the   2009) state for the wood that the second peak occurs when the thermal wave reaches the back-insulating surface and the original material has already been pre-heated to the pyrolysis temperature, thereby effectively reducing the heat of the pyrolysis.
After the fl ame burning, the rate of heat release decreases sharply.There are stages of smouldering and glowing, in which the heat is released mainly by the oxidation of carbon residues.The same course of burning in cone calorimetry testing has also been reported by Janssens (1991).
As the temperature of the thermal treatment increases, the value of both the fi rst and the second peak decreases (Table 2).This decrease was also noted by Martinka et al. (2016) for spruce timber treated by different processes.Due to the high susceptibility of hemicelluloses to thermal degradation compared to lignin described by various authors (Mburu et al., 2008;Mburu et al., 2007;Calonego et al., 2016), this phenomenon can be attributed to the gradual thermal decomposition of hemicelluloses and cellulose in wood during the preparation of samples.In this way, the proportion of volatile fl ammable fuel, which is a fuel for homogeneous combustion, is declining.Burning of the specimen heated at 300 °C has a specifi c course, where both the fi rst and the second peak do not reach any sharp maximum and only slightly differ from the steady burning phase.
In the samples treated thermally at temperatures of 250 °C and 300 °C, relatively faster rate of combustion starts to show up, which has also been stated by Pentananunt et al. (1990) for wood treated at similar temperatures (250 °C -270 °C).
In terms of the initiation phase, an important parameter is the average rate of heat release in the fi rst 60 seconds.For a spruce panel measured on a cone calorimeter, Östman et al. (1985) indicate a value of 172 kW•m -2 .The dried spruce was measured with a very similar result at a level just above 166 kW•m -2 (Figure 3).It is clear from the other measurements that the fi rst-minute average heat release rate decreases, with  the temperature growth of the thermal spruce treatment, which is mainly due to the declining heat release rate.The decrease in the fi rst peak of the heat release rate is also directly related to the decreasing maximum heat release energy (Table 2), which is an important indicator for the fi rst phase of fi re propagation.
The treatment temperature had a strong effect on total smoke release (Table 3).While dried wood released a smoke during the fl ame corresponding to an optical density of 345 m 2 •m -2 , the wood torrefi ed at 300 °C no longer showed any smoke.Similar results were described by different authors.Pentananunt et al. (1990) indicate that the torrefi ed wood showed significantly less smoking during combustion compared to untreated wood.Felfi et al. (2005) state that, from the combustion point of view, decreases in O/C and H/C ratios are favorable since less smoke and water vapour are formed, improving the performance of briquettes and contributing to energy loss reduction.The release of less smoke can be attributed to a gradual change in the soot formation mechanism in samples undergoing the process of thermal treatment (Mitchell et al., 2016).
Results of individual measurements are shown in Table 2.The time for sample ignition does not show a large diffusion in the specimens modifi ed to 250 °C.The value of 8 s to 9 s is comparable to that reported for spruce by Östman et al. (1985) Initiation of the sample modifi ed at 300 °C occurred slightly later than in the previous samples.Due to the very low peak of heat release rate and the visual observation at which the fl ame did not differ from other samples, it is clear that the amount of volatile combustible in the sample was small.Initiation of volatile fl ammable mixtures with air was, therefore, delayed.
Mean heat release rate per unit area and hence a total heat release does decrease with the increasing temperature of the treatment.However, since the heat treatment changes the weight of the samples, the mean effective heat of combustion increases.This result is consistent with the calorifi c value of wood due to thermal treatment (Calonego et al. 2016).At the same time, the yield of carbon dioxide and carbon monoxide increases.This is probably due to the higher carbon content of thermally modifi ed wood compared to the orig-inal material described in the literature (Pentananunt et al., 1990).

ZAKLJUČAK
The previous thermal treatment in an inert atmosphere has a notable infl uence on burning spruce wood.The infl uence occurs partly already before the sample initiation, which is higher compared to other specimens in the case of the sample modifi ed at the temperature of 300 °C.The heat release rate reaches lower maximum values with the increasing adjustment temperature and its course is getting balanced.Intensity as well as fl ame burning time decrease as a result of decreasing fl ammable fuel content.However, in the case of thermal treatment, the weight of the sample also decreases, resulting in an increase in effective heat of combustion.The burning rate showed a sharp increase in samples treated at 250 °C or higher.In terms of fl ammability, the increase in the temperature of the heat treatment was confi rmed by its decrease.In the case of torrefi ed wood, this is almost a hundredfold change, which can be considered very positive when used as a fuel.It can be concluded that the notable change in the burning of the treated wood is caused mainly by temperatures of thermal treatment above 200 °C.The highest amount of heat released per unit mass, substantially lower total smoke release, and the most uneven heat release during burning was observed in the sample thermally treated at the temperature of 300 °C.However, since the heat release rate in this sample reaches values very close to the external thermal fl ux applied to its surface, from the point of view of the use of spruce wood as a fuel, the preferred treatment temperature appears to be 250 °C.

Figure 1 1 .
Figure 1 Dependence of the percentage mass loss on the temperature of thermal treatment Slika 1. Ovisnost postotnoga gubitka mase o temperaturi toplinske obrade

Figure 3 3 .
Figure 3 Infl uence of thermal treatment temperature on the fi rst-minute average heat release rate in combustion of spruce wood Slika 3. Utjecaj temperature toplinske obrade drva na brzinu otpuštanja topline u prvoj minuti izgaranja smrekovine

Table 1
Measurement conditions for individual samples rest of the sample.At the same time, it partially prevents the penetration of volatile fl ammable substances into the burning area.Both of these mechanisms cause a decrease in the rate of heat release up to the phase of combustion stabilization.This area is visible in the graph in the form of a constant course.Then, it is followed by the second, considerably lower and wider peak.Hagen et al. (