THERMODYNAMIC FORM OF KINETIC EQUATIONS AND AN EXPERIENCE OF ITS USE FOR ANALYZING COMPLEX REACTION SCHEMES Valentin N. Parmon Boreskov Institute of Catalysis Novosibirsk State University Novosibirsk, Russia, 630090, parmon@catalysis.ru May 30, 2012 Ghent, Belgium 1 Novosibirsk is the 3rd largest city in Russia (behind Moscow and St-Petersburg) Population >1,500,000 Universities and academies 30 Russia St-Petersburg Moscow Siberia Siberian Branch of the Russian Academy of Sciences Great logistic center (Trans-Siberian railway, International airport) Novosibirsk High-tech industries The highest density of science in Russia Novosibisrk Scientific Center – Akademgorodok 2 • Population 130,000 • 35 academic research institutes of SB RAS with ca. 9,000 employees • 7 chemical research institutes • Novosibirsk State University The Siberian school of mathematicians in chemistry (since the beginning of 60ths) M.G. Slin’ko (1914–2008) – initiator of the wide application of mathematical methods in catalysis G.S. Yablonsky A.N. Gorban’ – coupling of kinetic analysis with thermodynamics V.I. Bykov – analysis of reaction schemes with singularitues M.Z. Lazman 3 – generalization of analysis of complex reaction schemes G.S. Yablonsky et al. (1970s-2000s) • Rigorous results based on assumed detailed reaction mechanisms with the ideal mass-actionlaw dependences (1) Linear Theory (1970s-1980s) (2) Non-Linear Theory (1980s-2000s) 4 Linear theory • ONE-ROUTE CATALYTIC REACTION with the linear mechanism General expression for the steady state reaction rate (Yablonsky, Bykov, 1976) R Cy where Cy is a “cyclic characteristics” Cy = K+f+(C) – K–f–(C) Cy corresponds to the overall reaction presents a complexity of complex reaction pi K c j ji j i • MULTI-ROUTE LINEAR MECHANISMS (Yevstignejev, Yablonsky, Bykov, 1979) 5 Non-linear theory KINETIC POLYNOMIAL (Lazman, Yablonsky, 1980-2000s) It is considered as the most generalized form which includes Langmuir-Hinshelwood-Hougen-Watson equations and equations of enzyme kinetics as particular cases The kinetic polynomial of the one-route non-linear reaction scheme is BmRm +…+ B1R +B0Cy = 0 where R is the steady-state reaction rate, m are the integer numbers 6 V.N. Parmon is a lecturer of Novosibirsk State University in chemical kinetics and both classical (equilibrium) and, since 1995, non-equilibrium thermodynamics 7 Contents of the presentation: 8 • Introducing the thermodynamic form of kinetic equations • • Some interesting one consequences • Few practical application A problem of the “bottle neck” (limiting step) and “rate controlling step” of a stepwise reaction A Thermodynamic Form of Kinetic Equations is an Inevitable Step for the Successive Unifying the Languages of Chemical Kinetics and Chemical Thermodynamics Chemical Kinetics: the main parameters are concentrations, c, of reactants A and rate constants, ki, d cα vα = = f k i , c α f k i ,[A α ] dt Chemical Thermodynamics: the main parameters are chemical potentials, , of thermalized reactants A Note, however, that = f α α = o RTln c where is an activity coefficient 9 Introducing the Thermodynamic Form of Kinetic Equations 0 1 c [A ] exp RT For a substance A For an elementary reaction “ij” vij kij ci i i ij A j A 0 1 kij exp i i RT i 0 1 kij exp i RT where i i exp RT i exp ijni ij RT is a “chemical potential” of reaction group i 0 Gij0 Gi0 1 kB T i ij k ij exp exp i RT h RT Gij0 kB T exp RT h 10 0 1 exp i i i RT Properties of the Thermodynamic Form of Equations for an Elementary Reversible Reaction ij i A ji j A If ij is the chemical variable for reaction “ij” vij dij dt vij vij ij ni ji nj ij ni nj since ij = ji ! Indeed, for partial equilibrium of reaction ij when i i j j or vij 0 thus ij = ji 11 j exp i ni exp nj RT RT New proprietary definitions: ij0 kB T ij ji exp h RT – “truncated” rate constant of reaction ij which depends only on the properties of TS ni exp i RT – thermodynamic “rush” of reaction group i Why “rush” ? vij 0 if ni nj vij 0 if ni nj vij 0 if ni nj i i A 12 TS j j A Some Related Properties Inequality ni nj is equivalent to the positive value of affinity Arij of reaction ij: j i ni exp nj exp RT RT at i j i j i j Arij 0 ni nj is equivalent to Arij 0 ni nj is equivalent to Arij 0 Direction of reaction ij coincides with the sign of Arij ! 13 Thermodynamic Criterium of Kinetic Irreversibility of a Reaction i j j i j i i vij ij exp exp exp 1 ij RT RT RT RT Arij ijni 1 RT So, the rate vij of reverse reaction j i is negligible in respect to If Arij RT (far from equilibrium) " can be substituted by single " In this case double " " If Arij RT (the vicinity to equilibrium) then vrij 14 ijni RT Arij vij Consequence 1 For a set of stationary consecutive reactions which occur from left to right 1 2 N N1 P: R Y1 Y2 ... YN • obligatory R > Y1 > Y2 >...> YN > P! • the number of single arrows " " can not exceed the value of Ar m ! RT P Here Ar R – P is affinity of stoichiometric stepwise reaction R 15 Consequence 2 For any stoichiometric stepwise reaction “” P R which is linear in respect to its reaction intermediates Yi the rate is expressed in the same way as for elementary reaction: d v nR nP dt where nR exp R , nP exp P RT RT R and P are initial and final reaction groups; is an algebraic combination of ij and, in some cases, thermodynamic rushes of “external reactants” from either initial of final reactions groups Note: Stoichiometric stepwise reaction means steady state occurrence of the reaction in respect to its intermediates Yi This relation in catalysis is known as a Horiuti–Boreskov equation! 16 Simple Example 1 Stepwise reaction P R occurs according to scheme 1 Y R 2 P Y In the steady state in respect to Y d[Y] 1 R Y 2 Y P 0 dt R 2P Y 1 1 2 d d[R] d[P] 1 R Y 2 Y P dt dt dt 1R 2P 12 1 R R P R P 1 2 1 2 v where 17 12 1 2 Simple Example 2 P R1 R2 Stepwise reaction occurs according to scheme 1 Y R1 2 P Y R2 In the steady state in respect to Y d[Y] 1 R1 Y 2 Y R2 P 0 dt R 2P Y 1 1 1 2R2 1R1 2P d d[P] v 2 Y R 2 P 2 P R dt dt 1 2 2 12 R1R 2 P R1R 2 P 1 2R 2 where 18 12 1 2R 2 For an arbitrary set of monomolecular transformations of intermediates there is a total analogy with an electrotechnical equivalent scheme! Stepwise process occurs according to scheme Electrotechnical analog P R {Y} P R i vij ij Yi Yj v Yi d[Yi ] dt Iij Y Y 0 ij j i v R P j Ii 1 Ui Uj R ij 1 R U j I i ij Uj 0 1 UR UP R 1 is calculated in the same way as : R is an algebraic combination of ij like R is that of Rij 19 Main basis for “linear” non-equilibrium thermodynamics Flux Ji of a parameter ai dai Xi , dt Ji where Xi is thermodynamic driving force for ai For a complex system Jj L j ij Xi , where Lij are the Onsager’ coefficients of interrelation. A sequence: existence of the Raleigh-Onsager dissipation function P T diS dt JX i i i L i j ij Xi X j 0 According to the Prigogine theorem, P is the Lyapunov’ function which reaches a positively defined minimum at the stationary state of the system (when Jj = 0) 20 Consequence 3: Existence of the Lyapunov’ functions which are positively determined and minimazing at the steady state in respect to intermediates even far from equilibrium for any reaction schemes which are linear in respect to intermediates Example 1: Stepwise reaction P R occurs via the scheme {Y} P R i where {Yi} means an arbitrary set of monomolecular transformations of Yi i v Yi 21 Ri R Yi 2 1 ij Yi Yj 2 i j 2 jR Yj P j 2 d[Yi ] 1 Ri R Yi ij Yi Yj iP Yi P 0 dt 2 Yi j Physical meaning of the Lyapunov’ function for an arbitrary set of monomolecular reactions far from equilibria Stepwise process occurs according to scheme Electrotechnical analog P R {Y} P R i vij ij Yi Yj v Yi d[Yi ] dt Iij Y Y j ij v R P i j Ii 1 Ui Uj R ij 1 j R Ui Uj ij I Thus, the Lyapunov’ function 1 UR UP R Y Y 2 i j ij i j corresponds to the power W of the dissipation of Ohmic heat in the electrical circuit 2 1 W Iij Ui Uj Ui Uj i j i j R ij 22 Example 2: Stepwise reaction P occurs via the scheme R1 R2 1 Y R1 2 P Y R2 1 R1 Y vY 2 2 R2 Y R 2 P 2 d[Y] 1 1 R1 Y 2 Y R2 P 0 dt 2 Y Conclusions: 1. The Lyapunov function exists for any stepwise reactions which are linear in respect to intermediates 2. Steady state of above reaction is stable 23 Consequence 4: According to the Prigogine theorem all systems near thermodynamic equilibrium have the stable steady state. All stepwise reactions linear in respect to intermediates have their Lyapunov’ functions and thus are also stable A contrary example: Stepwise reaction R P occurs via the nonlinear autocatalytic scheme in respect to Y: 1 2Y R Y 2 Y P The Lyapunov function does not exist ! There are two steady states 0, any R d[P] v 2 Y 2 2 dt R , R 2 1 1 The steady state in respect to Y can be nonstable ! Thus, the necessary conditions for loosing the stability of the steady state of a kinetic scheme: 24 (1) (2) As least one elementary reaction has to be kinetically irreversible This elementary reaction has to be non-linear in respect to the intermediates Necessary conditions for oscillation of the concentration of reaction intermediates: 25 " " in the • far from equilibrium (at least one single reaction scheme) • at least two reaction intermediates • at least one step which is nonlinear in respect to intermediates Consequence 5: It is possible to write modified Onsager’ (the Horiuti-BoreskovOnsager) equations of interrelation of parallel stepwise chemical reactions A simple example: Parallel step-wise reactions occur via mechanism 1 P1 R R 2 P2 R At the steady state Thus, d[P1] 2 Y P1 11 R P1 12 R P2 dt d[P2 ] 3 Y P2 21 R P1 22 R P2 dt 11 = 12/(1 + 2 + 3) > 0 22 = 13/(1 + 2 + 3) > 0 12 = 21 = -23/(1 + 2 + 3) 26 Y d[Y] 1 R Y 2 Y P1 3 Y P2 0 dt where 1 2 P1 3 P2 In a general case for parallel stepwise reactions i Pi R j v i Here d i dt R ij j R i exp R j RT , X Pj j j ij j Pi exp Pi RT ii > 0 Note: ij is not obviously symmetrical in respect to indexes i and j as it is the case for reprocisity coefficients Lij in the classic Onsager equations in the vicinity of equilibrium Ji 27 L j ij Xj The problem of “rate controlling” (“rate determining”) step and “rate limiting” step (“bottle neck” of the stepwise reaction) An unambigous interpretation of the notion “rate determining” (rate controlling) step by IUPAC Rate controlling factor ln v CF lnk ij k ,li,m j lm but In the thermodynamic representation f Gij o kB T ij exp h RT ln v CF ln ij 28 ln v CF ln , – contains parameters of only the thermalized reactions groups and reactants v f ij ,n i lm ,li,m j vij ij n i nj – contains parameters of only the transient states ni exp i exp i RT RT Thus, r Gij0 kij k ji exp RT – rate controlling factor of transient states – rate controlling factor of the reactant How to define correctly the “rate limiting” step (the “bottle neck”)? 29 The “rate determining step” and “bottle neck” in a consequtive monomolecular reaction 1 2 3 n1 P R Y1 Y2 ... Yn v 1 where 1 1 lim i 1 i d d[R] d[P] R P dt dt dt n 1 lim mini So rate-determining step is the step with minimal i Note: in the steady state Thus Yi Yi1 v i Yi Yi1 v i For i = lim the value of Y Y i i 1 for i = 0,…, n+1 is maximal ! It means that the “bottle neck” (limiting step) is the step with the maximum drop of Yi Yi1 ! 30 Application to catalytic reactions A simple example: P R occurs via catalytic Michaelis–Menten scheme 1 K1 R K 2 KP K1 where K and K1 are free catalytically active site and the catalytic intermediate The balance equation [K] [K1] [K]0 can be rewritten: Thus where K1 Here 31 K0 K0 K0 1 K0 exp K exp K1 exp RT RT RT K K1 K0 K0 K0 and K0 corresponds to K at K 1 exp RT 1 K K0 RT ln 1 K1 K0 RT ln K K1 K01 RT ln K1 Finally, at the steady state in respect to K1 d v 1 2 dt 1 2 K0 R P 1R 2P 1 2 1 K0 R P 1 K1 K • At small extent of occupation of the active site with catalytic intermediates K K 1 1 12 K R P 1 2 0 v 0 and does not depend on standard thermodynamic parameters K1 of K1 • At large extent of occupation of the active site with catalytic intermediates K K 1 v 0 depends on K1 32 1 12 1R 2P K0 K0 K0 R P exp 1 RT Note that for K K 1 one can have the situation when the 1 rate determining step does not coincide with the bottle neck ! v Let: R 12 1R 2P P and 1R K0 K0 K0 R P exp 1 RT 2P In this situation v 2K0 K0 K0 R P exp 1 RT independently on whether 1 2 or 1 2 ! So, the rate determining step is always step 2 But obligatory the “bottle neck” is the step with minimum i ! 33 An unexpected conclusion: there are situations when the rate-limiting step can not be the rate-controlling step! 34 Conclusions • The thermodynamic form of kinetic equations allows a dramatic simplification of analysis of complex reaction schemes • Indeed, the main application of this approach is possibility to extend fruitful and systematic analysis of chemical reaction schemes for the area “far from equilibrium” • Among few principal problems which are resolving via this approach this is a mathematically correct definition of the “bottle neck” (the limiting step) of a stepwise reaction and “rate determing step” Unexpectedly, for some particular cases (e.g. for catalytic reactions) these steps can not coinside 35 Few examples of practical interest 36 An example of a practical application: Super low temperature of melting of active component of operating metal catalysts due to their oversaturation with carbon Electron microscopy “in situ” videotape of Fe–C fluidized particles migration over amorphous carbon support at 650 °C time (sec): 300 A o The melting temperature is 500 °C (!) lower than that of the Fe–C eutectics time (sec): 37 O.P.Krivoruchko, V.I.Zaikovskij, K.I.Zamaraev, Dokl.Akad.Nauk, v.329, 744 (1992) (in Russian) Formation of Metastable Oversaturated Solutions of Carbon in Metals at Catalytic Graphitization of Amorphous Carbon Camorph Cgraphite met G –12 kJ/mol ( >RT) Melting Temperatures, oC Camorph solution of C in metal Cgraphite c(amorf) > c(in metal) > c(graphite) Result: steady-state concentration of xC in metal >> concentration of C in stable eutectics equilibrium steady pure eutectics metal metal with graphite state Fe 1539 1145 640 Co 1493 1320 600 Ni 1453 1318 670 If the rate determining step is formation of graphite from the melt: xC (eq. with amorph. C) = xC (eq. with graphite)exp(–GR/RT) 4xC(eq. with graphite) 4xC (eq. eutectics) ln 1 XC mH 1 1 T Teut Tst st 500 – 900 C ! R T T0 Hm and To are the melting heat and melting temperature for pure metal 38 V.N.Parmon. Catalysis Letters, 42, 195 (1996), O.P.Krivoruchko, V.I.Zaikovskij (1995) Metastable Phase Equilibria for Fe–C Systems during Occurrence of the Catalytic Reaction T, K T 2000 Graphite liquids Parameters, influencing the melting temperature: 1. Oversaturation: the Schröder equation Solution of Fe in C 1600 Tx = TomH/{mH – RToln(1 – x)} eutectics (T = 1420K, x = 0.173) 2. Metal particle size r: Schröder 1200 Solution of C in Fe steady state (920K) 3. Size r’ of a crystallization 800 center (= size of the catalyst particle) Fe3C Fe2C 0 50 Content of C (mol. %) 39 2Vm Tr T exp r mH 100 2 Vm xr x exp r RT Formation of Filamentous Carbon together with Hydrogen at the Moderate-Temperature Catalytic Pyrolysis of Methane and Low Hydrocarbons catalysis CnHm C 2H2 450650 C o 500A o Ni-catalyst 1000 A Ni/Cu-catalyst The weight of the catalyst can be increased by a factor of 400 due to formation of carbonaceous filamentous material The growth of the filament corresponds to diffusion of carbon through the active component with D>10–10 cm2/s 40 L.B.Avdeeva, V.A.Likholobov, G.G.Kuvshinov, at al. (1994) Size effects in catalysis over metal nanoparticles 2O3 CО + O2Au/Al CO2 Pt/Al2O3 CН4 + 2 O2 CO2 + 2 Н2О 430 °C (s–1))10 TOF УКА 10–4 (сек -4 0.07 0.05 -1 TOF (s–1) 0.06 0.04 0.03 0.02 0.01 400 °C 140 120 100 80 60 40 20 0.00 0 0 2 4 <d>, nm 6 8 10 0 5 10 15 <d>, nm <d>, нм 20 Conclusion: There may occur size effects in catalytic reactions, which are many time increase in the activity of metal catalysts upon a decrease of the active component particles in size to several nanometers I.41Beck, V.I. Bukhtiyarov, I.Yu. Pakharukov, V.I. Zaikovsky, V.V. Kriventson, V.N. Parmon, Journal of Catalysis 268 (2009) 60-67 25 Influence of the active component particle size on the catalytic activity (an energy correlation approach) A B Mechanism: A+K K1 K1 B+K (1) (2) The increment of chemical potential of a nanoparticle of radius r 2 V r = r Here – surface excess energy, V – molar volume of the catalyst active phase, TS1 = æ r æ < 1 is the Brønsted-Polyany correlation coefficient (1 æ )Δr v exp Result: TOF vΣr d[B] dt Σ RT æ Δr v exp Σ RT 42 V.N. Parmon, Doklady Physical Chemistry, vol. 413 (2007) 42-48 TS2 at low coverage with K1 at large coverage with K1 Correlation of the measured TOF values for the complete CH4 oxidation over Pt/Al2O3 with the Pt size at temperature 700 K and of the apparent activation energies Ea with the reciprocal to the active component size (diameter) d 0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,0 –1,0 140 120 log(TOF) (TOF in s–1) –1,4 100 80 –2,2 60 40 –2,6 lg (TOF) = 3,304 (1/d) – 2,981 Ea, kJ/mol –1,8 20 0,0 0 0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,0 1/d, nm–1 For both lines the correlation coefficient is the same: æ 0.75 I.43Beck, V.I. Bukhtiyarov, I.Yu. Pakharukov, V.I. Zaikovsky, V.V. Kriventson, V.N. Parmon, Journal of Catalysis 268 (2009) 60-67 Thermodynamic conjugation of parallel chemical transformations via a common catalytic intermediate K B A B K C A C A X The Horiuti-Boreskov-Onsager coupling equations: A C* d[B] [B] [C] kBB[A] 1 kBC[A] 1 dt K [A] K [A] B C {X} B B C C Reaction coordinate {X} – catalytic intermediates d[C] [B] [C] k CB[A] 1 k [A] 1 CC dt K [A] K [A] B C kij – formal rate constants kB and KC – equilibrium constants Conclusion: To change selectivity of a catalytic process one has to generate some thermodynamic driving forces 44 V.N. Parmon, Thermodynamics of Non-Equilibrium Processes, Elsevier, 2010 Thermodynamic control of selectivity at the decomposition of methanol Two independent stepwise reactions – two channels of decomposition: 2 CH3OH CH3OH CH3OOCH + 2 H2 CO + 2 H2 {X} Methylformate + 2 H2 (I) 2 CO + 4 H2 (II) Reaction coordinate Gibbs energy Gibbs energy 2 Methanol (I) (II) 2 Methanol {X} 2 CO + 4 H2 (II) Methylformate + 2 H2 (I) Reaction coordinate {X} – catalytic intermediate Conclusion: an increase in the partial pressure of CO has to result in improving the selectivity in respect to methylformate 45 An example of an important practical application: Development of principally new one-step catalytic processes of direct insertion of methane higher hydrocarbons Usually: hydrocarbons main products CH4 as a byproduct Due to existence of Onsager’s interrelation, one can reverse the direction of the process of CH4 formation Now: hydrocarbons + CH4 heavier hydrocarbons Examples: Process “Bicyclar” CH4 + C3,C4 alkanes aromatics + 5 H Process “Biforming” CH4 + linear C5+ aromatics + 5 H2 46 Putative one-stage processes for conversion of light paraffins CH4 and C3–C4 (methane and propanebutanes) to aromatic compounds Reactions of light hydrocarbons T*, K 1. 6 CH4 C 6H 6 + 9 H 2 1630 2. 2 C3H8 C 6H 6 + 5 C 2 H 6 760 3. 2 n-C4H10 p-C6H4(CH3)2 + 5 H2 4. C3H8 + n-C4H10 5. 3 C2H6 C6H5CH3 + 5 H2 C 6H 6 + 5 H 2 6. CH4 + 2 C3H8 8. CH4 + C2H6 + C3H8 9. CH4 + 3 C3H8 710 930 C6H5CH3 + 6 H2 7. CH4 + C3H8 + n-C4H10 800 p-C6H4(CH3) + 6 H2 C 6H 6 + 6 H 2 C10H8 + 10 H2 880 1060 940 830 Observation: Aromatization of C2–C4 paraffins is accompanied by the methane co-production. 47 Yield of aromatic hydrocarbons, t/t C4 Performance of the BICYCLAR process depending on the C1/C4 ratio CH4 + 2 C3H8 C6H5CH3 + 6 H2 CH4 + 3 C3H8 C10H8 + 10 H2 1.5 1.0 0.5 0 0 3 6 9 12 Molar ratio C1/C4 15 18 Catalyst Zn-ZSM-5, temperature 550 °C The coupled conversion of butane and methane allows the yield of aromatic hydrocarbons to be 2.5 times increased – up to 1.7 tonn per 1 tonn of involved C4 48 G.V. Echevsky, E.G. Kodenev, O.V. Kikhtyanin, V.N. Parmon, Appl. Catal. A: General 258 (2004) 159-171 An example of a practical application: Natural selection in simple autocatalytic systems at diminishing the concentration of food R follows in one-directional progressive evolution of the system i 2Y R Yi i ti Yi P There are two steady states: Yi(1) R ti R R cri i Yi(2) 0 At diminishing the concentration of food R, one Yi proceeds a consecutive and irreversible (due to disappearance of seeds) “death” of all autocatalysts with the larger values of R cr ti i Y3(1) Y2(1) i Thus, a one-directional and progressive (toward diminishing the parameter Rcri) natural selection takes place in the system. This is analogous to appearance of a prototype of biological memory Yi(2) An extremely important conclusion: existence of a prototype of biological memory in the absence of RNA or DNA ! 49 V. Parmon, Doklady Phys. Chem., 377, 4 (2001) 510-515 Y1(1) Thank you for your attention ! 50