**2.4 Thermogravimetric analysis (TGA)**

The combustion characteristics of isolated celluloses were studied using thermogravimetric analysis (TGA). The analysis was done on a TGA (Q500, TA instrument). Samples of 6 mg (1.0) were first equilibrated at 25°C for 5 min and then heated at specific heating rates of 10, 15, 20, and 25°C/min to 900°C. The process was performed under constant nitrogen environment flowing at 20 mL/min. As the thermal decomposition progressed, the change in weight was recorded continuously as a function of temperature and time. **Figure 3** shows the isoconversion versus temperature at different heating rates for the isolated celluloses from DPW. The conversion curves for acid-base (colored) and Organosolv (black) methods below 300 and 340°C, respectively, showed similar thermal decomposition patterns at all heating rates. There was a slight shift toward higher temperature side with increasing heating rates, possibly due to the increasing thermal energy in the system [17]. However, at higher temperatures, the conversion pattern changed for both methods, possibly due to the change in the degradation chemistry of components under pyrolysis. It is worth to note that Organosolv cellulose showed better thermal stability than the acid-base cellulose. **Figure 4** shows the differential thermogravimetric (DTG) results against temperature at different heating rates for the DPW and the isolated celluloses. The results showed a typical thermal degradation of lignocellulose biomass. The curves of both samples moved downward as the heating

**Figure 3.** *The relationship of conversion against temperature for acid-base cellulose (colored) and Organosolv cellulose (black).*

#### **Figure 4.**

isolation from the complex lignocellulose matrix of DPW. Stage III had total average mass loss of 8.68 1.2 and 33.08 0.8%, for acid-base and Organosolv methods, respectively. This represented combustion of the carbonaceous and some part of char oxidation [19]. In addition, the higher mass loss for Organosolv cellulose was plausibly due to residual lignin. Moreover, the FTIR results showed some lignin functional groups for this method. The last stage was associated with charring process and ash formation. The average total mass loss for acid-base and Organosolv methods in this stage were 9.57 1.3 and 5.20 0.4%. TGA analysis data was used

**) Tmin (°C) T1 (°C) T2 (°C) T3 (°C) Tmax (°C)**

**)**

**10 15 20 25**

**Acid-base cellulose** 10 30 192 266 402 900

*Investigation of Nonisothermal Combustion Kinetics of Isolated Lignocellulosic Biomass…*

**Organosolv cellulose** 10 30 220 318 422 900

15 197 273 406 20 199 279 411 25 201 381 415

15 226 324 435 20 227 328 441 25 228 330 450

**Stages Temperature Heating rate (°C min<sup>1</sup>**

**Acid-base cellulose** Stage I, WL% Tmin–T1 10.06 9.17 9.33 9.26 Stage II, WL % T1–T3 71.59 72.76 73.16 72.52 Stage III, WL % T3–Tmax 10.18 8.88 7.30 8.35 Final residue at 900-100 °C (%) 7.99 9.19 11.21 9.87 **Organosolv cellulose** Stage I, WL % Tmin–T1 4.35 5.24 5.72 5.79 Stage II, WL % T1–T3 54.42 56.16 57.16 55.55 Stage III, WL % T3–Tmax 35.57 33.30 32.24 33.69 Final residue at 900-100 °C (%) 5.66 5.30 4.88 4.97

*Characteristic temperatures associated with mass loss during pyrolysis of cellulose.*

The TGA data were used to calculate the nonisothermal kinetic and thermody-

namic parameters using model-free equations of Flynn-Wall-Ozawa (FWO), Kissinger-Akahila-Sunose (KAS), and Starink reported in the literature [20].

for kinetic modeling using the model-free methods.

*Mass loss (%) during different stages of cellulose pyrolysis.*

**2.5 Nonisothermal kinetic analysis**

**Heating rate (°C min<sup>1</sup>**

*DOI: http://dx.doi.org/10.5772/intechopen.93549*

**Table 2.**

**Table 3.**

**585**

rate increased, owing to the shorter reaction time at increasing temperature, a phenomenon known as thermal hysteresis. However, **Figure 4(a)** showed a noticeable difference between the peak mass loss patterns compared to **Figure 4(b)** for the two cellulose methods, which suggests a difference in the degradation chemistry. **Table 2** shows the temperature ranges that define the major stages of mass loss in response to increasing temperature for isolated celluloses from both methods, as given in **Table 3**. Stage I started from minimum temperature, Tmin to T1, the total average celluloses mass loss for acid-base and Organosolv were 9.46 0.1 and 5.28 0.1%, respectively. This was attributed to the inherent moisture and water molecules embedded in the intercellular and intracellular void spaces of the celluloses. Stage II, from T1 to T3 for both methods, there was only one major clear peak (**Figure 4(b)**) and the average mass loss in this region was 72.51 0.7 and 55.82 1.1%, for acid-base and Organosolv celluloses, respectively. The weight loss in this stage is associated with pyrolysis of mainly cellulose and to a lesser extent hemicellulose [18]. Compared to the three peaks observed in **Figure 4(a)** for the original DPW, this clearly shows that both methods were effective for cellulose


*Investigation of Nonisothermal Combustion Kinetics of Isolated Lignocellulosic Biomass… DOI: http://dx.doi.org/10.5772/intechopen.93549*

#### **Table 2.**

*Characteristic temperatures associated with mass loss during pyrolysis of cellulose.*


#### **Table 3.**

rate increased, owing to the shorter reaction time at increasing temperature, a phenomenon known as thermal hysteresis. However, **Figure 4(a)** showed a noticeable difference between the peak mass loss patterns compared to **Figure 4(b)** for the two cellulose methods, which suggests a difference in the degradation chemistry. **Table 2** shows the temperature ranges that define the major stages of mass loss in response to increasing temperature for isolated celluloses from both methods, as given in **Table 3**. Stage I started from minimum temperature, Tmin to T1, the total average celluloses mass loss for acid-base and Organosolv were 9.46 0.1 and 5.28 0.1%, respectively. This was attributed to the inherent moisture and water molecules embedded in the intercellular and intracellular void spaces of the celluloses. Stage II, from T1 to T3 for both methods, there was only one major clear peak

*The relationship of DTG against temperature for (a) date palm waste and (b) isolated celluloses (acid-base,*

*The relationship of conversion against temperature for acid-base cellulose (colored) and Organosolv cellulose*

**Figure 3.**

*Biotechnological Applications of Biomass*

*(black).*

**Figure 4.**

**584**

*colored; Organosolv, black).*

(**Figure 4(b)**) and the average mass loss in this region was 72.51 0.7 and

55.82 1.1%, for acid-base and Organosolv celluloses, respectively. The weight loss in this stage is associated with pyrolysis of mainly cellulose and to a lesser extent hemicellulose [18]. Compared to the three peaks observed in **Figure 4(a)** for the original DPW, this clearly shows that both methods were effective for cellulose

*Mass loss (%) during different stages of cellulose pyrolysis.*

isolation from the complex lignocellulose matrix of DPW. Stage III had total average mass loss of 8.68 1.2 and 33.08 0.8%, for acid-base and Organosolv methods, respectively. This represented combustion of the carbonaceous and some part of char oxidation [19]. In addition, the higher mass loss for Organosolv cellulose was plausibly due to residual lignin. Moreover, the FTIR results showed some lignin functional groups for this method. The last stage was associated with charring process and ash formation. The average total mass loss for acid-base and Organosolv methods in this stage were 9.57 1.3 and 5.20 0.4%. TGA analysis data was used for kinetic modeling using the model-free methods.

#### **2.5 Nonisothermal kinetic analysis**

The TGA data were used to calculate the nonisothermal kinetic and thermodynamic parameters using model-free equations of Flynn-Wall-Ozawa (FWO), Kissinger-Akahila-Sunose (KAS), and Starink reported in the literature [20].

$$\text{FWO model : } \ln(\beta) = \ln\left(\frac{AE\_a}{\text{g}(a)R}\right) - 5.331 - 1.052\left(\frac{-E}{RT}\right) \tag{1}$$

$$\text{KAS model}: \ln\left(\frac{\beta\_i}{T\_{a,i^2}}\right) = \ln\left(\frac{AR}{E\_a}\right) - \left(\frac{E\_a}{RT\_a}\right) + \ln\left(\frac{df(a)}{da}\right) \tag{2}$$

*Starink model* : ln *<sup>β</sup><sup>i</sup> T<sup>α</sup>*,*<sup>i</sup>* 1*:*92 ! <sup>¼</sup> ln *AR*0*:*<sup>92</sup> *<sup>g</sup>*ð Þ *<sup>α</sup> <sup>E</sup>*0*:*<sup>92</sup> *α* ! � <sup>1</sup>*:*<sup>0008</sup> *<sup>E</sup><sup>α</sup> RT<sup>α</sup>* � � � 0*:*312 (3)

where *T<sup>α</sup>*,*<sup>i</sup>* is the time to reach a given extent of conversion at temperature *Ti*. At *α*, the value of *E<sup>α</sup>* is determined from the slope of the plot ln ð Þ *β* , ln *β=T<sup>α</sup>*,*<sup>i</sup>* 2 � ), and ln (*β*/*T<sup>α</sup>*,*<sup>i</sup>* <sup>1</sup>*:*<sup>92</sup> Þ versus 1, 000*=T<sup>α</sup>*,*i*.

$$a = \frac{\mathbf{m}\_1 - \mathbf{m}\_t}{\mathbf{m}\_1 - \mathbf{m}\_\ast} \tag{4}$$

difference in the cellulose structure between the two methods. In addition, the residual lignin fractions detected in Organosolv method could also have resulted in the increased energy of activation, *Eα*, and bond dissociation, *ΔH*, needed to overcome the carbon number distribution from other components other than cellulose. It was noted that the *ΔH* values for all samples were positive, an indication of energy consumed during pyrolysis process, and are used to release various volatile and biochar products. Furthermore, the calculated solid-state process parameters were different due to the fundamental differences in the model-free methods [12, 13]. The first difference arises from the slope, S, of straight lines which is directly proportional to the activation energy, that is, FWO, KAS, and Starink slope,

*Investigation of Nonisothermal Combustion Kinetics of Isolated Lignocellulosic Biomass…*

*DOI: http://dx.doi.org/10.5772/intechopen.93549*

*Activation energy and enthalpies of (a) acid-base cellulose and (b) Organosolv by three model-free methods.*

**Tables 4**–**6** show other thermodynamic parameters from the three model-free methods for the acid-base and the Organosolv cellulose samples, respectively. The *ΔG* values for Organosolv cellulose for all model-free methods were lower than those of acid-base cellulose samples. Gibb's free energy gives the measure of how favorable a reaction is to reach chemical equilibrium [24]. In context of the first and second laws of thermodynamics, the sample with higher values of *ΔG* (acid-base cellulose), the further its reaction is from equilibrium and the further its reaction must shift to reach equilibrium. However, the entropy, *ΔS* values were lower for the acid-base celluloses for all model-free methods, with negative entropy values for the FWO model. This implies that the degree of disorder of initial reactants was higher than that of the products formed by bond dissociations [22]. In addition, it was already discussed previously that the heat input during the thermal decomposition was for bond dissociation of the reactants. In the context of reaction energy, the acid-base cellulose sample required lower activation energy and enthalpy to form products than Organosolv cellulose samples. On the other hand, the preexponential factor of Organosolv cellulose was ca. two times higher than that of acid-base cellulose. This was plausibly because the activation energy had a similar trend as already discussed above. The preexponential factor and activation energy both influence chemical kinetics and reaction dynamics in pyrolysis of biomass involving complex heterogeneous reactions [25]. The R<sup>2</sup> of all model-free parameters was

*RT*<sup>α</sup> , respectively. The second difference is in

*<sup>n</sup>* ) at a given temperature, which is

*<sup>n</sup>* ) term, n = 0, 2, and 1.92 for FWO, KAS,

*<sup>S</sup>* ¼ � <sup>1</sup>*:*052*<sup>E</sup>*

**587**

**Figure 5.**

*RT*<sup>α</sup> , *<sup>S</sup>* ¼ � *<sup>E</sup>*

and Starink, respectively).

*RT*<sup>α</sup> *and S* ¼ � <sup>1</sup>*:*0008*<sup>E</sup>*

the time to reach the extent of conversion (*T*α,*<sup>i</sup>*

above 0.98, signifying accuracy of the models.

different across different models (for (*T*α,*<sup>i</sup>*

where m1 is the initial biomass weight, mtis the change in weight at a particular time during the experiment, and m<sup>∞</sup> is the residual weight after time of the experiment.

The choice for these model-free methods is because no previous knowledge about reaction mechanism is required to determine the reaction activation energy [21]. The preexponential factor (A) and thermodynamic parameters [enthalpy (*ΔH*), entropy (*ΔS*), and Gibb's free energy (*ΔG*)] were calculated using equations in literature [22].

$$A = \frac{\left[\beta.\exp\left(\frac{E\_v}{RT\_m}\right)\right]}{\left[RT\_m^2\right]}\tag{5}$$

$$
\Delta H = E\_a - RT \tag{6}
$$

$$
\Delta G = E\_a + RT\_m \ln \left(\frac{K\_B T\_m}{hA}\right) \tag{7}
$$

$$
\Delta \mathcal{S} = \frac{\Delta H - \Delta G}{T\_m} \tag{8}
$$

where *β* is the heating rate, *E<sup>α</sup>* is the activation energy, *Tm* is the maximum peak temperature, *KB* is the boltzman constant, and *h* is the plank constant.

The activation energies for both sample methods were calculated using the three models, namely FWO, KAS, and Starink. These model-free methods avoid the shortcomings during model fitting and kinetic compensation effects. The FWO model-free method compensates the experimental measurement errors. However, the KAS and Starink methods depend on the good constant degree of conversion from the derivative mass loss function to provide precision of the kinetic data [12]. Therefore, application of different model-free methods involves a wide conversion range that allows for the study of change in mechanism during a reaction and reduces mass transfer limitations by using multiple heating rates [13]. **Figure 5** shows the relationship of activation energy and enthalpy from the three model-free methods for acid-base and Organosolv celluloses. Results showed little or no difference between *E<sup>α</sup>* and *ΔH*. This closeness in *E<sup>α</sup>* and *ΔH* values signifies the formation of activation complex and little extra energy might be required to achieve product formation [23]. Organosolv cellulose *E<sup>α</sup>* and *ΔH* values were higher than acid-base cellulose especially at higher temperatures (α >0.6). This was possibly due to a

*Investigation of Nonisothermal Combustion Kinetics of Isolated Lignocellulosic Biomass… DOI: http://dx.doi.org/10.5772/intechopen.93549*

**Figure 5.** *Activation energy and enthalpies of (a) acid-base cellulose and (b) Organosolv by three model-free methods.*

difference in the cellulose structure between the two methods. In addition, the residual lignin fractions detected in Organosolv method could also have resulted in the increased energy of activation, *Eα*, and bond dissociation, *ΔH*, needed to overcome the carbon number distribution from other components other than cellulose. It was noted that the *ΔH* values for all samples were positive, an indication of energy consumed during pyrolysis process, and are used to release various volatile and biochar products. Furthermore, the calculated solid-state process parameters were different due to the fundamental differences in the model-free methods [12, 13]. The first difference arises from the slope, S, of straight lines which is directly proportional to the activation energy, that is, FWO, KAS, and Starink slope, *<sup>S</sup>* ¼ � <sup>1</sup>*:*052*<sup>E</sup> RT*<sup>α</sup> , *<sup>S</sup>* ¼ � *<sup>E</sup> RT*<sup>α</sup> *and S* ¼ � <sup>1</sup>*:*0008*<sup>E</sup> RT*<sup>α</sup> , respectively. The second difference is in the time to reach the extent of conversion (*T*α,*<sup>i</sup> <sup>n</sup>* ) at a given temperature, which is different across different models (for (*T*α,*<sup>i</sup> <sup>n</sup>* ) term, n = 0, 2, and 1.92 for FWO, KAS, and Starink, respectively).

**Tables 4**–**6** show other thermodynamic parameters from the three model-free methods for the acid-base and the Organosolv cellulose samples, respectively. The *ΔG* values for Organosolv cellulose for all model-free methods were lower than those of acid-base cellulose samples. Gibb's free energy gives the measure of how favorable a reaction is to reach chemical equilibrium [24]. In context of the first and second laws of thermodynamics, the sample with higher values of *ΔG* (acid-base cellulose), the further its reaction is from equilibrium and the further its reaction must shift to reach equilibrium. However, the entropy, *ΔS* values were lower for the acid-base celluloses for all model-free methods, with negative entropy values for the FWO model. This implies that the degree of disorder of initial reactants was higher than that of the products formed by bond dissociations [22]. In addition, it was already discussed previously that the heat input during the thermal decomposition was for bond dissociation of the reactants. In the context of reaction energy, the acid-base cellulose sample required lower activation energy and enthalpy to form products than Organosolv cellulose samples. On the other hand, the preexponential factor of Organosolv cellulose was ca. two times higher than that of acid-base cellulose. This was plausibly because the activation energy had a similar trend as already discussed above. The preexponential factor and activation energy both influence chemical kinetics and reaction dynamics in pyrolysis of biomass involving complex heterogeneous reactions [25]. The R<sup>2</sup> of all model-free parameters was above 0.98, signifying accuracy of the models.

*FWO model* : ln ð Þ¼ *<sup>β</sup>* ln *AE<sup>α</sup>*

*T<sup>α</sup>*,*<sup>i</sup>* 2

!

*KAS model* : ln *<sup>β</sup><sup>i</sup>*

*T<sup>α</sup>*,*<sup>i</sup>* 1*:*92

!

*Starink model* : ln *<sup>β</sup><sup>i</sup>*

*Biotechnological Applications of Biomass*

<sup>1</sup>*:*<sup>92</sup> Þ versus 1, 000*=T<sup>α</sup>*,*i*.

(*β*/*T<sup>α</sup>*,*<sup>i</sup>*

experiment.

in literature [22].

**586**

*g*ð Þ *α R* � �

*<sup>g</sup>*ð Þ *<sup>α</sup> <sup>E</sup>*0*:*<sup>92</sup> *α*

where *T<sup>α</sup>*,*<sup>i</sup>* is the time to reach a given extent of conversion at temperature *Ti*. At

*<sup>α</sup>* <sup>¼</sup> m1 � mt m1 � m<sup>∞</sup>

The choice for these model-free methods is because no previous knowledge about reaction mechanism is required to determine the reaction activation energy [21]. The preexponential factor (A) and thermodynamic parameters [enthalpy (*ΔH*), entropy (*ΔS*), and Gibb's free energy (*ΔG*)] were calculated using equations

where m1 is the initial biomass weight, mtis the change in weight at a particular

*β:* exp *<sup>E</sup><sup>α</sup> RTm* h i � �

> *RT*<sup>2</sup> *m*

*<sup>Δ</sup><sup>G</sup>* <sup>¼</sup> *<sup>E</sup><sup>α</sup>* <sup>þ</sup> *RTm* ln *KBTm*

*<sup>Δ</sup><sup>S</sup>* <sup>¼</sup> *<sup>Δ</sup><sup>H</sup>* � *<sup>Δ</sup><sup>G</sup> Tm*

where *β* is the heating rate, *E<sup>α</sup>* is the activation energy, *Tm* is the maximum peak

The activation energies for both sample methods were calculated using the three

models, namely FWO, KAS, and Starink. These model-free methods avoid the shortcomings during model fitting and kinetic compensation effects. The FWO model-free method compensates the experimental measurement errors. However, the KAS and Starink methods depend on the good constant degree of conversion from the derivative mass loss function to provide precision of the kinetic data [12]. Therefore, application of different model-free methods involves a wide conversion range that allows for the study of change in mechanism during a reaction and reduces mass transfer limitations by using multiple heating rates [13]. **Figure 5** shows the relationship of activation energy and enthalpy from the three model-free methods for acid-base and Organosolv celluloses. Results showed little or no difference between *E<sup>α</sup>* and *ΔH*. This closeness in *E<sup>α</sup>* and *ΔH* values signifies the formation of activation complex and little extra energy might be required to achieve product formation [23]. Organosolv cellulose *E<sup>α</sup>* and *ΔH* values were higher than acid-base cellulose especially at higher temperatures (α >0.6). This was possibly due to a

!

<sup>¼</sup> ln *AR Eα* � �

<sup>¼</sup> ln *AR*0*:*<sup>92</sup>

*α*, the value of *E<sup>α</sup>* is determined from the slope of the plot ln ð Þ *β* , ln *β=T<sup>α</sup>*,*<sup>i</sup>*

time during the experiment, and m<sup>∞</sup> is the residual weight after time of the

*A* ¼

temperature, *KB* is the boltzman constant, and *h* is the plank constant.

� <sup>5</sup>*:*<sup>331</sup> � <sup>1</sup>*:*<sup>052</sup> �*<sup>E</sup>*

� <sup>1</sup>*:*<sup>0008</sup> *<sup>E</sup><sup>α</sup>*

� *<sup>E</sup><sup>α</sup> RT<sup>α</sup>* � �

*RT* � �

<sup>þ</sup> ln *df*ð Þ *<sup>α</sup> dα* � �

*RT<sup>α</sup>* � �

� � (5)

*ΔH* ¼ *E<sup>α</sup>* � *RT* (6)

*hA* � � (1)

(2)

(4)

(7)

(8)

� 0*:*312 (3)

2 � ), and ln


*Acid-base. b Organosolv.*
