Thermal decomposition behaviors and kinetics of carrageenan-poly vinyl alcohol bio-composite film

Accepted Manuscript Title: Thermal decomposition behaviors and kinetics of carrageenan-poly vinyl alcohol bio-composite film Authors: Fanrong Meng, Yijun Zhou, JunYu Liu, Jun Wu, Guoqing Wang, Ruisong Li, Yucang Zhang PII: DOI: Reference:

S0144-8617(18)30887-7 https://doi.org/10.1016/j.carbpol.2018.07.095 CARP 13895

To appear in: Received date: Revised date: Accepted date:

7-5-2018 28-7-2018 30-7-2018

Please cite this article as: Meng F, Zhou Y, Liu J, Jun W, Wang G, Li R, Zhang Y, Thermal decomposition behaviors and kinetics of carrageenan-poly vinyl alcohol bio-composite film, Carbohydrate Polymers (2018), https://doi.org/10.1016/j.carbpol.2018.07.095 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Thermal decomposition behaviors and kinetics of carrageenan-poly vinyl alcohol biocomposite film Fanrong Meng a,b, Yijun Zhou a, JunYu Liu a, Jun Wu a,b, Guoqing Wang a,b, Ruisong Li a,b,

a

IP T

Yucang Zhang a,b,

Key Laboratory of Advanced Materials of Tropical Island Resources of Ministry of

SC R

Education,

Hainan University, Hai Kou 570228, Hainan, China

State Key Laboratory of Marine Resource Utilization in South China Sea, College of

U

b

N

Materials & Chemical Engineering, Hainan University, Haikou 570228, China

M

A

Abstract

Pyrolysis characteristics of carrageenan-polyvinyl alcohol (CG-PVA) composite

ED

films were studied on a thermo gravimetric analyzer in N2 atmosphere. A stepwise

PT

procedure based on model-free Flynn-Wall-Ozawa (FWO), Kissinger-AkahiraSunose (KAS) and Friedman-Reich-Levi (FRL) methods were applied to calculate the

CC E

apparent activation energies (𝐸). The range of 𝐸 for CG-PVA/LBP/K film was 16.92~171.53 kJ/mol. Coats-Redfern and master-plots methods were utilized to

A

investigate the most probable mechanisms for CG-PVA/LBP/K film. Further kinetic analysis was performed and revealed that five independent parallel reactions were



Corresponding author. Tel: +86 0898 66279219 E-mail address: [email protected] (Y. Zhang)

1

supposed reasonably by deconvoluting DTG curves. Kinetic parameters of the corresponding five pseudo-components for CG-PVA/LBP/K film were separately calculated via Kissinger’s method using the peak maximum temperature. It’s concluded that lignin derivatives contributed the major part of degradation process.

IP T

The orders of activation energy for pseudo components are 𝐸(PVA) > 𝐸(carrageenan) > 𝐸(biomass derivatives). The comprehensive analysis of pyrolysis kinetics may

SC R

expand application fields of carrageenan based biocomposite films.

Keywords: Biocomposite film; slow pyrolysis; kinetic modeling; iso-conversional

N

U

methods; independent parallel reactions

A

1. Introduction

M

In the past few decades, almost all polymers produced and used in the market are

ED

petroleum based, from polyethylene (PE) to polypropylene (PP) (Zhang, Fevre, Jones & Waymouth, 2018). More than 100 million tons of carbon emission are brought

PT

about during plastic production and at least 275 million metric tons of plastic waste

CC E

are generated per year (Jambeck et al., 2015). Academic and industrial scientists have been investing significant efforts to solve the dilemma caused by shortage of fossil fuels and environmental problems. Exploiting alternative renewable sources to replace

A

non-biodegradable synthetic petrochemical materials is one strategy (Roudsari, Mohanty & Misra, 2017; Wang, Qian, He, Xiong, Song & Wang, 2017) and recycling the polymer wastes is another (Ragaert, Delva & Van Geem, 2017; Zhu, Romain & Williams, 2016). 2

Terrestrial crops and aquatic biomass have been extensively utilized as feedstock to obtain bio-derived polymers (Pan, 2011). Banana pseudo-stem, mainly consist of cellulose, hemicellulose and lignin, which are rich in hydroxyl groups, can be converted into bio-polyols using liquefaction method (D'Souza, Camargo & Yan,

IP T

2017). As a red seaweed species, Eucheuma muricatum is the most abundant source of polysaccharides such as carrageenan. Carrageenan is usually utilized as base

SC R

material for packing films due to its gelling capacity and anionic properties (H.P.S et al., 2017). Degradation recycling is one of the most feasible approaches to dispose

U

waste polymer because thermal degradation of polymers and/or biomass can generate

N

valuable chemical products, such as bio-oil (Ranzi, Debiagi & Frassoldati, 2017),

A

syngas (Czajczyńska, Krzyżyńska, Jouhara & Spencer, 2017) and char (Sharma,

M

Pareek & Zhang, 2015). Reaction mechanisms and multi-scale modeling of

ED

lignocellulosic biomass pyrolysis was reviewed in detail by Anca-Couce, including pyrolysis on molecular, particle and reaction levels (Anca-Couce, 2016). Physical and

PT

chemical transformation pathways for cellulose and seaweed polysaccharides during

CC E

co-pyrolysis process was deeply shed light on employing thermogravimetric analysis (TGA) (Wang et al., 2017). Pyrolysis characteristics and kinetics of flexible polyurethane foam was studied combining TGA with infrared spectrometry (IR) and

A

mass spectrometry (MS) techniques to gain an overall knowledge of the aging and combustion mechanism (Garrido & Font, 2015). Found on the TGA basis, there are two main mathematical approaches for the kinetic prediction: model-fitting and model-free methods (Vyazovkin, Burnham, 3

Criado, Pérez-Maqueda, Popescu & Sbirrazzuoli, 2011). In view of iso-conversional procedure, model-free methods can be used to calculate the characteristic parameters such as apparent activation energy (𝐸) and pre-exponential factor (𝐴) without knowing the corresponding reaction order as well as pyrolytic mechanism beforehand.

IP T

Wu et al. (Wu et al., 2014) studied the pyrolysis process of three species of aquatic biomass using Flynn-Wall-Ozawa (FWO) and master-plots methods, the results

SC R

revealed that the activation energy of each zone had a complex relationship with

biomass components. Kissinger-Akahira-Sunose (KAS) methods were applied to

U

analyze the pyrolytic and kinetic characteristics for different mass ratios of

N

microalgae and low-rank coal, the average values of 𝐸 for both LCNA and LCCH

A

samples changed with microalgal biomass ratio (Wu, Yang & Yang, 2018). The

M

amorphous morphologies of biomass components and actual highly heterogeneous

ED

reaction process lead to the reaction mechanism extremely complicated (Hu et al., 2016). Thereby, more complex models, for example, parallel reaction schemes, are

PT

commendably employed to provide better understanding of the pyrolysis mechanisms

CC E

of multiple component materials (Ali, Bahaitham & Naebulharam, 2017; Bach, Tran, Skreiberg & Trinh, 2015). Bach et al. (Bach & Chen, 2017) found that the sevenreaction model offered the highest fit quality and was applicable for microalgal

A

biomass pyrolysis. The main subject of this paper is to gain a deep knowledge of the pyrolysis kinetics of the carrageenan-polyvinyl alcohol composite film (CG-PVA/LBP/K), which has more desirable mechanical properties than many other bionanocomposite 4

coating films in the open literatures listed in Table A1. Exploring suitable mathematical methods to obtain thermal behaviors and precise pyrolysis kinetic parameters of every individual components of composite polymers is very essential for further optimizing the overall synthesis scheme, upgrading modifications and

IP T

broadening application fields for bio-based composite films.

SC R

2. Experimental 2.1 Materials

U

Eucheuma muricatum and banana pseudo-stem collected from Hainan, China

N

were dried and cut into suitable size before used. Chemical compositions of two raw

A

materials are demonstrated in Table A2. Poly vinyl alcohol (PVA), potassium

M

persulphate (K2S2O8), potassium chloride, sodium hydroxide, polyethylene glycol 400

ED

(PEG400) and glycerol were purchased from Xilong Chemical Co., Ltd (China). H2SO4 (98%) was selected as catalyst. All chemicals were reagent grade without

PT

further purification.

CC E

2.2 Preparation of modified carrageenan- poly (vinyl alcohol) composite films Extraction of carrageenan (CG) and liquefaction of banana pseudo-stem (LBP)

A

were presented by our previous work (Meng, Zhang, Xiong, Wang, Li & Zhang, 2018). The physical properties of LBP were exhibited in Table A3. CG (3.75 g) and PVA (15 g) were dissolved simultaneously in distilled water (or potassium chloride solution at 0.05 wt%) with continuously stirring at 90 °C for 1 h. K2S2O8 (0.3 g) acting as initiator was added into the obtained uniform and transparent sol under 5

stirring at 65 °C for another 15 min. The 3wt% LBP was mixed homogeneously with the CG-PVA mixture. 50 mL of the sample solution was casted into a preheated glass sheet and dried at 40 °C for 6 h. The obtained samples were named as CG-PVA, CGPVA/LBP and CG-PVA/LBP/K, respectively. It was understood that the hydrogen

IP T

bonds generated between LBP and CG-PVA copolymer softened the rigidity of the film; the presence of cation (K+) promoted copolymer chains helix transition.

SC R

Therefore, the mechanical properties of the hybrid film were significantly increased, as shown in Fig. A1.

U

Table 1 shows results of characterization of CG-PVA composite films. The

N

proximate analysis was determined using ASTM standards E870-82. A Flash 2000

M

A

Thermo Fisher CHNS/O elemental analyzer was utilized to perform the elemental

a

by difference

PT

2.3 Kinetic analysis

ED

analysis; the oxygen content was obtained by difference.

CC E

The non-isothermal pyrolysis of CG-PVA composite films were performed on a thermogravimetric analyzer (NETZSCH, Germany). Approximately 6 mg of the dried samples was placed in an alumina crucible; the nitrogen flow rate of 20 mL/min was

A

employed for each experimental run. The TGA analyses were conducted while the samples were being heated from 40 to 700 °C with five different heating rates of 5, 10, 20, 30 and 40 °C/min. The thermal analysis was performed using the Pyris™ software. 6

Thermal decomposition is a common preliminary step in the combustion and gasification process. Although involving various different complex reactions, the entire pyrolysis process can be simplified as the global reaction scheme (Bui, Tran & Chen, 2016). K

IP T

Solid biomass → Biochar (s) + Volatiles (g)

SC R

The kinetic equation for the global reaction scheme is generally based on

Arrhenius Eq. (1), which has been considered as the most reliable kinetic analysis

𝐸 𝑅𝑇

)

(1)

N

𝐾(𝑇) = 𝐴𝑒𝑥𝑝 (−

U

method.

A

where 𝐴 is the pre-exponential factor, 𝐸 is the apparent activation energy, 𝑅 is the

M

universal gas constant, 8.314 J/(mol·K) and 𝑇 is the absolute temperature,

ED

respectively.

PT

Assuming that the conversion rate is proportional to the concentration of reactants, heterogeneous and non-isothermal reactions are universally described by

CC E

the following equations: 𝑑𝛼 𝑑𝑡

=𝛽

𝑑𝛼 𝑑𝑇

= 𝐾(𝑇)𝑓(𝛼)

(2)

A

with 𝛼=

𝑤0 −𝑤𝑡 𝑤𝑡 −𝑤∞

,𝛽=

𝑑𝑇 𝑑𝑡

= 𝑐𝑜𝑛𝑠𝑡.

(3)

where 𝛼 is the degree of conversion, 𝑤0 , 𝑤∞ are the initial and final masses of solid and 𝑤𝑡 is the mass of solid at a given time t. 𝛽 is the certain heating rate. The function 7

𝑓(𝛼) is the mechanism model and considers that the rate constant is related to both conversion and temperature (Vyazovkin & Sbirrazzuoli, 2006). By combining the Eq. (1) and (2), the reaction rate can be modified to the

𝛽

𝑑𝛼 𝑑𝑇

= 𝐴𝑒𝑥𝑝 (−

𝐸 𝑅𝑇

) 𝑓(𝛼)

IP T

following Eq.: (4)

SC R

Integration with respect to temperature with the initial condition of 𝛼 = 0 at 𝑇 = 𝑇0 to yields Eq. (5): 𝛼 𝑑𝛼 𝑓(𝛼)

𝐴

𝑇

𝐸

= ∫𝑇 𝑒𝑥𝑝 (− ) 𝑑𝑇 𝛽 𝑅𝑇

(5)

0

U

𝑔(𝛼) = ∫0

A

N

The most common forms of 𝑔(𝛼) and corresponding 𝑓(α) are listed in Table 2.

M

2.3.1 Iso-conversional method

ED

Taking Eq. (4) as the fundamental formulae to interpret, the kinetic analysis method is classified into differential and integral method.

PT

The FRL method (Friedman, 1964) supposes that the pyrolysis is independent of

CC E

the temperature but dependent on the rate of the mass loss; thus, 𝑓(𝛼) is constant at any given 𝛼. 𝑑𝛼

𝑑𝛼

𝑑𝑡

𝑑𝑇

A

𝑙𝑛 ( ) = 𝑙𝑛 (𝛽

) = 𝑙𝑛[𝐴𝑓(𝛼)] −

𝐸

(6)

𝑅𝑇

𝐸 is easy to be calculated by plotting 𝑙𝑛 (𝛽

for each degree of conversion (𝛼).

8

𝑑𝛼 𝑑𝑇

) versus the inverse temperature

The integral solution of Eq. (5) by KAS (Akahira & Sunose, 1971; Kissinger, 1957) results in a linear Eq. (7): 𝐴𝑅 𝐸 𝑙𝑛𝑇𝛽2 = 𝑙𝑛𝐸𝑔(𝛼) − 𝑅𝑇

(7)

IP T

For a given degree of conversion (𝛼), the plot 𝑙𝑛𝑇𝛽2 versus 𝑇1, obtained from thermo grams recorded at five different heating rates, is a straight line, the slope of

SC R

which is used to evaluate 𝐸.

According to FWO method (Flynn & Wall, 1966; Ozawa, 1965), the 𝐸 can be

U

derived by plotting the natural logarithm of β versus 𝑇1 as represented in Eq. (8). 𝐴𝐸 𝐸 𝑙𝑛𝛽 = 𝑙𝑛𝑅𝑔(𝛼) −5.331−1.052𝑅𝑇

N

(8)

M

A

The Kissinger method (Kissinger, 1957) is widely utilized for solving Eq. (4) under the condition of maximum decomposition rate, as shown in Eq. (9): 𝛽 𝑇𝑚 2

= 𝑙𝑛 (−

𝐴𝑅 𝐸

𝑓′()) − 𝛽

𝑇𝑚 2

(9)

𝑅𝑇𝑚

against

PT

The plot of 𝑙𝑛

𝐸

ED

𝑙𝑛

1

𝑇𝑚

is a straight line, 𝐸 can be estimated from the

CC E

premise that 𝑓′() is independent of the heating rate. Then the intercept (b) of the plot can be used to calculate 𝑙𝑛𝐴, as displayed in Eq. (10):

A

𝑙𝑛𝐴 = 𝑏 − 𝑙𝑛(−𝑅𝑓′()) + 𝑙𝑛𝐸

(10)

This method is only available for analysis of single-step kinetics with reactions

that occur under linear-heating-rate conditions (Ali & Bahadar, 2017). 2.3.2 Model- fitting method

9

Coats-Redfern method (Eq. (11)) is the most commonly used method for speculating the reaction mechanism (Çepelioğullar, Haykırı-Açma & Yaman, 2016; Chen, Wang, Lang, Ren & Fan, 2017). 𝑔(𝛼) 𝑇2

= 𝑙𝑛

𝐴𝑅 𝛽𝐸

𝐸

−

(11)

𝑅𝑇

IP T

𝑙𝑛

Master-Plots method is another ubiquitously utilized approach to determine the

SC R

kinetic model for every pyrolysis stage (Aslan, Parthasarathy, Goldfarb & Ceylan, 2017). Combining the Eq. (4) with Eq. (11), Eq. (12) can be shown: 𝑇 𝑇0.5

)

2

𝑑𝛼/𝑑𝑡 (𝑑𝛼/𝑑𝑡)0.5

=

𝑓(𝛼)𝑔(𝛼)

(12)

𝑓(0.5)𝑔(0.5)

U

𝑌(𝛼) = (

𝑑𝛼/𝑑𝑡 (𝑑𝛼/𝑑𝑡)0.5

associated with the reduced rate

M

𝑇0.5

2

)

A

𝑇

The left side of the Eq. (12) (

N

where 0.5 is corresponding to the conversion in 𝛼 = 0.5

which can be obtained from experimental data. The right side of the equation denotes

ED

a reduced theoretical curve, which is determined by each reaction mechanism listed in

PT

Table 2. The kinetic model which is in good conformity to those obtained from experiments can depict an experimental reactive process.

CC E

Iso-conversional method is based on an ideally hypothesis that the thermal

decomposition of polymer or biomass is only a single reaction to generate char and

A

volatiles. Likewise, Coats-Redfern method is suitable for a certain pyrolysis stage in an entire thermal process. The overall conversion rate of the material with partial reactions can be summed up as shown in Eq. (13): 𝑑𝛼 𝑑𝑡

= ∑𝑁 𝑖=1 𝑖

𝑑𝛼𝑖 𝑑𝑡

−𝐸𝑖

= ∑𝑁 𝑖=1 𝑖 𝐴𝑖 𝑒𝑥𝑝 (

𝑅𝑇

) 𝑓(𝛼) 10

(13)

where 𝑖 is the mass loss contribution factor of each 𝑖 pseudo-component, 𝑁 is the number of the components. Curve fitting, which is based on non-linear least squares method, is usually employed for different models to fit the experimental DTG data. The peaks

IP T

deconvolution and kinetic parameters extraction were conducted using Origin® Pro 2016 and Microsoft Excel® 2010 as the main software.

𝑑𝛼𝑖 𝑑𝑡

)

𝑒𝑥𝑝

−(

𝑑𝑡

𝑑𝛼𝑖

𝑒𝑥𝑝

and (

𝑑𝑡

)

)

𝑐𝑎𝑙

𝑐𝑎𝑙

]

2

(14)

represent the experimental and calculated conversion

U

where (

𝑑𝑡

𝑑𝛼𝑖

)

SC R

𝑑𝛼𝑖

𝑆 = ∑𝑛𝑖=1 [(

𝑑𝛼𝑖 ) ] 𝑑𝑡 𝑒𝑥𝑝

) 100% 𝑚𝑎𝑥

M

[(

𝑆 𝑛

(15)

ED

𝐹𝑖𝑡 (%) = (1 −

√

A

measured and modeled values is defined as:

N

rates, respectively; 𝑛 is the number of experimental points. The fit quality between

3. Results and discussion

PT

3.1 Pyrolysis behavior

CC E

Fig. 1(a), (b), (c) reveals that increasing heating rate extends TG and DTG curves of CG-PVA composite films to a higher temperature region, without affecting the

A

total mass loss and the pattern of the thermal decomposition. It is probably owing to better heat transfer effect between the furnace and sample at lower heating rate. As polymers are weak heat conductor, large temperature gradient would be induced throughout the cross-section of the samples with the increase of heating rate. Herein, the thermal hysteresis created causes the temperature shifts. The similar phenomenon 11

has also been observed in pyrolysis of aquatic biomass (Wu et al., 2014), oil-plant residues (Chen, Wang, Lang, Ren & Fan, 2017) and low-lipid microalgae (Gai, Zhang, Chen, Zhang & Dong, 2013). Taking the thermal data obtained at 10 °C/min as an example (Figure 1(d)), pure

IP T

CG-PVA pyrolysis undergoes three thermal decomposition stages. The first mass loss

SC R

stage takes place in low temperature region, which is associated with the moisture

volatilization and oligomer pyrolysis of the CG-PVA. Simultaneously, a fairly broad endothermic peak in the heat flow curve (Fig. 1(e)) is observed around 91 °C. Glass

N

U

transition phenomenon and decomposition of hard segment occur in the second stage

A

at 200~300 °C. 29% of the mass loss during this degradation could be due to the bond

M

rupture of PVA branch grafted onto CG backbones (Sukhlaaied & Riyajan, 2013) and elimination of side groups from the main chain of PVA (Alexy, Káchová, Kršiak,

ED

Bakoš & Šimková, 2002). It is noticeable that a sharp endothermic peak with shoulder

PT

induced in the heat flow curve (Fig. 1(e)) due to the energy required for fracture of the carbon chains (Shah, Jan, Mabood & Jabeen, 2010). It is also assumed that the

CC E

temperature of peak maximum indicates the melting point of the film. When the temperature is higher than 300 °C in the last stage, soft segments such as polyether

A

between CG molecules and carbon double bonds in eliminated PVA molecular chains were thermal degraded, 45% mass loss was triggered. Notably, a new distinct pyrolysis peak at temperatures within 250~350 °C appears after introduced LBP into the film system (Fig. 1(b)) and another endothermic peak around 165 °C with

12

hysteresis is observed in the heat flow curve. This is mainly corresponds to the decomposition of ether bonds or ester bonds in the matrix of the film. These soft segments are generated by crosslinking with residual solvents and bio-polyols in LBP. Furthermore, light volatile compounds generated during liquefaction process moved

IP T

the thermolysis under 250 °C to lower temperature side. For the K+ added film, a large endothermic peak in the first stage exhibited in DSC curves. The second

SC R

decomposition stage shifted from 120~249 °C to 150~249 °C and the conversion rate

increased deeply (Fig. 1(d)). It’s believed to be ascribed to that small concentration of

U

potassium ion plays a pivotal role in increasing interfacial interaction and creating

N

helical dimers' aggregation in the carrageenan molecular chains that improved thermal

M

A

stability of polymer (Meng, Zhang, Xiong, Wang, Li & Zhang, 2018). 3.2 Calculation of apparent activation energy

𝑑𝛼 𝑑𝑇

), ln𝑇𝛽2 and lnβ vs. 1000 demonstrate quite good linear fits at high 𝑇

PT

plots of 𝑙𝑛 (𝛽

ED

As depicted in Fig. 2, in conversion range of 0.1~ 0.9 with a 0.1 step-size, the

CC E

conversion. A systematic error introduced by data smoothing or temperature error that depends on heating rate will take place under low starting heating temperature when using differential iso-conversion model (Starink, 2007). The values of apparent

A

activation energy for CG-PVA, CG-PVA/LBP and CG-PVA/LBP/K films can be separately calculated based on the three slopes, as shown in Fig. 3. It can be seen that apparent activation energy on conversion value calculated by FRL, KAS and FWO methods exhibit almost the same tendency in the 0.1~ 0.9 range. However, the values 13

of activation energy determined with the FRL method show little difference with the two other methods, which may due to systematic error for KAS and FWO methods does not exist in the FRL method (Vyazovkin, 2001). These trends are in agreement with results when processing TGA data of PE and PP (Aboulkas, Harfi & Bouadili,

IP T

2010), macro algae (Ceylan, Topcu & Ceylan, 2014) and typical Indian coal (Jain, Mehra & Ranade, 2016). The apparent activation energy of CG-PVA, CG-PVA/LBP

SC R

and CG-PVA/LBP/K films are in the range of 41.04 to 154.79 kJ/mol, 36.08 to

201.43 kJ/mol and 16.92 to 171.53 kJ/mol with the increasing conversion rate from 10

U

to 90 wt%. The rise of apparent activation energy is presumably related to that more

N

than one step chemical reactions in the pyrolysis process of composite films need

3.3 Reaction mechanisms

M

A

more energy to overcome potential barrier and gain higher conversion fraction.

ED

Given the deviations and relatively poor linear fits at low pyrolysis temperature

PT

mentioned above, whereby the main three stages defined according to Fig. 1(c), (d) are laid special stress on analyzing the thermal decomposition process for CG-PVA

CC E

based composite films. As displayed in Table 3, the initial decomposition temperature of LBP added sample is much lower than that of pure CG-PVA in the second stage,

A

while CG-PVA/LBP/K begins to decompose later. And the maximum decomposition temperature (Tmax) of CG-PVA/LBP/K is the highest. Nevertheless, CG-PVA film possesses only one main pyrolysis peak in the third and fourth stage. New pyrolysis stage appearing and temperature shifting support hypothesis that individual

14

component such as bio-polyol or K+ in the composite film significantly influence its thermal degradation characteristics. a

T0 is the initial temperature of each main decomposition stage.

b

Tm is the temperature of maximum decomposition rate.

c

Te is the end temperature of the main decomposition stage, which is the same as T0 of next

IP T

zone.

Coats–Redfern method and twelve kinds of reaction model in solid-state

SC R

reactions (Table 2) are adopted to focus on analyzing the pyrolysis kinetics of CGPVA/LBP/K film. The obtained activation energies for the three studied stages at

U

constant heating rate of 10 °C/min are presented in Table A4. It reveals that activation

N

energy determined employing most mechanism models exhibit relatively high

A

correlation coefficients (R) and low standard deviation (SD) except D1, D2, D3, D4

M

and F3 models. Based on the results of R2 and SD calculated, it’s suggested that

ED

respective mechanism function for the three stages of CG-PVA/LBP/K film degradation are F1, F2 and F1 model. Apparent activation energy corresponding to

PT

these pyrolysis stages are 131.79 kJ/mol, 178.02 kJ/mol and 102.61 kJ/mol,

CC E

respectively.

Generalized master-plots method is further conducted to confirm the most

A

suitable thermal degradation mechanisms. Master and experimental curves of three decomposition stages are illustrated in Fig. 4. It’s obvious that the mechanisms for three stages give better match with a certain order model at lower conversion, the mechanisms reveal as order model between reaction n = 1 and n = 2 at higher conversion (α > 0.5). This phenomenon may be connected to the inner molecular 15

structure and amorphous component of the film that effect heating transfer throughout the sample (Hu et al., 2016). Hence, although the overall kinetic model for the composites film is in accordance with the result generated from Coats-Redfern method, mechanism changes in the process of pyrolysis verify that the kinetic

IP T

mechanism of the CG-PVA/LBP/K film does not possess a single reaction.

SC R

3.4 Kinetic evaluation by pseudo-component model

On the basis of existing research, it can be concluded that various constituents with reactivity in series or in parallel result in more than one decomposition and

N

U

devolatilization stage in the pyrolysis of the CG-PVA/LBP/K film. In light of model

A

fitting, the DTG curves in Fig. 1(c) was deconvoluted into five pseudo-component

M

peaks and then the peaks are simulated by Gaussian distribution, as demonstrated in Fig. 5. Thus, the pyrolysis of CG-PVA/LBP/K film process was divided into five

ED

single reactions. Results show that the fit quality between the experimental and

PT

calculated data is within 97.0 - 98.9%, which indicates the goodness of model variant

CC E

representing the composites film decomposition. GC-MS analysis of the LBP was conducted and presented in Fig. A2. Aliphatic

compounds such as 1-propanol, 2-(2-methoxypropoxy)-, levulinic acid butyl ester and

A

other ester, which were derived from the decomposition of the hemicellulose or cellulose, were detected. Meanwhile, phenolic monomers, vanillin, fused aromatic chemicals and other phenolic compounds, which stemmed from lignin degradation, were also achieved. As a consequence, it’s possible to allocate the five modeled 16

components of CG-PVA/LBP/K film to hemicellulose derivatives (Hemicellulose-D), cellulose derivatives (Cellulose-D), lignin derivatives (Ligin-D), carrageenan and PVA. Referring to the decomposition temperature ranges of components in different polymers listed in Table 4, the five peaks in the temperature ranges of 40~210 °C,

IP T

153~308 °C, 207~352 °C, 40~605 °C and 309~609 °C can be assigned to the thermal decomposition of hemicellulose-D, carrageenan, cellulose-D, ligin-D and PVA,

SC R

respectively. Among the five simulated pseudo-components, Hemicellulose-D, Cellulose-D and Ligin-D exactly exhibit the typical features of lignocellulosic

U

biomass pyrolysis: hemicellulose degradation at lower temperature; cellulose is more

N

stable than hemicellulose and decomposes quickly once up to the sufficient

A

temperature; lignin has the slowest pyrolysis rate and cover a wider temperature range

M

(Hu et al., 2016). As presented in Table 4, different polymers possess different

CC E

PT

ED

activity

A

energy. It disclosed that the distinction in the kinetic parameters is related to the distinction in physical and chemical heterogeneity of the sample. Furthermore, different operation conditions, systematic errors and all sorts of mathematical

17

methods may cause the obtained kinetic parameters to distinct from each other (White, Catallo & Legendre, 2011). The mass loss fraction (i) and the extracted kinetic parameters of the pseudocomponents are presented in Table 5. The average mass fraction of lignin-D, PVA,

IP T

CG, hemicellulose-D and cellulose-D are approximately 34.7%, 32.0%, 19.0%,

SC R

10.3%, and 4.0%, respectively. Ligin-D contributes the major part of degradation

process, which may ascribed to that lignin-D degradation covers a broad temperature range because of its hard destroyed benzene ring. Previous reports concluded that the

U

solvolysis of hemicellulose, lignin and amorphous cellulose occurs earlier than

N

decomposition of crystalline cellulose in the process of liquefaction (Zhang, Ding,

M

A

Luo, Xiong & Chen, 2012). Accordingly, it would be reasonable to suppose that low content of cellulose-D in liquefied product cause the least mass fraction in

ED

degradation.

PT

The degradation peaks of other 4 substances nearly overlap the pyrolysis peaks

CC E

of lignin-D at lower heating rate, so the R2 value of the linear fit for Kissinger's plot is only 0.88. Vaporization and devolatilization occur in the low temperature (< 150 °C) resulted in the low R2 value (0.87). It also perhaps the fact that some lighter fraction

A

existed in hemicellulose-D led to its apparent energy is the lowest (=30.60 kJ/mol). The estimated values of the apparent activation energy of carrageenan, cellulose-D, lignin-D and PVA are respectively 142.27, 115.98, 82.42 and 231.48 kJ/mol. The high activation energy of PVA means that plastic is more stable than polysaccharide 18

or biomass derivatives. The result is in agreement with previous studies on the decomposition of LDPE and cellulose (Gunasee, Danon, Görgens & Mohee, 2017). Overall, it can be observed that the five-reaction model is more capable to precisely predict physical meanings of the involved components.

IP T

4. Conclusions

SC R

Thermal decomposition characteristics of CG-PVA composite films were fully investigated utilizing TGA and kinetic modeling. The apparent activation energy of CG-PVA films calculated by iso-conversional methods (FRL, KAS and FWO) was

N

U

found to increase from 31.75 to 261.60 kJ/mol with conversions proceeding.

A

Introducing bio-polyol and K+ into the film plays a vital role in significantly influence

M

the pyrolysis behaviors. Coats-Redfern and master-plots methods employed in combination confirmed that the pyrolysis process of CG-PVA/LBP/K film did not

ED

possess a single reaction. The assumed multi pseudo-component kinetic model was

PT

further developed by deconvoluting DTG curves to simulate five pseudo-components (derivatives of cellulose, hemicellulose and lignin in liquefied products, carrageenan

CC E

and PVA) for CG-PVA/LBP/K film. Results show that carrageenan ranking second largest in activation energy distribution for pseudo components may improve the

A

thermal properties of bio-based films; lignin derivatives degradation covered a broad temperature range. The multiple reaction models with high fit quality can be applied for other biocomposite polymers pyrolysis kinetics design. The overall thermolysis of

19

the carrageenan based biocomposite film is expected to guide its further elevated temperature resistance or flame resistance modification. Acknowledgements

IP T

This study was supported by the National Natural Science Foundation of China (No. 51263006), the Hainan Province of Key Project (ZDYF2017005), the Ministry

SC R

of PhD Education and the Hainan International Science and Technology Cooperation

Specific (KJHZ2014-02). The authors wish to thank the Analytical and Testing Center

U

of Hainan University.

N

Reference

A

Aboulkas, A., Harfi, K. E., & Bouadili, A. E. (2010). Thermal degradation behaviors of polyethylene and polypropylene. Part I: Pyrolysis kinetics and mechanisms. Energy

M

Conversion and Management, 51(7), 1363-1369.

Akahira, T., & Sunose, T. (1971). Joint convention of four electrical institutes. Science

ED

Technology, 16, 22-31.

Alexy, P., Káchová, D., Kršiak, M., Bakoš, D., & Šimková, B. (2002). Poly(vinyl alcohol) 421.

PT

stabilisation in thermoplastic processing. Polymer Degradation and Stability, 78(3), 413-

Ali, I., & Bahadar, A. (2017). Red Sea seaweed (Sargassum spp.) pyrolysis and its

CC E

devolatilization kinetics. Algal Research, 21, 89-97.

Ali, I., Bahaitham, H., & Naebulharam, R. (2017). A comprehensive kinetics study of coconut shell waste pyrolysis. Bioresour Technol, 235, 1-11.

A

Anca-Couce, A. (2016). Reaction mechanisms and multi-scale modelling of lignocellulosic biomass pyrolysis. Progress in Energy and Combustion Science, 53, 41-79.

Aslan, D. I., Parthasarathy, P., Goldfarb, J. L., & Ceylan, S. (2017). Pyrolysis reaction models of waste tires: Application of Master-Plots method for energy conversion via devolatilization. Waste Management, 68, 405-411. Bach, Q. V., & Chen, W. H. (2017). A comprehensive study on pyrolysis kinetics of microalgal biomass. Energy Conversion and Management, 131, 109-116.

20

Bach, Q. V., Tran, K. Q., Skreiberg, Ø., & Trinh, T. T. (2015). Effects of wet torrefaction on pyrolysis of woody biomass fuels. Energy, 88, 443-456. Bui, H. H., Tran, K. Q., & Chen, W. H. (2016). Pyrolysis of microalgae residues – A kinetic study. Bioresour Technol, 199(Supplement C), 362-366. Çepelioğullar, Ö., Haykırı-Açma, H., & Yaman, S. (2016). Kinetic modelling of RDF pyrolysis: Model-fitting and model-free approaches. Waste Management, 48, 275-284. Ceylan, S., Topcu, Y., & Ceylan, Z. (2014). Thermal behaviour and kinetics of alga

IP T

Polysiphonia elongata biomass during pyrolysis. Bioresour Technol, 171, 193-198. Chen, J., Wang, Y., Lang, X., Ren, X. e., & Fan, S. (2017). Comparative evaluation of

thermal oxidative decomposition for oil-plant residues via thermogravimetric analysis:

SC R

Thermal conversion characteristics, kinetics, and thermodynamics. Bioresour Technol, 243, 37-46.

Czajczyńska, D., Krzyżyńska, R., Jouhara, H., & Spencer, N. (2017). Use of pyrolytic gas

U

from waste tire as a fuel: A review. Energy, 134, 1121-1131.

D'Souza, J., Camargo, R., & Yan, N. (2017). Biomass Liquefaction and Alkoxylation: A

N

Review of Structural Characterization Methods for Bio-based Polyols. Polymer Reviews,

A

57(4), 668-694.

M

Fernandes, D. M., Hechenleitner, A. A. W., & Pineda, E. A. G. (2006). Kinetic study of the thermal decomposition of poly(vinyl alcohol)/kraft lignin derivative blends.

ED

Thermochimica Acta, 441(1), 101-109.

Flynn, J. H., & Wall, L. A. (1966). A quick, direct method for the determination of activation energy from thermogravimetric data. Journal of Polymer Science Part B: Polymer

PT

Letters, 4(5), 323-328.

Friedman, H. L. (1964). Kinetics of thermal degradation of char‐forming plastics from thermogravimetry. Application to a phenolic plastic. Journal of Polymer Science Part C:

CC E

Polymer Symposia, 6(1), 183-195.

Gai, C., Zhang, Y., Chen, W., Zhang, P., & Dong, Y. (2013). Thermogravimetric and kinetic analysis of thermal decomposition characteristics of low-lipid microalgae. Bioresour

A

Technol, 150(Supplement C), 139-148.

Garrido, M. A., & Font, R. (2015). Pyrolysis and combustion study of flexible polyurethane foam. Journal of Analytical and Applied Pyrolysis, 113, 202-215.

Gunasee, S. D., Danon, B., Görgens, J. F., & Mohee, R. (2017). Co-pyrolysis of LDPE and cellulose: Synergies during devolatilization and condensation. Journal of Analytical and Applied Pyrolysis, 126, 307-314.

21

H.P.S, A. K., Saurabh, C., Tye, Y., Lai, T. K., Easa, A., Rosamah, E., Fazita, M., Syakir, M. I., Adnan, A. S., Fizree, H. M., Aprilia, S., & Banerjee, A. (2017). Seaweed based sustainable films and composites for food and pharmaceutical applications: A review. Renewable and Sustainable Energy Reviews, 77, 353-362. Hu, M., Chen, Z., Wang, S., Guo, D., Ma, C., Zhou, Y., Chen, J., Laghari, M., Fazal, S., Xiao, B., Zhang, B., & Ma, S. (2016). Thermogravimetric kinetics of lignocellulosic biomass slow pyrolysis using distributed activation energy model, Fraser–Suzuki deconvolution,

IP T

and iso-conversional method. Energy Conversion and Management, 118, 1-11. Jain, A. A., Mehra, A., & Ranade, V. V. (2016). Processing of TGA data: Analysis of isoconversional and model fitting methods. Fuel, 165, 490-498.

SC R

Jambeck, J. R., Geyer, R., Wilcox, C., Siegler, T. R., Perryman, M., Andrady, A., Narayan,

R., & Law, K. L. (2015). Plastic waste inputs from lan into the ocean. Science, 347, 768771.

Kissinger, H. E. (1957). Reaction Kinetics in Differential Thermal Analysis. Analytical

U

Chemistry, 29(11), 1702-1706.

N

Ma, S., Chen, L., Liu, X., Li, D., Ye, N., & Wang, L. (2012). Thermal Behavior of Carrageenan: Kinetic and Characteristic Studies. International Journal of Green Energy,

A

9(1), 13-21.

M

Meng, F., Zhang, Y., Xiong, Z., Wang, G., Li, F., & Zhang, L. (2018). Mechanical, hydrophobic and thermal properties of an organic-inorganic hybrid carrageenan-

ED

polyvinyl alcohol composite film. Composites Part B: Engineering, 143, 1-8. Ozawa, T. (1965). A New Method of Analyzing Thermogravimetric Data. Bulletin of the

PT

Chemical Society of Japan, 38(11), 1881-1886. Pan, H. (2011). Synthesis of polymers from organic solvent liquefied biomass: A review. Renewable and Sustainable Energy Reviews, 15(7), 3454-3463.

CC E

Ragaert, K., Delva, L., & Van Geem, K. (2017). Mechanical and chemical recycling of solid plastic waste. Waste Management, 69, 24-58.

Ranzi, E., Debiagi, P. E. A., & Frassoldati, A. (2017). Mathematical Modeling of Fast

A

Biomass Pyrolysis and Bio-Oil Formation. Note II: Secondary Gas-Phase Reactions and Bio-Oil Formation. ACS Sustainable Chemistry & Engineering, 5(4), 2882-2896.

Roudsari, G. M., Mohanty, A. K., & Misra, M. (2017). Green Approaches To Engineer Tough Biobased Epoxies: A Review. ACS Sustainable Chemistry & Engineering, 5(11), 95289541.

22

Shah, J., Jan, M. R., Mabood, F., & Jabeen, F. (2010). Catalytic pyrolysis of LDPE leads to valuable resource recovery and reduction of waste problems. Energy Conversion and Management, 51(12), 2791-2801. Sharma, A., Pareek, V., & Zhang, D. (2015). Biomass pyrolysis—A review of modelling, process parameters and catalytic studies. Renewable and Sustainable Energy Reviews, 50, 1081-1096. Starink, M. J. (2007). Activation energy determination for linear heating experiments:

IP T

deviations due to neglecting the low temperature end of the temperature integral. Journal of Materials Science, 42(2), 483-489.

Sukhlaaied, W., & Riyajan, S. A. (2013). Synthesis and properties of carrageenan grafted

SC R

copolymer with poly(vinyl alcohol). Carbohydrate Polymers, 98(1), 677-685.

Vyazovkin, S. (2001). Modification of the integral isoconversional method to account for variation in the activation energy. Journal of Computational Chemistry, 22(2), 178-183.

U

Vyazovkin, S., Burnham, A. K., Criado, J. M., Pérez-Maqueda, L. A., Popescu, C., &

Sbirrazzuoli, N. (2011). ICTAC Kinetics Committee recommendations for performing

N

kinetic computations on thermal analysis data. Thermochimica Acta, 520(1), 1-19.

A

Vyazovkin, S., & Sbirrazzuoli, N. (2006). Isoconversional Kinetic Analysis of Thermally Stimulated Processes in Polymers. Macromolecular Rapid Communications, 27(18),

M

1515-1532.

Wang, J., Qian, W., He, Y., Xiong, Y., Song, P., & Wang, R. (2017). Reutilization of

ED

discarded biomass for preparing functional polymer materials. Waste Management, 65, 11-21.

PT

Wang, S., Xia, Z., Hu, Y., He, Z., Uzoejinwa, B. B., Wang, Q., Cao, B., & Xu, S. (2017). Copyrolysis mechanism of seaweed polysaccharides and cellulose based on macroscopic experiments and molecular simulations. Bioresour Technol, 228, 305-314.

CC E

White, J. E., Catallo, W. J., & Legendre, B. L. (2011). Biomass pyrolysis kinetics: A comparative critical review with relevant agricultural residue case studies. Journal of Analytical and Applied Pyrolysis, 91(1), 1-33.

A

Wu, K., Liu, J., Wu, Y., Chen, Y., Li, Q., Xiao, X., & Yang, M. (2014). Pyrolysis characteristics and kinetics of aquatic biomass using thermogravimetric analyzer. Bioresour Technol, 163, 18-25.

Wu, Z., Yang, W., & Yang, B. (2018). Thermal characteristics and surface morphology of char during co-pyrolysis of low-rank coal blended with microalgal biomass: Effects of Nannochloropsis and Chlorella. Bioresour Technol, 249, 501-509.

23

Zhang, H., Ding, F., Luo, C., Xiong, L., & Chen, X. (2012). Liquefaction and characterization of acid hydrolysis residue of corncob in polyhydric alcohols. Industrial Crops and Products, 39, 47-51. Zhang, X., Fevre, M., Jones, G. O., & Waymouth, R. M. (2018). Catalysis as an Enabling Science for Sustainable Polymers. Chemical Reviews, 118(2), 839-885. Zhu, Y., Romain, C., & Williams, C. K. (2016). Sustainable polymers from renewable

A

CC E

PT

ED

M

A

N

U

SC R

IP T

resources. Nature, 540(7633), 354-362.

24

IP T SC R U N A M ED PT

Fig. 1. TG and DTG curves of CG composites film at different heating rate (a. CG-PVA, b. CG-PVA/LBP, c. CG-PVA/LBP/K, d. β=10 °C/min) and (e) heat flux curves of CG

A

CC E

composites film.

25

IP T SC R

U

Fig. 2. Least square regression lines to apparent activation energy proposed by (a) Friedman;

A

CC E

PT

ED

M

A

N

(b) KAS and (c) FWO methods for CG-PVA/LBP/K using TG data shown in Fig. 1(c).

26

Fig. 3. Activation energy of CG-PVA composites film as a function of conversion (α). (a.

ED

M

A

N

U

SC R

IP T

CG-PVA, b. CG-PVA/LBP, c. CG-PVA/LBP/K, d. mean values)

Fig. 4. Master-plots of different kinetic models and experimental data at 10 °C/min for CG-

A

CC E

PT

PVA/LBP/K film degradation during the 2nd (a), 3rd (b) and 4th (c) stage.

27

IP T

Fig. 5. Curve fitting and Kissinger’s plot for pseudo-components in CG-PVA/LBP/K at

SC R

10°C/min.

U

Table 1

A

Proximate analyses /%

C

H

S

Oa

73.55

14.85

5.37

45.76

7.61

0.71

44.44

CG-PVA/LBP

14.27

65.15

15.90

4.68

44.51

7.69

0.76

42.15

CG-PVA/LBP/K

10.26

68.45

15.27

6.02

43.78

7.57

0.53

41.47

Fixed carbon

CC E

PT

ED

6.23

Table 2

Ash content

Volatile matter

CG-PVA

Element /%

M

Code Moisture

N

Ultimate and proximate analyses of CG-PVA composites films

Integral and differential functions of the most common solid state mechanism model (Gai,

A

Zhang, Chen, Zhang & Dong, 2013) Mechanism model Avarami -Erofe’ve (A2) Avarami -Erofe’ve (A3) Avarami -Erofe’ve (A4) Contracting Sphere (R2) Contracting Cylinder (R3) One-dimensional diffusion (D1) Two-dimensional diffusion (D2) Three-dimensional diffusion (Jander，D3) Ginstling -Brounshtein (D4) First-order (F1) Second-order (F2)

𝑔(𝛼) [-ln(1-α)]1/2 [-ln(1-α)]1/3 [-ln(1-α)]1/4 [1-(1-α)1/2] [1-(1-α)1/3] α2 [(1-α)ln(1-α)]+α [1-(1-α)1/3]2 1-(2α/3)-(1-α)2/3 -ln(1-α) (1-α)-1-1

28

𝑓(𝛼)=1/𝑔′(𝛼) 2(1-α) [-ln(1-α)]1/2 3(1-α) [-ln(1-α)]2/3 4(1-α) [-ln(1-α)]3/4 2(1-α)1/2 3(1-α)2/3 1/2α [-ln(1-α)]-1 3(1-α)2/3/[2(1-(1-α)1/3)] 3/2((1-α)-1/3-1) (1-α) (1-α)2

[(1-α)-2-1]/2

Third-order (F3)

(1-α)3

Table 3 Main temperature range (°C) of pyrolysis stages at heating rate of 10 °C/min. The 2nd stage Tmb Tec 166 218 280 122 200 244 155 221 243

The 3rd stage T0 Tm Te 280 443 526 244 291 329 243 289 325

T0a

SC R

CG-PVA CG-PVA/LBP CG-PVA/LBP/K

The 4th stage T0 Tm Te -- -- -329 435 545 325 433 535

IP T

Codes

Table 4

U

Degradation temperature ranges (T0~Te) (°C) and activation energy (𝐸) (kJ/mol) of pseudo-

PVA/kraft lignin derivative film CG-PVA /LBP/K film

--

160, 157

T0~Te

150~335

--

𝐸

104, 138

T0~Te

--

𝐸

--

T0~Te 𝐸

40~210 31

Cellulose --

Lignin --

A

𝐸

CC E

Table 5

CG 216~332

PVA --

--

--

248~360

160~456

--

M

Spruce and birch

Hemicellulose --

--

--

222, 194

66, 84

--

--

--

30~900

180~520

--

--

18~63

160~270

153~308 142

207~352 116

40~605 82

309~609 231

ED

Carrageenan

T0~Te

Methods FWO, KAS Model fitting KAS Model fitting

Refs. (Ma, Chen, Liu, Li, Ye & Wang, 2012) (Bach, Tran, Skreiberg & Trinh, 2015) (Fernandes, Hechenleitner & Pineda, 2006) Current study

PT

Polymers

N

components in different biopolymers.

The fitting results of the pseudo-components in CG-PVA/LBP/K film. i

Kissinger method

5 °C/min

10 °C/min

20 °C/min

30 °C/min

40 °C/min

Average±SD

𝐸 (kJ/mol)

𝑙𝑛𝐴 (1/min)

R2

Hemicellulose-D

0.169

0.066

0.103

0.107

0.070

0.103±0.041

30.596

2.5178

0.869

Carrageenan

0.166

0.216

0.186

0.198

0.186

0.190±0.018

142.269

28.025

0.996

Cellulose-D

0.026

0.055

0.037

0.033

0.048

0.040±0.011

115.980

18.371

0.990

Lignin-D

0.287

0.378

0.324

0.357

0.390

0.347±0.041

82.417

10.300

0.880

PVA

0.353

0.286

0.349

0.304

0.306

0.320±0.029

231.478

32.901

0.999

Fit (%)

0.970

0.981

0.989

0.989

0.988

--

--

--

--

A

Reactions

29

30

A ED

PT

CC E

IP T

SC R

U

N

A

M