6126
J. Phys. Chem. B 2005, 109, 6126-6129
Thermal Decomposition Kinetics of Hydrazinium Cerium 2,3-Pyrazinedicarboxylate Hydrate: A New Precursor for CeO2 Thathan Premkumar,†,‡ Subbiah Govindarajan,*,† Andrew E. Coles,§ and Charles A. Wight§ Department of Chemistry, Bharathiar UniVersity, Coimbatore 641046, India, and Department of Chemistry, UniVersity of Utah, 315 South 1400 East, Room 2417, Salt Lake City, Utah 84112-0850 ReceiVed: December 2, 2004
The thermal decomposition kinetics of N2H5[Ce(pyrazine-2,3-dicarboxylate)2(H2O)] (Ce-P) have been studied by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC), for the first time; TGA analysis reveals an oxidative decomposition process yielding CeO2 as the final product with an activation energy of ∼160 kJ mol-1. This complex may be used as a precursor to fine particle cerium oxides due to its low temperature of decomposition.
1. Introduction Coordination compounds containing positively charged ligands are rare and are therefore of considerable interest.1-3 Among the known cationic polyamine ligands, the hydrazinium (+1) cation (N2H5+) is unique since the site of the positive charge is immediately adjacent to the N-donor atom. There has been a significant amount of interest in the chemistry of rare earth complexes of hydrazine with carboxylic acid, due to their wide use as precursors for fine particle metal oxides.4,5 Our recent interest in hydrazine chemistry began with the preparation and thermal degradation of lanthanide complexes of nitrogencontaining heterocyclic dicarboxylic acids, particularly 2,3pyrazinedicarboxylic acid, 1. What makes this acid interesting and versatile is that it has six sites for potential coordination to a metal atom, but without severe distortion it can only serves as a mono- or bidentate chelating (tetradentate) ligand.6-8
Several lanthanide (La, Ce, Pr, Nd, and Dy) hydrazine complexes with carboxylate anions such as oxalate,9 phthalate, 4and N H COO- 5 have been reported in the literature, and their 2 3 spectral and thermal properties have been studied. While all of these compounds undergo normal or mild exothermic decomposition giving the respective metal oxides (M2O3) as the end product, the cerium compounds show high exothermicity to give CeO2. Further, to understand this, no systematic study has been done on the kinetics of thermal decomposition of lanthanide hydrazine compounds in general and cerium compounds in particular. Hence, with this in view, for the first time, this work was undertaken to study the decomposition kinetics of hydrazinium cerium 2,3-pyrazinedicarboxylate, and the results are presented. * Author to whom correspondence should be addressed. E-mail:
[email protected]. † Bharathiar University. ‡ Present address: Laboratory of Applied Macromolecular Chemistry, Department of Materials Science and Engineering, Gwangju Institute of Science and Technology, 1 Oryong-dong, Buk-gu, Gwangju 500-712, South Korea. § University of Utah.
Cerium oxide is widely used as a promoter in the so-called “three-way catalyst” for the elimination of toxic exhaust gases in automobiles.10 The promoting effect of cerium oxide was originally attributed to the enhancement of the metal dispersion and the stabilization of the support toward thermal sintering.11 It was later shown that cerium can also act as a chemically active component, working as an oxygen store, releasing oxygen in the presence of reductive gases and removing it by interaction with oxidizing gases,12 and participating in the water-gas shift reaction13 or the decomposition of nitrogen oxides.14 More recent efforts are devoted to elucidating the participation of cerium in an important metal/cerium interaction induced by the establishment of contacts between both components, which strongly affects their redox and, as a consequence, catalytic properties.15,16 2. Experimental Section Following our previous work on the synthesis of N2H5[Ce(pyrazine-2,3-dicarboxylate)2(H2O)] (Ce-P),17 now we describe its thermal decomposition kinetics properties. This complex is insoluble in water and most of the common organic solvents. The hydrazine content of the complex was determined volumetrically using a standard KIO3 solution (0.025 M) under Andrews’ condition.18 The cerium, after destroying the organic part and hydrazine by treatment with concentrated HNO3 and evaporating the excess HNO3, was determined volumetrically by EDTA titration.18 The IR spectrum was recorded as a KBr pellet with a Perkin-Elmer model 597 spectrophotometer in the 4000-400 cm-1 range. Yield: 80%. IR (KBr pellet): νasy(COO), 1630 cm-1; νsy(COO), 1385 cm-1; νN-N (N2H5+), 960 cm-1. Hydrazine: found, 5.48%, calcd, 5.51%. Cerium: found, 24.58%; calcd, 24.66%. Thermogravimetric analysis (TGA) experiments were performed on a Rheometric Scientific TGA 1000M+ module. The experiments were performed on ∼0.5 mg Ce-P samples, which were placed in an open Pt pan. Compressed air flowing at a rate of 50 mL min-1 served as the purge gas. Several experiments were carried out at heating rates of β ) 1.0, 2.0, 5.0, 10.0, and 20.0 °C min-1 over a temperature range of 25600 °C. DSC curves were collected on a Mettler-Toledo DSC821 module. Differential scanning calorimetry (DSC) traces were collected for ∼0.3 mg Ce-P samples sealed in high-
10.1021/jp0445223 CCC: $30.25 © 2005 American Chemical Society Published on Web 03/12/2005
TGA and DSC of N2H5[Ce(C4N2H2(COO)2)2(H2O)]
J. Phys. Chem. B, Vol. 109, No. 13, 2005 6127
pressure crucibles. Al pans (40 µL) were used in DSC experiments with compressed air flowing at a rate of 50 mL min-1. Non-isothermal DSC experiments employed heating rates of β ) 1.0, 2.0, 5.0, 10.0, and 20.0 °C min-1. All of the uncertainties given in this report represent 95% confidence limits for the indicated quantity. In most cases, the uncertainties arise from statistical uncertainties in measured quantities. Uncertainties in the activation energies also include systematic errors arising from errors in assumptions of the kinetic equations. Kinetic Analysis. The decomposition kinetics of a heterogeneous solid-state reaction are typically described by the equation19
dR ) k(T)f(R) ) A e-E/RT f(R) dt
R
e-E/RT(t) dt ) AJ[E,T(tR)] (2)
where R describes the values pertaining to a particular extent of conversion. Model-free isoconversional methods20 use g(R) to determine activation energies from thermal analysis data as a function of R. This approach assumes that the reaction model is independent of the temperature program. From this assumption, the right-hand sides of eq 2 are all equal for experiments conducted under different heating programs. To solve the equation for n such experiments, ER can be determined for any chosen R by solving for the activation energy that minimizes the function
S2(ER) )
1
n
n
∑∑ n(n - 1) i-1 j*i
(
J[ER,Ti(tR)]
J[ER,Tj(tR)]
)
∫t t -∆R e-E /RT (t) dt R
R
R
i
1 Af
-1
Typically, the extent of reaction is broken down into 10-50 different segments, and the analysis is repeated for each segment to obtain the activation energy as a function of the extent of reaction. For a complete description of the kinetics (e.g., to make kinetic rate predictions), it is also necessary to extract the values of A and f(R) from the data. If the number of reaction segments is sufficiently large that f(R) can be considered to be constant,
R
R
Af )
(5)
∆R J[ER,Ti(tR)]
(6)
where J is the Arrhenius integral for the given segment from eq 4. Thus, for a given segment, it is straightforward to determine the product Af. Repeating this procedure for each segment builds up the R-dependence over the entire reaction, Af(R). Because the separation of the constant A and the function f(R) is arbitrary, we represent this combined quantity as A(R). Confidence intervals for the activation energies have been estimated using the following statistical treatment.21 The Jintegrals (eq 4) for any given segment (tR-∆R f tR) should be equal for all experiments under the assumption that the reaction model is independent of the temperature program, T(t) (eq 2). Thus, the variance, S2, in eq 3 should be independent of the number of experiments, apart from statistical fluctuations in the experimental data. The optimal value of the activation energy, Emin, is determined by minimizing S2. For this case, the statistics
Ψ(ER) )
(3)
(4)
R t dR′ ) ∫t -∆R e-E/RT(t) dt ∫R-∆R
and the integration over the segment and rearrangement yields
2
where the subscripts i and j denote indices for two experiments performed under different heating programs. The integral, J, in this equation is solved numerically with the trapezoid rule. To account for the variation of activation energy with the extent of conversion, the experimental kinetic traces are broken up into segments (e.g., each 5% of the extent of reaction is analyzed independent of other segments). ER for each segment is calculated by numerical integration of the T(t) data over that segment
J[ER,Ti(tR)] ≡
f, within a given segment, then separation of variables in eq 1 gives
(1)
where R is the extent of reaction and f(R) is the reaction model, which gives the dependence of the reaction rate on the extent of reaction. For gas-phase reactions, the reaction model usually takes a simple first-order or second-order form. In the second part of eq 1, the Arrhenius equation has been substituted for the temperature-dependent rate constant, k(T). Rearranging this equation by separating R and t and integrating leads to an equation that describes the integrated reaction model, g(R) t dR ) g(R) ) A ∫0 ∫0R f(R)
Figure 1. Open-pan non-isothermal TGA traces of Ce-P samples acquired at β ) 1.0, 2.0, 5.0, 10.0, and 20.0 °C min-1.
S2(ER) Smin2
(7)
have the F-distribution. These statistics allow the confidence limits for Emin to be determined from estimating the confidence limits for the variance, S2min. For the condition
Ψ(ER) < Fl-p,n-l,n-l
(8)
where Fl-p,n-l,n-l is a percentile of the F-distribution for (1 p) × 100% confidence probability, the p × 100% confidence interval for S2min can be calculated. Then the upper and lower confidence limits (ERlow and ERup, respectively) for ERmin can be estimated as the value of ER for which Ψ(ER) ) Fl-p,n-l,n-l. Error limits given for ER in this paper were determined using p ) 0.05 and therefore represent 95% confidence intervals. 3. Results TGA curves exhibit a three-step mass loss that approaches 33% of the initial mass (Figure 1). The first step of mass loss occurs between 125 and 220 °C; the second step takes place between 220 and 318 °C; the major mass loss in the final step occurs between 318 and 350 °C and includes around 66% of the total mass loss.
6128 J. Phys. Chem. B, Vol. 109, No. 13, 2005
Figure 2. ER versus R plot generated by applying isoconversional analysis to open-pan TGA traces acquired at β ) 1.0, 2.0, 5.0, 10.0, and 20.0 °C min-1.
Premkumar et al.
Figure 5. Third step ER versus R plot using DSC traces acquired at β ) 1.0, 2.0, 5.0, 10.0, and 20.0 °C min-1
Figure 3. Specific ER versus R plot of Figure 2 for R > 0.4 Figure 6. Plot of the integrated and normalized exotherm as a function of β
3750 J g-1. The fairly constant heat release of ∼3000 J g-1 is largely independent of β and the sample mass. 4. Discussion
Figure 4. Non-isothermal DSC trace collected for open-pan Ce-P at β ) 10.0 °C min-1.
Application of the isoconversional method22 to the collected TGA data for β ) 1-20 °C min-1 yields an ER versus R dependence that exhibits a climbing activation energy to a maximum of ∼225 kJ mol-1 for R < 0.4 and corresponds to activity during the first two steps of mass loss. As the reaction proceeds for 0.4 < R e 0.95, ER decreases to ∼160 ( 35 kJ mol-1, indicating the third and final step of mass loss (Figures 2 and 3). A typical DSC curve collected in an open 40 µL Al pan for β ) 10.0 °C min-1 is shown in Figure 4. The common thermal feature in all of the DSC curves is a similarly structured exotherm appearing between 325 and 375 °C; it is strongly temperature related to the third step of mass loss observed from the TGA traces. Application of the isoconversional method to the exotherm for β )1.0-20.0 °C min-1 gives an ER dependence with a practically constant activation energy23 of ER ≈ 135 ( 25 kJ mol-1 for 0.2 < R < 1.0 (Figure 5). Figure 6 shows that the integrated and normalized heat release of the exotherm in DSC traces remains fairly constant during experiments of varying β. The maximum total heat release is around
The first endotherm, consistent in all DSC traces and exhibited in Figure 4, concludes at ∼130 °C and is attributed to dehydration of a H2O from the starting chemical species. Heat flow is relatively stable during which intermediate mass loss is observed. The occurrence of the exotherm about 375 °C is consistent with the third mass loss decomposition. Activation Energies of ER ≈ 160 kJ mol-1 (Figure 3) determined by the isoconversional method for the last and major step of mass loss in open-pan TGA experiments (Figure 1) are very similar to those determined in open-pan DSC experiments. The abrupt increase in ER for R < 0.1 to ER ≈ 225 kJ mol-1 and subsequent decrease to ER ≈ 160 kJ mol-1 coincides with the second mass loss step and may be due to the decomposition of a nonvolatile organic residue. Both TGA and DSC analysis performed over identical temperature ranges produce agreeable activation energies. 5. Conclusions In summary, the thermal decomposition kinetics of Ce-P have been studied by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC). Activation energies as a function of the extent of reaction, R, have been determined by model-free isoconversional analysis of these data. TGA analysis reveals an oxidative decomposition process yielding CeO2 as the final product with an activation energy of ∼160 kJ mol-1. This temperature range is significantly less than the metal oxide formation temperature of simple lanthanide salts and other hydrazine lanthanide complexes.1,24,25 DSC traces estimate an activation energy of this decomposition reaction of ∼135 kJ
TGA and DSC of N2H5[Ce(C4N2H2(COO)2)2(H2O)] mol-1. In open pans, Ce-P generates a heat release of ∼3750 J g-1 that is independent of both the heating rate, β, and the mass. This complex may be used as a precursor to fine particle cerium oxides due to its low temperature of decomposition. This method could also be used to prepare other metal oxides from a thermal decomposition process. Further studies to elucidate the full potential of the CeO2 are in the progress. Acknowledgment. The authors thank Mettler-Toledo, Inc. for donating the DSC instrument used in this study. Partial funding for this project was provided by the University of Utah. T. P. thanks the council of Scientific and Industrial Research, New Delhi, India, for the award of a Senior Research Fellowship. References and Notes (1) Govindarajan, S.; Patil, K. C.; Manohar, H.; Werner, P. E. J. Chem. Soc., Dalton Trans. 1986, 119. (2) Kumar, N. R. S.; Mathews, I. I.; Patil, K. C. Polyhedron 1991, 10, 579. (3) Edwards, E. D.; Keilly, J. F.; Mahon, M. F.; Molloy, K. C.; Thompsett, D. J. Chem. Soc., Dalton Trans. 1993, 3471. (4) Kuppusamy K.; Govindarajan, S. Thermochim. Acta 1996, 279, 143. (5) Sivasankar, B. N.; Govindarajan, S. Mater. Res. Bull. 1996, 31, 47. (6) Richard, P.; Tran Qui, D.; Bertaut, E. F. Acta Crystallogr., Sect. B 1974, 30, 628. Richard, P.; Tran Qui, D.; Bertaut, E. F. Acta Crystallogr.,
J. Phys. Chem. B, Vol. 109, No. 13, 2005 6129 Sect. B 1973, 29, 1111. (7) Ptasiewicz-Bak, H.; Leciejewicz, J. Pol. J. Chem. 1999, 73, 717. (8) Wenkin, M.; Devillers, M.; Tinant, B.; Declereq, J.-P. Inorg. Chim. Acta 1997, 258, 113. (9) Bezdenezhnykh, W. et al. Russ. J. Inorg. Chem. 1970, 15 (3), 324. (10) Trovarelli, A. Catal. ReV. Sci. Eng. 1996, 38, 439. (11) Dictor, V.; Roberts, S. J. Phys. Chem. 1989, 93, 5846. (12) Miki, T.; Ogawa, T.; Haneda, M.; Kakuta, N.; Ueno, A.; Tateishi, S.; Matsuura, S.; Sato, M. J. Phys. Chem. 1990, 94, 6464. (13) (a) Shido, T.; Iwasawa, Y. J Catal. 1992, 136, 493. (b) Shido, T.; Iwasawa, Y. J Catal. 1993, 141, 71. (14) Martinez-Arias, A.; Soria, J.; Conesa, J. C.; Seoane, X. L.; Arcoya, A.; Cataluna, R. J. Chem. Soc., Faraday Trans. 1995, 91, 1679. (15) Cordatos, H.; Ford, D.; Gorte, R. J. J. Phys. Chem. 1996, 100, 18128. (16) Martinez-Arias, A.; Soria, J.; Conesa, J. C. J. Catal. 1997, 168, 364. (17) Premkumar, T.; Govindarajan, S. Inorg. Chem. Commun. 2003, 6, 1385. (18) Vogel, A. I. A Text Book of QuantitatiVe Inorganic Analysis, 4th ed.; Longman: London, 1986. (19) Galwey, A. K.; Brown, M. E. Thermal Decomposition of Ionic Solids; Elsevier: Amsterdam, 1999. (20) Vyazovkin, S. J. Comput. Chem. 1997, 18, 393. (21) Long G. T.; Wight, C. A. J. Phys. Chem. B 2002, 106, 2791. (22) Long, G. T.; Vyazovkin, S.; Brems, B. A.; Wight, C. A. J. Phys. Chem. B 2000, 104, 2570. (23) Sell, T.; Vyazovkin, S.; Wight, C. A. Combust. Flame 1999, 119, 174. (24) Bukovec, N. Thermochim. Acta 1985, 88, 391. (25) Bukovec, N.; Milic´ev, S. Inorg. Chim. Acta 1987, 128, L25.
