Thermodynamics and kinetics of carbon dioxide chemisorption on

potential function.15 The absence of zig-zag relationships suggests that the potential well around the bending structure is rather deep and the observ...
5 downloads 0 Views 519KB Size
4052

J. Phys. Chem. 1984,88, 4052-4055

46 42

r

t

t

i

/

a'

A_:I 30

,p'

+

by calculating the values of Z, I , - I , = A, for different ringpuckering angles. In calculating the moments of inertia for methylenecyclohexane, we used the same structural parameters as those previously mentioned except that the dihedral angle between the C3C4Csand the c2c3c5c6 planes was assumed to be 50.0°, and the angle (e) between the cIcZc6 and czc3c5C6 planes was varied. Calculations for cyclohexanone were carried out in the same fashion except y(C=O) = 1.21 A. The results of these calculations are plotted in Figure 3 as a variation of Ac vs. 0 where the solid circles and triangles represent the calculated data for methylenecyclohexane and cyclohexanone, respectively. If the shape of the potential functions governing the ring-bending vibration for methylenecyclohexane and cyclohexanone are assumed to be similar around the well, the experimental values of Ac may yield indications of the relative flattening of chair conformations and of relative barriers to the ring-bending motion. Assuming the same angle of puckering for the six-membered rings of methylenecyclohexane and cyclohexanone, the value of Ac is calculated to be larger for the former than for the latter molecule. For instance, a puckering angle of 50" gives the values of 38.10 and 36.78 wA2 for methylenecyclohexane and cyclohexanone, respectively. From the experimental values of Ac for methyleit may necyclohexane (38.3 wA2) and cyclohexanone (34.3 dZ), be suggested that the extent of ring puckering and the barrier to ring inversion are higher in the former than in the latter molecule. This result is in accord with those determined from N M R spectra in s 0 1 u t i o n ~ ~and J ~ is attributed mainly to differences in the torsional barriers about the Cl-C2 and C& bonds as reflected by the experimental barriers to internal rotation in acetone (0.78 kcal/mol)'* and isobutene (2.2 kcal/m01).'~

Acknowledgment. I thank the Department of Chemistry, Memphis State University, for research support. Registry No. Methylenecyclohexane, 1192-37-6. (16) J. T. Gerig and R. A. Rimerman, J . Am. Chem. SOC.,92, 1219 (1 ' 970). (l?) F.A. L. Anet, G. N. Chmurny, and J. Krane, J . Am. Chem. SOC., 95,4423 (1973). (18) R. Nelson and L. Pierce, J . Mol. Spectrosc., 18, 344 (1965). (19) V. W.Laurie, J . Chem. Phys., 34, 1516 (1961).

Thermodynamics and Kinetics of Carbon Dioxide Chemisorption on Calcium Oxide Dario Beruto,t*Rodolfo Botter,*and Alan W. Searcy*t Materials and Molecular Research Division, Lawrence Berkeley Laboratory, and Materials Science and Mineral Engineering, University of California, Berkeley, California 94720, and Istituto Chimica Facolta di Ingegneria, Universita di Genova, Genova, Italy (Received: August 31, 1983: In Final Form: March 13, 1984)

Equilibrium chemisorption of COz on CaO was measured at 923-1013 K, and the kinetics of rapid adsorption and of slow dissolution of C 0 2 in the CaO, perhaps as CO,,- at grain boundaries, were measured in 1333 Pa of COz at 983-1033 K. Surface areas were measured by the BET method. From the assumptions that only (100) surfaces of the CaO were exposed, it was calculated that -55-99% of the surface 02-ions reacted with C 0 2 to form C032-ions. The enthalpy of adsorption was -199 & 8 kJ/mol and the thermal entropy of adsorption was -153 f 8 J/(mol deg). Both values were somewhat less negative at S O % coverage. The apparent activation enthalpy for adsorption was 61 i 10 kJ/mol. The apparent activation enthalpy of the slow dissolution process was 303 f 15 kJ/mol.

Studies of the thermodynamics and kinetics of chemisorption of gases on metal surfaces have long been, and remain, an active area of investigation.' In contrast, very few quantitative studies have been made of chemisorption on oxide surfaces. We became

interested in the chemisorption of CO, on CaO because of observations that CO, catalyzes a rapid reduction in CaO surrface area at temperatures as low as 900 K and pressures as low as 0.2 (1) See, for example, J. Benard, Ed., "Adsorption on Metal Surfaces: An Integrated Approach", Elsevier, Amsterdam, 1983.

+University of California. "niversita di Genova.

0022-3654/84/2088-4052$01.50/0

0 1984 American Chemical Society

CO, Chemisorption on C a O Busca and Lorenzelli4 have recently reviewed infrared studies of carbon oxides on metal oxide surfaces. These studies are in general agreement that at the temperatures of our interest-800 K and above-the C 0 2 on oxide surfaces is chemisorbed as C0:ions. Fukuda and TanabeS conclude that C03,- ions may be present on CaO surfaces in two different coordinations. On Scz03, COz is reported6 to be only weakly chemisorbed. (The measured heat of adsorption is -30 kJ/mol.) The kinetics of desorption of COz from silica surfaces have been studied,’ but the reported activation energies are a function of the heating programs used in the study. We know of no measurements of either the thermodynamics or the kinetics of CO, chemisorption on CaO. We report such measurements here, and we also report measurements of the rate of slow dissolution of COz in CaO.

Experimental Section Sample Preparation. Single crystals of high-purity calcites were cut along the natural cleavage planes into slices with average dimensions of 3.8 mm X 4.5 mm X 0.8 mm. Four such pieces were placed in a Pt crucible. It was hung from a symmetrical microbalance in a quartz resistance furnace? A chromel-alumel thermocouple tip was placed 5 mm from the bottom of the crucible. Calibration with another thermocouple introduced from the top of the furnace showed the temperature to remain constant within k 1 OC over a length of 5 cm. Double-wound resistance wires were designed to minimize the temperature gradient along the furnace cross section. Pressure was controlled by a single-sided absolute sensor head plus a control system placed at the top of the thermobalance. Vacuum decompositions of the calcite crystals to CaO were carried out at 959 K, while the recorded pressure at the ion gauge was 5 X lo-’ Pa. Two sets of calcite crystals were used for these experiments. The total initial weights were 160.8 and 156.5 mg. The specific surface area of the porous CaO obtained was in both cases 118 f 2 m2/(g of CaO). These high surface area samples of CaO are stable under vacuum, but sinter when exposed to C 0 2 at 900 K or highere2V3 The rate decreases as the surface area is reduced and becomes slow enough when the area is reduced below 20 m2/(g of CaO) to permit adsorption isotherm measurements. Accordingly, one sample was sintered at 959 K in CO, at 1333-Pa pressure for 100 h. After the CO, was desorbed at the same temperature under vacuum, BET measurements10,” with N2 at 78 K showed the sample specific surface area to be 19 m2/(g of CaO). The surface area of this sample decreased over the total time of the adsorption measurements to 15 m2/g. For the other sample, the C 0 2 treatment was carried out at 959 K and 1333 Pa for 50 h, and then at 973 K and 2000 Pa for an additional 50 h. This yielded CaO of 12.5 m2/g specific surface area. This area did not change significantly during adsorption studies. Isotherm Measurements. Tests showed that at 913 K the measured C 0 2 pressure at which calcite formation started was 5% below the reported equilibrium value.I2 Isothermal adsorption equilibria were studied at 923-1013 K and in the pressure range (2) J. Ewing, D. Beruto, and A. W. Searcy, J . Am. Ceram. SOC.,62, 580 (1979). (3) A. W. Searcy, M. Kim, and D. Beruto, “ P r d i n g s of the Symposium on High Temperature Materials Chemistry-I”, Z. A. Munir and D. Cubicciotti, Eds., The Electrochemical Society, Pennington, NJ, 1983, p 133. (4) G. Busca and V. Lorenzelli, Mater. Chem., 7 , 89 (1982). (5) Y . Fukuda and K. Tanabe, Bull, Chem. SOC.Jpn, 46, 1616 (1973). (6) J. E. Gonzalez De Prado, L. Gonzalez Tejuca, J. A. Pajares, and J. A. Soria, An. Quim.,69, 1239 (1973). (7) J. F. Antonini and G. Hochstrasser, Surf. Sci., 32, 665 (1972). (8) D. Beruto and A. W. Searcy, J . Chem. SOC.,Faraday Trans. 1 , 70, 2195 (1976). (9) D. Beruto, L. Barco, G. Belleri, and A. W. Searcy, J. Am. Ceram. Soc., 64, 76 (1981). (10) J. M. Thomas and W. J. Thomas, ‘Introduction to the Principles of Heterogeneous Catalysis”, Academic Press, London, 1967, Chapter 2. (1 1) S . Brunauer, P. H. Emmett, and E. Teller, J . Am. Chem. SOC.,60, 309 (1938). (12) K. H. Stern and E. L. Weise, “High Temperature Properties and Decomposition of Inorganic Salts, Part 2. Carbonates”, National Bureau of Standards, Washington, DC, 1969, Natl. Stand. Ret. Data Ser. (US.Natl. Bur. Stand.) No. 30.

The Journal of Physical Chemistry, Vol. 88, No. 18, 1984 4053

14

c

I

I

I

I

I

I

1

I

3

P

Lo

0 x

E

9

f

I

I

I

I

loo0

0

I

I

2000 P

COY

I

3000

4000

(Pal

Figure 1. CO, desorution isotherms obtained with CaO of 19 mz/a initial surface area and 15m2/g final area. nco2 is the moles of C 0 2 taken up by the sample and grcaOis the grams of CaO. ‘I

I

I



I

I

I

I

I

3’

x

8 9 Y

,s“

2‘ 973 3‘ 993 4‘ 1013

I

0

I

I

I

I

2000

4000

6000

1

Pco, (PO)

Figure 2. COz adsorption isotherms obtained with CaO of 12.5 mz/g surface area.

from 65 Pa to 70-80% of the reported equilibrium calcite decomposition pressure. For each temperature, weight variations were recorded continuously by using the microbalance with a sensitivity of 2 x g/cm. As soon as a sample reached the desired temperature, CO, was admitted. Constant pressure was established without observable delay. Runs with CO,, but without samples, were made to calibrate the system. The sample volume was too small to affect the pressure significantly. Blank runs with N2 gave buoyancy corrections of g when the pressure was 65 Pa, and less than 2 X g when the presence was greater than 150 Pa. To obtain an equlibrium isotherm, the CaO was heated under vacuum and held in CO, under the maximum planned CO, pressure for the isotherm until the thermogravimetric trace was constant to within its sensitivity. Weight changes were then less g/min. The pressure was decreased in steps, and than 1.5 X the points where the T G trace showed the weight changes to be no more than g/min were recorded. After each isotherm, the sample was pumped under vacuum until the initial sample weight was recorded. To test for reversibility of the isotherms, in two runs pressures were then increased by the same step increments. As shown in Figure 3, the weights reached when the changes became too slow for practical measurement were reproducible in desorption and adsorption. Kinetic Measurements. Adsorption rates were measured on the samples of 12.5 m2/g surface area in the temperature range 983-1033 K, and at a constant CO, pressure of 1333 Pa. The measured rates depended strongly upon the starting surface state. Reproducible runs were obtained by introducing initially 7 X g of C 0 2 , a quantity equivalent to 0.09% of the weight of C 0 2 required to transform all the CaO to CaCO,. At each temper-

4054

The Journal of Physical Chemistry, Vol. 88, No. 18, 1984 I

I

I

I

Beruto et al. TABLE I: Adsorption Thermodynamic Data 103(mol of co2/ isomol of -AHadsorpn: thermsa CaO) P kJ/mol Sample 1 (at 923-1000 K)

T=923 K

3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4

41

0.56-0.70e 0.60-0.75 0.64-0.80 0.67-0.84 0.71-0.89 0.74-0.93 0.78-0.98 0.82-1.00

199 197 197 199 198 186 186 16@

-Macisorpn,d

J/(mol deg) 156.2-1 5 1. l e 155.2-149.5 157.0-150.3 160.6-152.6 153.5-143.6 142.6-129.5 140.9-118.7 127.9-45.0

Sample 2 (at 953-1013 K)

I

I

0

200

I 400 600 pc0, (Pa) I

I

000

1000

Figure 3. Left axis, moles of COz, nco2, taken up per gram of CaO as a function of COz pressure in pascals. The triangles show values measured in two runs with decreasing pressure; the boxes are values measured with increasing pressure. ature, the weight was recorded from the initial time when CO, was introduced to the time when the weight change was less than g/min. Weight increases were followed with the thermog/cm. The initial chart speed was balance sensitivity at 2 X 30 cm/min and the final speed was 2 cm/h.

6 6 6 5

0.55 0.69 0.83 0.99

2.0 2.5 3.0 3.7

199 199 200 186

150.2 150.6 160.9 113.1

aNumber of isotherms used for enthalpy derivation. *Mole fraction of surface Oz- sites occupied by COj2-. CAdsorption enthalpy. dThermal entropy of adsorption. e Uncertainties reflect uncertainty in

surface area which changed from 19 m2/g initially to 15 m2/g finally. /Some data points above the deflection points of Figures 1 and 2 were used.

Results and Discussion When CO, was first introduced, 4-6 h was required before the rate of weight increases became too low to measure. But when pressures were then decreased in steps (Figures 1 and 2), apparent equilibration at each step was rapid. Periods of the order of 4-h heating at the temperature of the experiment under vacuum were required to remove about the final 20% of the COz that had been taken up by the sample. When pressures of CO, were reintroduced in steps to obtain adsorption isotherms, the initial approach to equilibrium was again slow, but once the CO, concentration in the sample reached the level characteristic of the desorption isotherm the adsorption curve followed the desorption curve to within a 2% or 3% scatter in the data (Figure 3). The fact that essentially identical isotherms were thus measured with decreasing and increasing C 0 2 pressures (Figure 3) is evidence that equilibrium was achieved. The long terms required to reach nearly constant weights when CO, is first introduced or is pumped off are consequences of slow transfer of CO, into and out of the interior of the CaO. The weight of slowly absorbed C 0 2 was subtracted from the total weight of C 0 2before adsorption equlibria were calculated (Table 1). Infrared measurements indicate COz to chemisorb by reaction with surface 0,- ions to form C03,The concentration of surface 0’ ions was calculated from the surface areas measured on the assumption that the total area is by the BET accounted for by surfaces formed of (100) planes of the NaC1-type structure. Enough COz was chemisorbed a t the highest C 0 2 pressures to convert some 80-99% of the surface 0” ions to CO$. The enthalpies of adsorption then were calculated from variations in the logarithm of the pressure of C02with the reciprocal of the temperature at constant C0,2- surface concentrations. These enthalpies (Table I) are constant at 199 f 2 kJ/mol for surface coverages of 55-80%. The change in surface area during measurements with the first sample introduces less than 2% uncertainty into the calculated enthalpies. From consideration of all probable errors, the uncertainty is set at 8 kJ. Lower enthalpies are calculated when the high coverage points that showed upward curvature in Figures 1 and 2 are included. The lower enthalpies may reflect repulsive interactions between surface C032- ions. At the average temperature of the equilibrium adsorption measurements, the standard enthalpy of reaction of COz and CaO

0”0.4 T (K) Q

1033

I

01 0

20

IO

0

30

501

40

9e3 I

I

1000

I

I

2ooo

I

I

3000

I

I

4000

I

J

5000

TI ME (sed

Figure 4. Ratios of moles of COz, nco2, taken up by CaO to the maximum COz taken up, nco:, as function of time when Pco2 = 1333 Pa. to form calcite is -1 8 1 kJ/mol.12 Thus, the enthalpy of adsorption of CO, on CaO is, like the enthalpy of adsorption of sulfur on a transition-metal surface,’, more negative than the enthalpy to form the bulk solid. As pointed out by Kemball,I4 strong chemisorption can be expected to imply adsorption on fixed sites and low entropies of the adsorbed species in comparison with entropies of more weakly bound species. By analogy with observations for bulk and solid solution reaction^,'^ the thermal (Le., nonconfigurational) entropy of adsorption of COz as C032-on CaO could be expected to be nearly the same as the entropy of the reaction of C 0 2 with CaO to form calcite. Configurational entropies S, were calculated from the expression s,: -R In (X/(l - X)) where X i s the mole fraction of surface oxide sites converted to C032-ions (Table I). Subtraction of S, from the total entropy of C 0 2 adsorption gave the thermal entropies reported in the last column of Table I. The values found near X = 0.5-0.7, for which errors in the calculated configurational entropy should be relatively small, gave for the thermal entropy of adsorption -153 f 8 J/(mol deg), compared to -150 J/(mol deg) for calcite formation.I2 (13) J. Benard, J. Oudar, N. Barbouth, E. Margot, and Y.Berthier, Surf. Sci., 88,L35 (1979). (14) C. Kemball in “The Heterogeneity of Catalyst Surfaces for Chemisorption”, H. S. Taylor, Ed., Academic Press, New York, 1948, p 233. (15) A. W.Searcy in “Progress in Inorganic Chemistry”, Vol. 3, F. A. Cotton, Ed., Interscience, New York, 1960, p 49.

J. Phys. Chem. 1984, 88, 4055-4058 Because C 0 2 must chemisorb on fixed surface sites, the adsorption equilibria would be expected to fit Langmuir adsorption isotherms in the range of coverage for which a constant enthalpy of adsorption is measured.16 In fact, reasonable fits to Langmuir isotherms are found up to the inflection points of Figures 1 and 2. Figure 4 shows, in the insert, the very rapid uptake of C 0 2 by CaO when C a O is first exposed to C 0 2 at a constant pressure at various temperatures. The much slower later pickup of COz is shown by the other part of the same figure. The rates of adsorption were calculated from the expression Jco, = kPco,, where Jco, is the moles of CO, reacted per second and k is the rate constant for the initial, essentially linear portions of each plot. In the linear range, back-reaction must be negligible, and the rate of the forward reaction must be little affected by the fraction of (16) I. Langmuir, J . Am. Chem. Soc., 38, 2221 (1916).

4055

surface sites already occupied by C032- ions. From the temperature dependence of the adsorption rates, the apparent enthalpy of activation for the adsorption reaction was calculated to be 61 f 10 kJ/mol. This activation enthalpy lies within the range of values found for C 0 2 adsorption on ScZO3, from 35 to 105 kJ, depending on treatment of the Sc203adsorption data.6 From the temperature dependence of the slow absorption of C 0 2 into the CaO, an apparent activation enthalpy of 303 f 15 kJ/mol is calculated for the unidentified rate-limiting step of this slow process. Acknowledgment. U. Auselmi Tamburini kindly provided computer calculations of data in the early part of this work. This work was supported by the Director, Office of Energy Research, Office of Basic Energy Sciences, Materials Sciences Division of the U S . Department of Energy, under contract no. DE-AC0376SF00098. Registry No. C 0 2 , 124-38-9; CaO, 1305-78-8.

Polymorphism and Liquid-Crystalline Behavior of Lithium n -Hexadecanoate Vincenzo Busico,* Angelo Ferraro, and Michele Vacatello Zstituto Chimico dell’Uniuersitd, 80134 Naples, Italy (Received: September 30, 1983; In Final Form: February 3, 1984)

The phase behavior of anhydrous lithium n-hexadecanoate has been carefully reinvestigated by differential scanning calorimetry (DSC) and X-ray diffraction, and the results have been compared with the largely conflicting literature data. Melting of the room-temperature LC1 (LC = lamellar crystalline) phase is achieved through the stepwise mechanism LC1-LC2-BR (BR = bidimensional rectangular). In the LC2 phase chain packing resembles that in the “rotator” phase of the n-paraffins. In the BR phase the ionic groups are organized in ribbonlike double sheets arranged in a rectangular bidimensional superlattice, while the conformationally disordered alkyl groups fill up the rest of the space. The melt is a smectic mesophase of the kind found for the n-alkylammonium chlorides.

Introduction n-Alkanes crystallize in layers of all-trans chains with the methyl ends in the basal planes.’ Chain-end segregation, which allows complete crystallization of the (CH,), sequences, is lost upon melting.2 Amphiphilic compounds CH3(CH2),X (with X an ionic end group) may crystallize with similar layered structures.M Relative to the n-alkanes, extra stabilization to the layers is provided by Coulomb interactions between the charged chain termini, particularly effective in the bilayer arrangement6 (pairs of layers facing each other with the ionic planes). Should such interactions compensate a reduced entropic gain on melting, double sheets of ionic chain ends could be maintained in the molten state and smectic mesophases result. Actually, smectic liquid crystals of potas~ium,~ rubidium,10and cesium n-alkanoates” have long been known (and comprehensively denoted as “neatn phases). Nevertheless, until (1) A. I. Kitaigorodsky, “Molecular Crystals and Molecules”, Academic Press, New York, 1973, pp 48-62. (2) M. Vacatello, G. Avitabile, P. Corradini, and A. Tuzi, J. Chem. Phys., 73, 548 (1980). (3) B. Gallot and A. Skoulios, Kolloid Z . Z . Polym. 209, 164 (1966). (4) A. D. Skoulios and V. Luzzati, Acta Crystallogr., 14, 278 (1961). (5) “Biological Membranes”, D. Chapman, Ed., Academic Press, London, 1968. (6) V. Busico, P. Cernicchiaro, P. Corradini, and M. Vacatello, J . Phys. Chem., 87, 1631 (1983). (7) A. Skoulios and V. Luzzati, Nature (London), 183, 1310 (1959). (8) B. Gallot and A. E. Skoulios, Acta Crystallogr., 15, 826 (1962). (9) B. Gallot and A. Skoulios, Kolloid Z . Z . Polym., 210, 143 (1966). (10) M. Sanesi, P. Ferloni, and P. Franzosini, Z . Naturforsch., A , 32A, 1173 11977). (1 i)M. Sanesi, P. Ferloni, M. Zangen, and P. Franzosini, 2. Naturforsch., A , 32A, 627 (1977).

0022-3654/84/2088-4055$01.50/0

recently, no other such mesophases of similar compounds had been described12 nor had there been any attempt to rationalize the mesomorphic behavior of anhydrous amphiphilic compounds. In previous p a p e r ~ , we ~ J ~presented results relative to a series of primary n-alkylammonium chlorides, n-CnHz,+,NH3Cl. Stepwise melting was pointed out for these compounds from the crystalline bilayer structure through a plastic phase to two increasingly disordered smectic mesophases (smectic I and smectic 11). Smectic 11, in particular, is isotropic in visible light, due to the reduced dimensions of the smectic domains, and stable up to thermal decomposition.6 Comparison with literature data on the alkali-metal n-alkanoates and observations on other compounds from our laboratory suggested a common mesomorphic behavior for a wide class of largely unknown smectogens. This prompted us to reinvestigate some systems for which controversial results are reported in the literature. Lithium nhexadecanoate (palmitate) was our first choice. In ref 14, a brief review concerning lithium n-alkanoates is found, in which largely conflicting data from different laboratories are presented. Lithium n-hexadecanoate, in particular, has been the subject of detailed structural and/or calorimetric investigations. Most authors agree in detecting three main phase transitions on heating the freshly prepared compounds from room temperature in the ranges 374-381 K,460-488 K , a n d 497-502 K. (12) H. Helker and R. Hatz, “Handbook of Liquid Crystals”, Verlag Chemie, Weinheim, 1980. (1 3) V. Busico, A. Scopa, and M. Vacatello, Z . Naturforsch., A, 37A, 1466 (1982). (14) P. Franzosini, M. Sanesi, A. Cingolani, and P. Ferloni, Z . Nufurforsch., A , 35A, 98 (1980).

0 1984 American Chemical Society