2.1. Basic Properties of WPS
As received, the WPS is slurry-like containing a very large amount of water. Each sample has been prepared by open-airsun-drying over 3–4 days and then by oven treatment at 378 K over 24 h. Therefore, its water content was not measured. Instead, dry-basis contents were determined. As shown in
Table 1, the values of dry-basis combustibles content (M
C), fixed carbon (M
FC), volatile matters (M
VM), and ash (M
A) (i.e., M
CD, M
FCD, M
VMD, and M
AD) are 95.19, 20.79, 74.41, and 4.81 wt % for WPS1, and 80.30, 18.75, 61.55, and 19.70 wt % for WPS2, respectively. The dried sample when exposed to air gained equilibrium moisture content (M
WE). For WPS1 and WPS2, the values of M
WE are 12.75 and 3.77 wt %, respectively. It is noted that the initial moisture content of the WPS batches would vary depending on the source and pretreatment procedures. The results of this study showed two cases of WPS samples with high and low amounts of initial equilibrium moisture.
The dry-basis high heating value (HHVD or HHD), wet-basis HHV of a sample with equilibrium moisture (HHVWE), and wet-basis low heating value of a sample with equilibrium moisture (LHVWE)are 18.30, 15.97, and 14.60 wt % for WPS1, and 15.72, 15.13, and 14.19 wt % for WPS2, respectively. These results indicate that WPS1 possesses a higher amount of combustibles and a higher heating value while possessing lower ash content than WPS2. Thus, the quality of WPS1 is better than WPS2 and would yield a better torrefied biomass. The dry-basis metal compositions of Na, Ca, and Mg of WPS1 are 0.059, 0.761, and 0.089 wt %, respectively, which are much lower than those of WPS2, which are 0.168, 5.27, and 0.842 wt %, respectively. The high contents of Na, Ca, and Mg of WPS2 are attributed to its high ash content which was discharged from the process using alkaline oxides and would hinder the pyrolysis and torrefaction.
2.2. TGA Characteristics of Pyrolysis and Torrefaction
Figure 1 depicts the diagrams of thermal gravimetric analysis (TGA) for WPS1 and WPS2. It presents the variation of the residual mass fraction of the WPS (M
R, wt %) with temperature (T) during the pyrolysis process. The T started from 378 K (105 °C) because the M
R was calculated by the initial WPS on a dry-basis. Before the pyrolysis process, the WPS was maintained at 378 K in order to remove the moisture. According to the study by Bergman et al. [
20], the pyrolysis process can be divided into three regions. Under 473 K, it involves de-polymerization, re-condensation, and softening. Before 523 K, the process consists of limited de-volatilization and carbonization. After 523 K, the biomass begins to proceed with extensive de-volatilization and carbonization.
The results of
Figure 1 show that the WPS exhibits no extensive decomposition at a temperature below 523 K. It starts to decompose significantly (e.g., mass loss of 5 wt %) at around 553 K and 573 K, respectively, for WPS2 and WPS1, which implies that the torrefaction process has begun. The WPS1 decomposed easier than WPS2 at 773 K by mass amounts of about 87 and 50 wt %, respectively. This is consistent with the assertion that WPS1 has a higher content of volatile matter. A significant residual mass is left in the case of WPS2, due to its high ash content.
The DTG (differential thermogravimetric) curves reveal a rough composition ratio of the constituents in WSP1 and WPS2, which allows for the estimation of the mild temperature for torrefaction [
20]. The value of −dM
R/dT is related to the time variation of M
R, i.e., the reaction rate −dM
R/dt is related by −dM
R/dt = (−dM
R/dT) HR, where HR represents the heating rate (10 K/min). A higher peak indicates a higher reaction rate. As reported by previous studies [
36], hemi-cellulose has a sharp pyrolysis rate around 550–585 K. As soon as 628 K is reached, the intense decomposition of cellulose can be observed. However, the decomposition of lignin is difficult to identify in the DTG curve because the temperature at which this occurs overlaps with that of cellulose [
36]. As shown in
Figure 2, the DTG curves indicated that both WPS1 and WPS2 possess low hemi-cellulose composition, and they are both predominantly composed of cellulose and lignin. This is because the WPS is a mixed waste of waste wood and pulp sludge with no significant hemi-cellulose content.
Figure 2 reveals that WPS1 and WPS2 exhibit four and three significant peaks, respectively, and that their first peaks appear at about 593 K at about 10 wt % of the WPS having been decomposed. Notice that the highest peak of WPS1 at around 683 K is twice as high as that of WPS2 at about 633–653 K. This again indicates that the properties of WPS1 make it easier and faster to decompose than WPS2. Furthermore, the significant peak at 820 K for WPS1 is similar to the DTG pattern of secondary pulp mill sludge obtained from the aerobic biological treatment [
37].
Simulated torrefactions by TGA were performed at T
T in the range of 533 to 593 K with 10 K intervals.
Figure 3 shows the influences of t
T on M
R at various T
T values for WPS1 and WPS2. For both WPS1 and WPS2, a higher T
T as well as a longer t
T induces higher mass reduction. Most of the biomass torrefaction tests have been set to retain a torrefied biomass of about 70 wt % while allowing for the decomposition of the hemi-cellulose and cellulose constituents. For WPS1, the M
R was about 72 wt % after 40 min of torrefaction at 573 K. However, the M
R decreased to about 68 and 61 wt % at 583 and 593 K, respectively. This suggests that the T
T for WPS1 may be properly operated at 573–583 K or lower. For WPS2, the mass reduced slightly after 60 min of torrefaction for the cases with T
T lower than 553 K (e.g., 533 and 543 K). At 40 min, the values of M
R were 75, 71, and 68 wt % at 573, 583, and 593 K, respectively, suggesting the appropriate T
T value of 583–593 K for WPS2.
2.3. Energy Densification
Key parameters reflecting the performances of the torrefaction system assessed are the mass yield (Y
M), energy yield (Y
E), and energy densification ratio (E
D). Y
M is the ratio of the mass of the product to the feed, as calculated by Equation (1). Y
E is the ratio of the energy output of the product to the input of the feed, as computed by Equation (2), or the multiplication of Y
M and the ratio of H
HD of the product to the feed (H
HDTD/H
HDRD). E
D can be calculated from Y
E/Y
M or H
HDTD/H
HDRD by Equation (3). A higher energy densification ratio implies a higher enhancement of the heating value of the torrefied product:
where
HHD = dry-basis high heating value,
HHDRD, HHDTD = HHD of dried raw and torrefied samples,
mRD, mTD = mass of dried raw and torrefied samples.
Referring to the TGA results of
Section 2.2, the lab-scale torrefactions using a tubular furnace were conducted at T
T = 533, 553, and 573 K for WPS1 and 553, 573, and 593 K for WPS2, both with t
T = 20, 40, and 60 min. The heating values for the torrefied WPS are shown in
Table 2. For WPST1, the H
HD generally increases with an increasing t
T at a moderate T
T of 533 and 553 K. However, at high a T
T of 573 K, the torrefaction is vigorous, resulting in a reduction of H
HD as t
T increases. For WPST2, similar enhancing trends at 553 and 573 K were observed, while a tendency to decrease at 593 K with time resulted. The T
T as H
HD decreased with the t
T at 573 K for WPST1 was lower than that at 593 K for WPST2, because WPS1 contains more combustibles making it easier to decompose. In general, a higher T
T produced a higher H
HD, except at T
T = 573 K with a t
T value of 60 min for WPST1. The H
HD was raised from 18.30 MJ/kg for WPS1 to its highest value of 27.49 MJ/kg for the WPST1, which occurred at 573 K with a t
T value of 20 min (WPST1-300-20). That of WPS2 was raised from 15.72 MJ/kg to 19.74 MJ/kg for the WPST2, which occurred att593 K with a t
T value of 20 min (WPST2-320-20). Again, the values of H
HD for WPS1 were higher than those of WPS2, as expected. The H
HD values of WPST1-280-60, WPST1-300-20, WPST1-300-40, and WPST1-300-60 were 25.19, 27.49 (highest), 24.98, and 24.58 MJ/kg, respectively, and are all within the H
HD range of bituminous coal which is 24–35 MJ/kg [
38].
Table 2 also presents the HHV
WE values of WPS1, WPS2, torrefied WPS1 (WPST1) and WPS2 (WPST2), and the LHV
WE values of WPS1, WPS2, WPST1-300-20, and WPST2-320-20. The wet-basis HHV
WE of the above mentioned cases of WPST1-280-60, WPST1-300-20, WPST1-300-40, and WPST1-300-60 were 23.75, 25.87 (highest), 23.51, and 23.13 MJ/kg, respectively, which also satisfy the Quality D coal specification of the Taiwan Power Co., (TPC) (City, Country), which is 20.92 MJ/kg minimum [
39].
Figure 4 presents the Y
M and Y
E values for WPST1 and WPST2 at various conditions. The values of Y
E are higher than those of Y
M, indicating that more energy was retained than mass for the torrefied product. In general, both Y
M and Y
E decrease with increasing t
T and T
T, revealing that more energy and mass are lost as t
T and T
T increase. The corresponding values of E
D are listed in
Table 3 and all listed values are higher than 1, indicating that the torrefaction is beneficial for the energy densification of the WPS. The E
D is generally enhanced with increasing T
T. At moderate T
T values, e.g., 533 and 553 K for WPST1 and 553 and 573 K for WPST2 (depending on the raw materials), a longer t
T yields a higher E
D. However, at a high T
T value such as 573 K for WPST1 and 593 K for WPST2, the extent of torrefaction is very significant resulting in a reduction of E
D with an increasing energy loss as t
T increases.
For the cases of WPST1-280-60, WPST1-300-20, WPST1-300-40, and WPST1-300-60, the values of HHVWE qualify the Quality D coal specification (20.92 MJ/kg minimum). The obtained values of HHD are also within the range of bituminous coal (24–35 MJ/kg), and these ED values are 1.38, 1.50, 1.36, and 1.34 while those of YM are 0.53, 0.56, 0.51, and 0.47, respectively. Excellent energy densification was observed for these cases, however, with a low mass yield of the torrefied product. Some cases, such as WPST1-260-60, WPST1-280-20, and WPST1-280-40, possess respective HHVWE of 21.83, 21.69, and 21.67 MJ/kg, satisfying the Quality D coal specification, whereas the respective HHD values of 23.20, 23.05, and 22.96 MJ/kg are slightly lower than that of bituminous coal. The respective ED values of 1.27, 1.26, and 1.26 are satisfactory with corresponding YM values of 0.58, 071, and 0.56. Hence, WPST1-300-20 with an HHD, HHVWE, ED, and YM of 24.79 MJ/kg, 25.87 MJ/kg, 1.50, and 0.56, respectively, seems to be the optimal case among all WPST1 samples. The highest ED for WPST1 is 1.50 for the WPST1-300-20 torrefied at 573 K for 20 min; however, this sample possesses a low YM value of 0.56.
As for the torrefaction of WPS2, the ED and YM values respectively range from 1.09–1.11 and 0.67–0.80 for WPST2-280, 1.03–1.14 and 0.59–0.69 for WPST2-300, and 1.15–1.26 and 0.57–0.61 for WPST2-320. The highest ED value of 1.26 was observed in the case of WPST2-320-20. The corresponding YM, HHD, and HHVWE values for this sample are 0.61, 19.74 MJ/kg, and 19.31 MJ/kg, respectively. The HHVWE value is close to the Quality D coal specification of the TPC.
2.4. Comparison of Results with Those of Others
The results of the proximate analysis, ultimate analysis, and equilibrium moisture content of WPS1, WPS2, WPST1-300-20, and WPST2-320-20 are listed in
Table 1. After drying, the dried WPS1 and WPS2 possessed the equilibrium moisture contents, M
WE, of 12.75 and 3.77 wt %, respectively. By applying torrefaction, the M
WE of torrefied WPST1-300-20 and WPST2-320-20 decrease to 5.91 and 2.15 wt %, respectively. The M
WE of both torrefied samples are lower than those of WPS1 and WPS2, suggesting that torrefaction beneficially reduces the equilibrium moisture content.
The respective MVMD values of WPS1 and WPS2, 74.41 and 61.55 wt %, were found to have decreased to those of torrefied WPST1-300-20 and WPST2-320-20, 48.67 and 53.75 wt %, respectively. On the other hand, the MFCD values of WPS1 and WPS2 of 20.79 and 18.75 wt % increase to those of torrefied WPST1-300-20 and WPST2-320-20 at 42.74 and 20.86 wt %, respectively. This signifies the role of torrefaction in removing moisture and volatile matters at mild conditions while retaining most of the fixed carbon in biomass. The decrease in MCD (the sum of MVMD and MFCD) then results in the increase of MAD.
The results of the ultimate analysis show an increase in the contents of carbon and nitrogen, while a decrease of hydrogen, oxygen, and sulfur were observed for torrefied WPST1-300-20 and WPST2-320-20. Since the WPS is hydrophilic, a decrease in the oxygen content makes the WPST comparatively more hydrophobic. From the results,it can be also inferred that the structure of the WPS is altered such that C-H, C-O, and C-S bonds are broken during torrefaction, releasing the H-, O-, and S-containing compounds. Thus, relatively larger amounts of carbon and nitrogen are retained. These findings suggest that the hygroscopic nature of the WPS can be destructed to yield hydrophobic products. Therefore, WPST would be easy to store and would not spoil easily. The characteristics of the relative contents of C, H, and O can be expressed in a Van Krevelen plot as atomic ratios of H/C vs. O/C as shown in
Figure 5. A comparison with other carbon-containing materials is also plotted in
Figure 5. The characteristics of H/C and O/C of coals are near the low left corner [
40], while those of the raw WPS, rice straw, and bio-fiber are in the upper right region. After the torrefaction process, the said characteristics of WPST are close to those of coal. Moreover, both the WPST1-300-20 and WPST2-320-20 samples exhibit better heating values and atomic ratios of O/C and H/C when compared with torrefied rice straw at 553 K and 50 min [
35], as shown in
Table 1. Thus, the torrefaction of WPS is suitable for improving the combustion and co-firing quality of the torrefied products, WPST1-300-20 and WPST2-320-20, for use as solid-fuels.