Limitations in Thermal Degradation Modelling and Kinetic Parameters - PowerPoint PPT Presentation

Limitations in Thermal Degradation Modelling and Kinetic Parameters Evaluation for Polymeric Blends in Dynamic Thermogravimetry Presented by: Dr. Abdul Rehman Khan – Consultant Environment & Life Sciences Research Center Kuwait Institute for Scientific Research 5 th Technology Innovations Conference & Exposition. 2 nd November 2014, Kuwait

Presentation Agenda 1. Introductory Remark. 2. Motivation of Work and Benefits to Clean Fuels. 3. Used Models in Literature. 4. Case Study: Degradation of PET/PMMA & Integral Method Development. 5. Conclusions & Future Work

Introductory Remark • Thermal degradation of polymers is arguably one of the hottest topics in engineering disciplines. • Typically, results are conflicted especially in micro-scale and in particular, dynamic thermogravimetric analysis (TGA). • Such discrepancies result from different factors, such as: – Experimental setups: Different inert atmospheres (at different scales), temperature ranges, sample amounts, heating rates (b) and pressures. – Adequacy of the kinetic model: Modelling approach and assumptions. – Thermal lag (  T): Heat transfers problems.

Introductory Remark • Waste is accumulating in Kuwait with NO GOVRMT scheme to handle. • Plastic solid waste (PSW) is estimated at 200 Mtpa. (2013). • PSW is typically shipped abroad (exported)/recycled in private company(ies) for profit. • THIS IS A WASTE!!! • Being a crude oil product, plastics encompass energy that should be taken advantage of.

Table 1: Calorific Value of Major Polymers in Comparison to Common Fuels. Item CV (MJ kg -1 ) Item CV (MJ kg -1 ) PE 43.3-46.5 Gas Oil 45.20 PP 46.50 Heavy Oil 42.50 PS 41.90 Petroleum 42.30 Kerosene 46.50 Household PSW mixture 31.80 Thermolysis Main advantages include: Hydrogenation Gasification Pyrolysis 1. Minimal pre-treatment. VEBA - 2. PCs production & Integration. Kiener Noell Oel Eisenmann Winkler ABB VKE Texaco BASF Lurgi SVZ BP 3. Waste disposal solution. 4. Sustainable energy source. Naphtha &High Oil Oil Boiling Oil

Motivation • Today’s refining capacity of Kuwait is around 936 mbpd divided. • Post CFP, the refining capacity of the country will decrease to a total of 800 mbpd. • It is anticipated that the NRP will process 615 mbpd of Kuwait Export Crude (KEC, API ≈ 30). Total refining capacity will be: 1,415 mbpd. Table 2: Major products specs post CFP (Sulfur ppm). • Products from TCT Product Current Spec. Post CFP units is the answer. Such include H2, Full Range Naphtha 700 500 C3, C4, etc. This Gasoline (All Grades) 500 10 will intensify Gas Oil 1 (Including Domestic Use) 2000-5000 10-500 production of Gas Oil 2 500 10 chemical feedstock Gas Oil 3 (New Grade) - 10 from a renewable energy source. Fuel Oil (%) 4.5 1

Problem Statement • Polymers, in the form of plastics, are fed to pyrolysis reactors as a fraction of MSW. They are a mixture of polymers, not just a single one . Hence, predicting their degradation behavior and evaluating their kinetic parameters in a blend is a must.

Established Models Most common kinetic degradation models are isoconversion ones: 1. Ozawa-Flynn-Wall: • One of the most used expressions in literature. • This method is considered to be the most exact. • Assumes a first order kinetics ( n =1). E m       a ln( ) ln( AE / R ) 5 . 33 ( 1 ) 1 . 05 a m RT o 2. Friedman’s method:  1 m    ln( )( dm / dt ) ln( A ) ln f ( ) E / RT m m o o

PET/PMMA Degradation as a blend • Blends of PET/PMMA (traded under the name of Ropet) are typically used in electrical applications. • Hence studying their thermal degradation and stability determines optimal operating condition of these blends avoiding electrical overshoots in electronics. • Thermogravimetric analysis (TGA) was carried out for the blends with pure dry nitrogen purge of 20 cm 3 /min. Four  were used: 5, 10, 15 and 20 o C/min. •

Mathematical Derivation The expression of degradation could be written after rearranging the denominator as: dx   P k dt n x p The results presented in this work reflect first order kinetics. Integrating the resulting expressions results in the following for each polymer in the blend: t dx  x    pA pA k dt A x 1 pA 0  x x N  P p (exp)  (th) Objective Function (O.F.) min x  i 1 p (exp) t dx  x    pB pB k dt B x 1 pB 0 For a given blend of known composition (x A is the PET fraction in the blend), the overall cumulative weight loss expression will be given as          ( E / RT ) ( E / RT ) x x exp A e t ( 1 x ) exp A e t a 1 a 2 p A 1 A 2

Results & Observations • Several non-isothermal pyrolitic degradation TG curves have been modelled in this study. • It was noticed that with the increase of PET fraction in the blends, the TG curve showed a shift to higher value in the degradation temperature till 75 wt%. • The 90/10 (wt/wt%) blend of PET/PMMA started decomposing at almost 600 K (beginning of the first shoulder incline). • Virgin PET and PMMA typically started decomposing at temperatures around 630 K and 560 K, respectively.

1 1 Exp. 0.9 0.9 PET/PMMA (50/50 wt/wt%) Theor. PET/PMMA (50/50 wt/wt%) 0.8 0.8 Exp. 0.7 0.7 Theor. 0.6 0.6 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 0 500 550 600 650 700 750 800 500 600 700 800 900 Temperature (K) Temperature (K) Model vs. experimental results for PET/PMMA blend Model vs. experimental results for PET/PMMA (25/75 wt/wt%) at  5= o C/min. blend (50/50 wt/wt%) at  10= o C/min. • The structural and physical properties of blends of both polymers affect the TG curves, where the abundance of PET in the 90/10 blend delays the degradation to a point where the material acts almost like a virgin PET (or pseudo virgin material).

Apparent activation energy evaluated for PET (E a1 ) and PMMA (E a2 ), pre- exponential factors (A 1 and A 2 ) and regression coefficient between experimental and theoretical fits. PET/PMMA (25/75) (wt/wt%) PET/PMMA (50/50) (wt/wt%) A 1 (min -1 ) A 2 (min -1 ) r 2 A 1 (min -1 ) A 2 (min -1 ) r 2 E a1 (kJ/mol) E a2 (kJ/mol) E a1 (kJ/mol) E a2 (kJ/mol)  = 5 o C/min 1.1 x 10 19 1.73 x 10 9 3.9 x 10 15 2.7 x 10 10 240 140 0.98 250 150 0.99 1.03 x 10 17 2 x 10 11 1.82 x 10 16 8.1 x 10 10  = 10 o C/min 230 130 0.98 240 140 0.99 3.2 x 10 18 2.7 x 10 8 2.2 x 10 17 2.28 x 10 7 220 120 0.98 230 130 0.97  = 15 o C/min 5.5 x 0 18 1.33 x 10 4 4.2 x 0 17 2.7 x 10 11 210 110 0.98 220 120 0.99  = 20 o C/min PET/PMMA (75/25) (wt/wt%) PET/PMMA (90/10) (wt/wt%) A 1 (min -1 ) A 2 (min -1 ) r 2 A 1 (min -1 ) A 2 (min -1 ) r 2 E a1 (kJ/mole) E a2 (kJ/mole) E a1 (kJ/mol) E a2 (kJ/mol)  = 5 o C/min 3.12 x 10 14 6.23 x 10 8 2.32 x 10 16 6.4 x 10 7 260 160 0.99 270 170 0.99 2.89 x 10 15 2.92 x 10 9 2.7 x 10 16 5.9 x 10 7  = 10 o C/min 250 150 0.99 260 150 0.99 7.6 x 10 16 6.62 x 10 7 6.23 x 10 15 2.1 x 10 9 240 140 0.98 250 150 0.98  = 15 o C/min 1.27 x 0 17 4.28 x 10 6 1.4 x 0 16 1.07 x 10 10 230 130 0.99 240 140 0.99  = 20 o C/min

Conclusions • A general mathematical expression based on the integral solution of different PET/PMMA blends for non-isothermal (dynamic) thermogravimtry (TG) has been developed. • The model results show good agreement with experimental values depicting the true pyrolitic reaction mechanism. • The apparent discrepancies are attributable to melt mixing resulting in the formation of different phases. • Thermal lag was caused by the evaporation of volatile degraded products (heat absorption) and also greatly influenced by thermal characteristics of blend of polymers ensuing in the observable deviation among experimental and model results.

Thank you Dr. Sultan Al-Salem Dr. Abdul R. Khan (Consultant-ELSRC) (PRC)