Preliminary torrefaction of oil palm empty fruit bunch pellets

Torrefaction of pelletised oil palm empty fruit bunches (OPEFBs) is a promising pretreatment technique for improving its solid biofuel properties and energy recovery potential. Therefore, this paper investigates the torrefaction of OPEFB pellets to examine the effects of temperature and purge gas flow rate on mass yield (MY), energy yield (EY), and mass loss (ML). The results revealed that MY and EY decreased due to significant ML during torrefaction. Furthermore, significant improvements in the higher heating value (HHV) and energy density (DE) were observed. The torrefaction temperature increased liquid (tar) and gas yields mainly above 300 °C at the expense of solid products. However, the effect of purge gas flow rate on the torrefaction products was found to be negligible. Consequently, the torrefaction of OPEFB pellets were limited to 250-300 °C, 30 min, and nitrogen (N2) gas flow rate of 200 ml min-1.


Introduction
The growing global demand for crude palm oil (CPO) has increased the cultivation of oil palm (Elaeis guineensis Jacq.) worldwide [1]. In Malaysia, over 20 million tonnes of CPO are processed from over 75 million tonnes of fresh fruit bunches (FFB) harvested annually [2]. However, with an estimated oil yield of 22%-23%, CPO production generates large quantities of oil palm waste (OPW) annually from oil palm mills in Malaysia [3]. Over the years, the growing stockpiles of OPW have become an environmental burden arising largely from uncontrolled tipping or dumping, crude burning, and landfilling of OPW [4]. In addition, poor efficiencies, secondary wastes, and emissions from current conversion technologies have resulted in increased air, land, and water pollution [3].
The outlined problems are largely ascribed to the poor solid biofuel properties of OPW, particularly oil palm empty fruit bunches (OPEFBs). High moisture, heterogeneity, ash, and alkali content of OPEFB along with its low calorific value, energy density, and grindability have hampered efficient energy recovery [5]. Furthermore, poor fuel properties of OPEFB are responsible for operational problems such as sintering, agglomeration, and fouling of gasifiers and biomass boilers [6]. Nonetheless, the outlined challenges can be addressed by biomass pretreatment technologies such as pelletisation and torrefaction. The process of pelletisation compacts pulverised biomass into uniform solid fuel with high energy density and calorific value but low moisture [7]. Similarly, torrefaction improves the solid fuel properties of biomass such as hydrophobicity and grindability [8,9]. In principle, torrefaction is a mild pyrolysis process used to pretreat biomass at low pressures, heating rates (< 50 °C min -1 ), gas flow rates (< 1000 ml min -1 ), and temperatures from 200 to 400 °C either under inert or oxidative conditions [10,11].
However, previous studies that have examined the torrefaction of OPWs revealed low mass and energy yields [12,13] due to loss of interparticle bonding and overoxidation of loose biomass particles [10,14]. However, these challenges can be addressed by the torrefaction of pelletised OPEFB. This will likely address the problems of loss of interparticle bonding, rapid decomposition (overoxidation), and poor yield of pulverised biomass after torrefaction and poor quality pellets after pelletisation. Therefore, this paper seeks to perform a preliminary analysis of the torrefaction of OPEFB pellets in a horizontal fixed bed tubular reactor (FBT). It also examines the effects of temperatures (250-350 °C) and purge gas flow rates (50-200 ml min -1 ) on mass yield (MY), energy yield (EY), energy density (DE), and higher heating value (HHV) of torrefied pellets.

Experimental methods
Pelletised OPEFBs were acquired from a palm oil mill in Kota Tinggi, Johor, Malaysia. Subsequently, OPEFB pellets were characterised to examine the physicochemical, thermal, and kinetic properties as presented in our previous studies [15][16][17]. Based on the results, it was observed that OPEFB pellets required further pretreatment to improve its solid biofuel properties for efficient energy recovery [5]. Therefore, OPEFB pellets were pretreated through torrefaction from 250 to 350 °C (in 25 °C steps) under non-oxidative conditions (nitrogen (N2)) at 15 °C min -1 in a stainless steel FBT reactor as depicted in Figure 1. Heat was supplied to the reactor during torrefaction through a tube furnace (Model: Lindberg Blue M, USA). For each test, 15 g of OPEFB pellets was placed in the FBT reactor and purged with N2 for 15 min at the selected flow rate of 200 ml min -1 . After flushing was completed, the reactor and OPEFB pellets were heated under non-isothermal (dynamic) conditions from room temperature (RT) to the selected torrefaction temperature (250-350 °C). Next, the heating programme was switched to isothermal mode to maintain heating at the selected torrefaction hold time of 30 min. Upon completion of torrefaction, the tube furnace was switched off and the FBT reactor and its contents were cooled to RT. The torrefied OPEFB pellets were subsequently retrieved, weighed, and stored in airtight vessels prior to characterisation. The torrefaction process was evaluated based on the parameters of MY, EY, DE, and HHV, calculated from Equations 1.1 to 1.4 [18,19]. Where the terms mTB represent the mass of torrefied OPEFB pellets, mRB is the mass of raw OPEFB pellets, MY is the mass yield, EY is the energy yield, DE is the energy density, HHVTB is the higher heating value of torrefied OPEFB pellets (MJ kg -1 ), and HHVRB is the higher heating value of raw OPEFB pellets (MJ kg -1 ).

Fig. 1. Schematic diagram for torrefaction of OPEFB pellets.
Lastly, the effect of gas flow rates (50, 100, and 200 ml min -1 ) on the MY of torrefaction was examined at 300 °C for 30 min. The tests were performed to examine the effect of torrefaction purge gas flow rate on the torrefaction of OPEFB pellets in a stainless steel FBT reactor. All tests were performed in duplicate to ensure the accuracy and reliability of the measurements.

Results and discussion
The values of MY and ML for the torrefaction of OPEFB pellets in the FBT reactor under N2 gas from 250-350 °C are presented in Figure 2.  As observed in Figure 2, MY decreased from 72.61% to 32.58% during torrefaction. However, ML increased from 27.39% to 67.42%. The ML is due to the thermal degradation and devolatilisation of hemicellulose, lignin, and volatile matter during torrefaction [20]. The increase in temperature promotes the thermal decomposition of lignocellulosic bonds and volatile matter in OPEFB pellets during torrefaction. Consequently, the pellets undergo drying, devolatilisation, decarboxylation, and depolymerisation, resulting in the loss of moisture, volatiles, CO2, and lignocellulosic components, respectively, during torrefaction. The outlined thermochemical reactions improve hydrophobicity, grindability, and porosity of biomass after torrefaction [8,9]. Similarly, torrefaction improves the EY, energy density DE, and HHV of the torrefied biomass. Hence, the effect of OPEFB pellets torrefaction on the parameters was examined as presented in Table 1. In addition, the HHV of torrefied pellets is significantly high and comparable to lignite and sub-bituminous coals reported in the literature [21,22]. This indicates the torrefied pellets have potential synergic characteristics that could promote co-firing or utilisation in boilers. The decrease in EY is ascribed to the significant ML resulting from the removal of water, volatiles, and lignocelluloses during torrefaction. ML positively affects HHV, thus resulting in higher DE. Furthermore, the torrefaction process generated high liquid (tar) and gas yields, particularly above 300 °C. The tar produced was deposited on the walls of the reactor with traces were also found in the liquid product (Figures 3 (a-d)), as also reported in the literature [23,24]. Tar is an undesirable torrefaction product that affects biomass conversion efficiency, product selectivity, and the costs of production and cleaning due to the fouling of equipment. The formation of tar is typically attributed to the decomposition of cellulose and lignin [25,26], which are major sources of condensable vapour and tar compounds during torrefaction [27]. Therefore, it is imperative to limit OPEFB pellets torrefaction to 250-300 °C to selectively yield HHV and energy dense solid products as opposed to liquid (tar) or gases. Lastly, the effect of purge gas flow rate for torrefaction on MY and ML was examined as presented in Table 2. The torrefaction tests were performed at 300 °C and 30 min to examine the effect of purge gas flow rate at 50, 100, and 200 ml min -1 . As observed in Table 2, MY and ML did not differ significantly during the torrefaction of the pellets. On average, the values of MY and ML were 44.57% and 55.44%, respectively. The results indicated that the change in flow rate of the purge gas had a negligible effect on MY and ML. The findings are in good agreement with Asadullah et al. [24] whose study examined the effect of gas flow rates on the torrefaction of palm kernel shell (PKS). Therefore, the gas flow rate for future OPEFB pellet torrefaction experiments should be fixed at 200 ml min -1 .

Conclusion
This paper presents preliminary test results for the torrefaction of OPEFB pellets. The tests examined the effects of temperature (250-350 °C) and torrefaction purge gas flow rate (50, 100, and 200 ml min -1 ) for 30 min. The results showed that torrefaction resulted in significant changes in MY and ML along with improvements in HHV, DE, and EY. However, significant liquid (tar) and gas were generated particularly above 300 °C during torrefaction. Hence, the 300 °C temperature mark is the limiting temperature for torrefaction of OPEFB pellets. The effect of torrefaction purge gas flow rate was also examined and found to be negligible on MY and ML. Therefore, the preliminary tests demonstrated that future experiments on OPEFB pellets torrefaction should be limited to 250 and 300 °C for 30 min under gas flow rate of 200 ml min -1 to selectively yield solid (MY) products.