Langmuir 2005, 21, 3933-3939
3933
The Adsorption of C4 Unsaturated Hydrocarbons on Highly Dehydrated Silica. An IR-Spectroscopic and Thermodynamic Study Giuliana Magnacca* and Claudio Morterra Department of Chemistry IFM, University of Turin, and Consortium INSTM, Research Unit of Turin University, Via P. Giuria 7, 10125 Torino, Italy Received October 25, 2004. In Final Form: February 1, 2005 The adsorptive interaction of 1-butyne and 1-butene with a highly dehydrated pyrogenic silica system has been studied to understand the thermodynamic behavior of the adsorption process by the application of the Langmuir model and of the Van’t Hoff equation. In situ FTIR spectroscopy allowed the characterization of the adsorption phenomenon in terms of involved surface sites, gas-volumetric determinations yielded quantitative information relative to the adsorption process, and microcalorimetric results allowed the comparison between calculated and experimental data. Keq and ∆adsG° were obtained from Langmuir’s model application; ∆adsH data were obtained from the Van’t Hoff equation and by the isosteric heats method and were compared with experimental values. The virtual constancy of ∆adsH with equilibrium pressure and surface coverage (Langmuir model) allowed us to obtain the ∆adsH° values and, consequently, the ∆adsS° values for the systems of interest.
Introduction It is known that adsorption phenomena are hard to interpret in terms of thermodynamic behavior. One of the few ways through which it is possible to obtain thermodynamic information on an adsorption process is to examine it in terms of a simple mechanism. The most convenient of such mechanisms is that described by Langmuir’s theory, but very seldom do the experimental data obtained with real systems fit in a sufficiently wide range of experimental conditions that the oversimplified Langmuir model actually predicts. In view of its convenience, Langmuir’s model has been applied in various fields. Most often the studies concern the kinetics of adsorption of complex molecules (sucrose, amino acids) or metal ions on solids from a liquid phase (e.g., see ref 1), because the average homogeneity of the solid/liquid interface guarantees a better applicability of the model; sometimes reasonable results have been obtained also with gas-solid adsorption (e.g., see ref 2). The Langmuir model for nondissociative adsorption3 concerns the one-to-one interaction between a gas molecule (A) and a surface adsorbing site (M) to form the surface complex (A-M), according to the equilibrium reported in the following scheme: ka
Agas + Msurf y\ z [A-M]surf k d
where ka is the rate constant for adsorption and kd is that for desorption. Other conditions imposed by the Langmuir model are: * Corresponding author. Phone: +39 011 670 7543. Fax: +39 011 670 7855. E-mail:
[email protected]. (1) Singh, K.; Mohan, S. Appl. Surf. Sci. 2004, 221, 308. Titus, E.; Kalkar, A. K.; Gaikar, V. G. Colloids Surf., A 2003, 223, 55. Sekar, M.; Sakthi, V.; Rengaraj, S. J. Colloid Interface Sci. 2004, 279, 307. (2) Pakseresht, S.; Kazemeini, M.; Akbarnejad, M. M. Sep. Purif. Technol. 2002, 28, 53. Choudhary, V. R.; Mayadevi, S. Zeolites 1996, 17, 501. (3) Atkins, P. W. Physical Chemistry, V ed.; Oxford University Press: Oxford, 1994; p 987.
(i) the adsorption process is reversible and cannot proceed beyond the monolayer coverage; (ii) the surface is uniform and all sites are equivalent (i.e., there is no intrinsic heterogeneity); and (iii) adsorbed molecules do not interact with one another (i.e., there is no induced heterogeneity). The first form of the Langmuir isotherm (i.e., the form first proposed by Langmuir) is:
θ)
bp 1 + bp
(1)
where p is the equilibrium gas pressure, θ is the surface coverage, and b (the so-called Langmuir constant) is b ) ka/kd ) Keq, that is, the thermodynamic equilibrium constant. On rearranging the first θ versus p equation, the second form of the Langmuir isotherm is obtained:
θ ) Keqp 1-θ
(2)
which is the form currently used to confirm the applicability of Langmuir’s model to the experimental data. Although the conditions imposed by the starting Langmuir hypotheses are quite strict and not very realistic on a physicochemical ground, some adsorbent/adsorbate systems have been reported that do agree with the Langmuir equations. In particular, as was recently demonstrated by Garrone et al.,4,5 the adsorption of several alkenes and alkynes on highly dehydrated nonporous silica is actually describable in terms of Langmuir’s equation because: (i) uptake is largely limited to a monolayer (the interaction implies a specific and reversible H-bonding to surface OH groups); (ii) all adsorption sites are of the same type (only isolated surface silanols are present); and (4) Garrone, E.; Barbaglia, A.; Onida, B.; Civalleri, B.; Ugliengo, P. Phys. Chem. Chem. Phys. 1999, 1, 4649. (5) Onida, B.; Allian, M.; Borello, E.; Ugliengo, P.; Garrone, E. Langmuir 1997, 13, 5107.
10.1021/la0473761 CCC: $30.25 © 2005 American Chemical Society Published on Web 03/30/2005
3934
Langmuir, Vol. 21, No. 9, 2005
Magnacca and Morterra
(iii) lateral interactions between admolecules are virtually absent (adsorption sites are well separated from one another). The accurate IR data by Garrone et al.4,5 were used to obtain: (i) Keq of the adsorption processes and, from that, the ∆G° figures; (ii) other thermodynamic functions by resorting to two empirical correlations between standard molar adsorption enthalpy (∆adsH°) and the spectral shift (∆νOH) undergone by surface silanols upon H-bonding.6,7 The present contribution, dealing with the adsorption on a highly dehydrated nonporous silica of 1-butyne (an emblematic alkyne species) and 1-butene (an emblematic alkene species), is aimed at obtaining the isostheric heats of adsorption (qst) by carrying out 1-butyne and 1-butene adsorption experiments at different temperatures and comparing the latter values with experimental molar adsorption enthalpies obtained by combined microcalorimetric and gas/volumetric measurements. The comparison of isosteric heats with experimentally obtained ∆adsH values will confirm the Langmuir model as adequate in dealing with the system of interest, but will also evidence the heavy limits of empirical relations of the type applied by other authors.6,7 It is recalled that, for systems behaving according to Langmuir’s model, the noncalorimetric experimental determination of isosteric heats is quite easy. In fact, when θ is constant (isosteric conditions), by applying the Van’t Hoff equation, eq 2 leads to:
d(ln Keq) ) [d(-ln p)]θ
(3)
and, from this, we obtain:
-
( ( )) ( ( )) ∂ ln K ∂ ln p ) 1 1 ∂ ∂ T T
)
∆adsH qst )R R
Figure 1. FTIR spectra of pyrogenic silica A200: outgassed in vacuo at 873 K for 1 h (curve a); outgassed in vacuo at 1073 K for 1 h (curve b); contacted, after step b, with 20 Torr of water for 30 min and then outgassed in vacuo at room temperature for 10 min (curve c); after step c, outgassed in vacuo at 873 K for 1h (curve d). Numbers and arrows indicate the halfbandwidths of curves a, b, and d.
adsorption enthalpy ∆adsh () -qst) reasonably constant with equilibrium pressure and surface coverage, and, therefore, it allows also the equivalence: ∆adsH ) ∆adsH°. The possibility of expressing, for the systems of interest, standard ∆adsH° figures leads then to the possible acquisition of standard ∆adsS°. Experimental Section
(4)
θ
By the use of adsorption isotherms run at different temperatures (chosen in a relatively limited range so that ∆adsH can be assumed as constant), the Van’t Hoff equation allows us to obtain the (average) value of ∆adsH, whereas a plot of [ln p]θ versus 1/T yields the isosteric heat of adsorption at any selected coverage (qθst). The independent acquisition of ∆adsG° (from Keq) and, when feasible, of ∆adsH° should allow the evaluation of ∆adsS°, an important thermodynamic function to be compared, for the adsorbing molecule(s) and the adsorbent system considered, with calculated ∆adsS° data4,5 and with tabulated standard molar data.8 Note that, dealing with adsorption processes, reaction enthalpies (∆adsH) are normally referred to in order to define the energetics of the process in specified conditions. Standard ∆adsH° figures are, on the contrary, very seldom referred to for several reasons, including the difficulty of defining, for the adsorbed species (i.e., the reaction products) a meaningful standard state. Yet, systems following Langmuir’s equation in a sufficiently wide range of equilibrium pressures and surface coverages represent a useful exception, in that the absence of both intrinsic and induced sites heterogeneity guarantees a molar (6) Hertl, W.; Hair, M. L. J. Phys. Chem. 1968, 72, 4676. (7) Curthoys, G.; Davydov, V. Ya.; Kiselev, A. V.; Kiselev, S. A.; Kuznetsov, B. V. J. Colloid Interface Sci. 1974, 48, 58. (8) CRC Handbook of Chemistry and Physics; West, R. C., Ed.; CRC Press: Boca Raton, FL, 1985. Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17. Stull, D. R.; Westrum, E. F., Jr.; Sinke, G. C. Gas-Phase Ion and Neutral Thermochemistry. The chemical thermodynamics of organic compounds; J. Wiley: New York, 1969. Harrison, A. G. Chemical ionization mass spectroscopy; CRC Press: Boca Raton, FL, 1983.
Instrumentation. IR spectra were recorded at 2 cm-1 resolution with a FTIR spectrophotometer (Bruker IFS88), equipped with MCT detector. A homemade metal variabletemperature cell, equipped with CaF2 windows, was employed to obtain spectra of adsorbed unsaturated hydrocarbons in a temperature range between 296 and 353 K. The cell can be connected to a conventional high-vacuum apparatus (residual pressure < 10-5 mbar) and allows in situ thermal treatments up to 873-923 K and gas adsorption/desorption runs at various temperatures ranging from 673 to 77 K. For comparison purposes, plain beam-temperature (nominally, ambient temperature) spectra were also obtained with a conventional all-quartz cell, equipped with KBr windows, that allows in situ thermal treatments at up to 1123 K. The elaboration of experimental spectra (e.g., background subtraction, peaks area integration, etc.) was performed with the standard software OPUS by Bruker. Microcalorimetric experiments were performed at 303 K with a heat-flow microcalorimeter of the Tian-Calvet type (by Setaram) equipped with a calibrated gas-volumetric apparatus, so that adsorbed amounts and evolved adsorption heats can be determined simultaneously.9 Materials and Adsorption Procedure. Self-supporting IR pellets (of ∼5 mg/cm2 thickness) were obtained by compacting under a pressure of ∼1 ton dosed amounts of a nonporous pyrogenic silica (Aerosil A200 by Degussa, Frankfurt A. M.; specific surface area ≈ 200 m2/g). As the highest vacuum activation temperature that can be safely reached with the metal variable-temperature cell may not be sufficient to lead to a highly dehydroxylated silica surface (i.e., a surface carrying few, well separated, and (mostly) single silanols), the following two-steps procedure was adopted. The silica samples were first outgassed in vacuo in a quartz cell at 1073-1123 K for 1-2 h, and then briefly exposed to the atmosphere and quickly transferred to the variable-temperature IR cell, where another vacuum activation at 873 K was carried out for (at least) an additional hour. Figure 1 (curve b) shows that the preliminary activation at 1073-1123 (9) Cerruti, M.; Magnacca, G.; Bolis, V.; Morterra, C. J. Mater. Chem. 2003, 13, 1279.
Adsorption of C4 Unsaturated Hydrocarbons K produces a well dehydroxylated sample, with a Si-OH band definitely weaker and thinner than in the case of a vacuum activation at 873 K (curve a of Figure 1). Due to the largely hydrophobic nature of pyrogenic silicas brought to temperatures higher than ∼950 K,10 the subsequent brief exposure to the atmosphere (the transfer takes less than 10 min) restores only to a very marginal extent the surface hydrated layer of the starting material (see curve c of Figure 1. Note that only a hydrothermal conditioning at high temperature in the presence of a hydrating agent could restore the surface OH layer). As a consequence, the temperature of 873 K, applied for the second vacuum activation run in the variable-temperature cell, turns out to be high enough to lead again to the typical spectrum of a well dehydroxylated silica, characterized (as in curve d of Figure 1) by a symmetrical sword-shaped profile of the νOH band at ∼3745 cm-1. In these conditions, the formation of weak H-bonds between isolated surface OH groups and adsorbing gas molecules can be easily observed in the IR spectrum. In the present study, variable-temperature adsorption experiments were carried out on the activated samples, starting from the highest temperature (353 K) to the lowest one (298 K). The two-steps activation treatment described above for IR samples was performed also on a portion of the A200 silica specimen for the gas-volumetric and microcalorimetric adsorption study. Also in this case, the fluffy silica powder was compacted under a pressure of ∼1 ton, both for similarity with the IR samples and to avoid losses during the vacuum activation. The molecules chosen for the present adsorption study were 1-butene and 1-butyne. Both adsorptives give reversible adsorption with the silica substrate. Different surface coverages were obtained by contacting at 303 K the activated adsorbing material with stepwise increasing pressures of the selected gas. Equilibrium pressures were read with a Hg manometer in the case of IR experiments, and with a Barocell pressure gauge (Edwards, operating in the 0-100 Torr pressure range) in the case of microcalorimetric/gas-volumetric experiments. Chemicals were supplied by Merck (Reagents for Analysis; purity: >99.5%).
Langmuir, Vol. 21, No. 9, 2005 3935
Figure 2. FTIR spectra in the spectral νOH region before (dotted line curve) and after admission of 1-butyne at beam temperature (solid line curves) onto A200 silica activated at 1073 K, briefly rehydrated at room temperature, and outgassed in vacuo at 873 K for 1 h. The bold arrows indicate the gradual disappearance of the free OH-groups of silica (sharp peak at ∼3745 cm-1) and the contemporary formation of a broad band due to H-bonded OH groups (the maximum is constantly located at ∼3570 cm-1).
Results and Discussion 1-Butyne Adsorption. Activated silica was contacted with butyne pressures up to ∼0.25 bar. The extent of adsorption, represented by the surface coverage θ, was determined considering the H-bonding formation between admolecules and OH groups present at the surface of the silica system. Figure 2 shows that the adsorption interaction causes the gradual disappearance of the 3745 cm-1 signal (the νOH vibration of free surface OH groups), and the parallel appearance of a weak and broad spectral component at lower wavenumbers (the νOH vibration of H-bonded OH groups; apparent maximum at ∼3570 cm-1; ∆νOH ≈ -175 cm-1. The constant shape and spectral position of this band is consistent with the Langmuir’s hypothesis of a constant energy of interaction). If the integrated absorbance Amax stands for the starting intensity of the νOH peak at 3745 cm-1, this figure is related to the maximum amount of free OH groups, that is, to the overall amount of surface sites available for adsorption. If Apeak represents the integrated absorbance of the 3745 cm-1 band still present after the adsorptive interaction, at any equilibrium pressure the difference (Amax - Apeak) is proportional to the number of sites interacting with the admolecules, and the surface coverage θ is thus represented by the (Amax - Apeak)/Amax ratio. The OH/butyne interaction was studied at seven different temperatures (from 353 to 296 K), by increasing at each adsorption temperature the butyne pressure up to ∼0.20-0.25 bar. (10) Legrand, A. P. The surface properties of silicas; J. Wiley & Sons: New York, 1999.
Figure 3. Isotherms of 1-butyne adsorbed on vacuum activated A200. Symbols represent the experimental data. Each curve corresponds to the adsorption temperature there indicated. Horizontal dotted-line traces correspond to the isosteric sections selected for isosteric heats calculation (the relevant θ figures are indicated in the figure).
The spectroscopic adsorption isotherms obtained for each temperature are reported in Figure 3. No desorption experimental data are reported in the figure, as the process is fully reversible. As expected of an exothermic process, the adsorption isotherms cover θ-intervals that become less and less extended with increasing adsorption temperature. The applicability of Langmuir’s model was then checked, to obtain indirect thermodynamic information on the adsorption process. Figure 4 reports the so-called Langmuir plots, in which spectroscopic θ/(1 - θ) figures are reported as a function of equilibrium pressure: the linear plots obtained allow us to conclude that the adsorption of 1-butyne on a highly dehydrated nonporous silica surface fits the Langmuir model, at least for each of the temperatures examined. The slope of the linear plots gives, for each adsorption temperature, the value of the thermodynamic equilibrium constant. Table 1 reports the (Keq)T figures obtained, and the corresponding (∆G°)T values, being ∆G° ) -RT ln Keq. Besides (∆G°)T values, also the molar adsorption enthalpies ∆adsh can be obtained. The Van’t Hoff equation yielded an average ∆adsh value of ∼ -28.5 kJ/mol. Figure 5 reports the values for ln p and 1/T, interpolated from the isotherms of Figure 3, at the selected constant
3936
Langmuir, Vol. 21, No. 9, 2005
Magnacca and Morterra
Figure 4. Langmuir plot applied to the experimental data of 1-butyne adsorption (square, circle, triangle symbols). The straight lines are linear fits of the data obtained at the adsorption temperatures indicated in the figure.
Figure 5. Plot of ln p versus 1/T for the adsorption of 1-butyne on activated A200 silica. The straight lines drawn in the figure are linear fits of the experimental data obtained for the coverage values indicated in the figure. Experimental data (solid square symbols) were obtained from Figure 3, by reading the (p-T) values along the isosteric dotted-lines there indicated. Table 1. Equilibrium Data Obtained, According to Langmuir’s Model, for the Adsorption of 1-Butyne on Activated A200 Silica T (K)
Keq
∆G° (kJ/mol)
296 303 313 323 333 343 353
9.50 6.20 4.37 3.31 2.48 1.92 1.25
-5.559 -4.596 -3.838 -3.214 -2.515 -1.86 -0.655
Table 2. Molar Adsorption Enthalpies in Isosteric Conditions for the Adsorption of 1-Butyne on Activated A200 Silica in the Temperature Range 296-353 K θ
∆adsh (kJ/mol)
0.10 0.20 0.30 0.39 0.40
-24.78 -28.41 -28.68 -29.51 -31.30
coverages θ ) 0.10, 0.20, 0.30, 0.39, and 0.40: all plots are indeed linear and all lines are almost parallel, indicating, on the basis of eq 4, that the isosteric heat of adsorption is actually virtually constant with coverage in the selected temperature interval, as expected on the basis of Langmuir’s theory. The values of ∆adsh are reported in Table 2, and the average molar adsorption enthalpy turns out to be -28.6 kJ mol-1, as was also obtained by the use of Van’t Hoff equation. In particular, it is noted that the values obtained for θ ) 0.20, 0.30, and 0.39 are affected by a very small dispersion, whereas for θ ) 0.40 the calculated ∆adsh figure is somewhat higher, but for this higher coverage only few
Figure 6. Molar adsorption heats versus coverage for 1-butyne adsorbed at 303 K on activated A200 silica.
ln p readings are available. As for the slightly lower isosteric heat obtained for the lowest coverage (θ ) 0.10), it is possible that ln p readings at low coverages are affected by a somewhat higher uncertainty. Certainly, at low coverages, a somewhat higher (and not lower) qst should be expected if a minor fraction of more energetic sites were actually active in the earliest stages of adsorption (moderate sites heterogeneity). To validate the applicability of Langmuir’s model to the 1-butyne/silica system and to confirm the accuracy of the ∆adsh figures obtained, differential molar adsorption heats were also obtained from a direct experimental determination. A silica sample was activated following the same procedure adopted for the spectroscopic measurements, and underwent the adsorption of increasing doses of 1-butyne at 303 K in a gas-volumetric and microcalorimetric apparatus. From the quantitative adsorption isotherm (amounts adsorbed per m2 vs equilibrium pressure) and the corresponding calorimetric adsorption isotherm (heats released per m2 vs equilibrium pressure), the differential molar energetic data reported in Figure 6 (heats released vs adsorbed amounts) were obtained by graphical differentiation. With the exception of the first 2-3 points of Figure 6, in which the qst value slowly declines from 38-40 to ∼33 kJ/mol (so confirming the existence of a moderate sites heterogeneity in the highly dehydrated silica system), the differential adsorption heat values remain fairly constant at a figure around 32-33 kJ/mol. This “direct” energetic datum is only slightly higher and thus in good agreement with the data calculated above with the “indirect” isosteric heats method. It also confirms the reasonable applicability of the Langmuir’s model to the 1-butyne/silica system. It is interesting to note that the average molar adsorption enthalpy obtained here by both calorimetric and noncalorimetric methods for the 1-butyne/silica system is comparable and, actually, some 15-30% higher than the standard molar enthalpy for condensation (-∆condH° ) 25 kJ mol-1). Unlike that, Onida et al.5 obtain, using an empirical correlation taken from the literature,6,7 a definitely lower figure (qst ≈ 16 kJ mol-1, i.e., a figure some 50% lower than the actual experimental qst). The comparability between qst and -∆condH° means that the formation in the adsorption process of a specific H-bonding interaction between isolated surface silanols and butyne triple bonds does actually compensate for the substitution of the whole-space liquid phase with the half-space-filling silica. Note that the difference between -qst (obtained through the indirect isosteric heats method) and the differential molar adsorption heat (obtained through the direct microcalorimetric method) is actually slightly larger than reported above, as a -RT term should be added to the
Adsorption of C4 Unsaturated Hydrocarbons
Langmuir, Vol. 21, No. 9, 2005 3937
Figure 7. Spectroscopic adsorption isotherm (left-hand curve; θ is expressed as Amax - Apeak) and the corresponding Langmuir plot (right-hand plot) obtained for the adsorption of 1-butyne on activated A200 silica, maintained at beam temperature by the IR radiation (i.e., without external heating).
latter, even if this aspect is normally neglected. In fact, the microcalorimetric apparatus operates at constant volume, and thus all (integral) heats obtained experimentally measure changes of internal energy ∆U rather than changes of enthalpy ∆H, and the difference between the two state function changes, for a mole of perfect gas passing in isothermal conditions from the gaseous phase to the condensed (adsorbed) phase, is given by the following relations:
∆H ) ∆U + ∆(nRT) ) ∆U + RT(∆n) ) ∆U - RT where ∆n ) - 1 is the number of moles passing, in the adsorption process stoichiometry, from the gaseous phase to the condensed (adsorbed) phase. Thus,
qst ) -∆h ) -∆u + RT For comparison purposes, we have also considered the IR spectroscopic description of 1-butyne adsorption on A200 silica at beam temperature (BT), that is, when the sample is not kept at a constant temperature in the variable-temperature cell but is heated to a constant temperature by the IR radiation in a conventional quartz cell. Data relative to this experiment are reported in Figure 7. The linear Langmuir plot in Figure 7 allows us to determine the thermodynamic equilibrium constant: (K)BT ) 3.12. This (K)BT value, intermediate between that obtained for the sample kept at 323 K and that at 333 K (see Table 1), indicates as between 323 and 333 K the temperature actually reached by the (white) silica sample in the IR beam. This is a temperature measurement of a “working system” not easily obtainable by other methods. The simultaneous knowledge of ∆adsG°, obtained through the experimental determination of the (K)T equilibrium constants, and of ∆adsH°, obtained through the isosteric heats method and/or through the direct microcalorimetric determination of molar adsorption heats, and in view of the mentioned Langmuir’s hypothesis of differential adsorption enthalpies constant with coverage, the equation ∆G° ) ∆H° - T∆S° leads us to the evaluation of ∆adsS°, that is, of the standard molar entropy loss caused by the adsorption process. If, for instance, we refer to the operative temperature of the calorimetric apparatus (T ) 303 K), when ∆G° ) -4.60 kJ mol-1 and ∆H° ranges between -33 kJ mol-1 (from microcalorimetric measurements) and -29 kJ mol-1 (from isosteric heats determination), a figure -∆S° ) 94-80 J mol-1 K-1 is obtained. In the limiting case of the very first adsorption steps, in which a small fraction of somewhat more energetic sites is likely to be involved and ∆adsh can be as high as -40
Figure 8. Isotherms of 1-butene adsorbed on highly dehydrated A200. Symbols correspond to the experimental data. The adsorption temperatures are indicated on the curves. Horizontal dotted-line traces correspond to the four isosteric sections selected for isosteric heats calculation (the relevant θ figures are indicated in the figure).
kJ mol-1, -∆adss can become as large as ∼117 J mol-1 K-1. These ∆adsS° figures are comparable or, in the limiting case, in moderate excess with respect to the standard molar entropy change for condensation, for 1-butyne (∆condS° ) 92 J mol-1 K-1, and should be expected to be ∼85 J mol-1 K-1 according to the well-known Trouton rule). It is thus concluded that, in the case of 1-butyne uptake on a highly dehydrated non porous silica, the formation of specific H-bonding interactions between silanols and the unsaturated hydrocarbon leads to a disorder/order transition comparable to that typical of the gas/liquidphase transition, despite the fact that the solid is filling only half of the available space so that more freedom should be allowed to adsorbed molecules than to condensed ones. 1-Butene Adsorption. The same procedure reported above for 1-butyne uptake was also applied to the study of 1-butene adsorption on the same silica system, and some of the results are here briefly outlined. All measurements are expected to be affected by a larger error because the interaction between surface OH groups and the doublebond C4 molecule should be less energetic than that observed for 1-butyne, as previously reported by Garrone et al.4 Reasonably accurate spectroscopic isotherms were obtained at 299, 303, 316, 323, and 343 K, as reported in Figure 8. Only a pressure range up to ∼0.1 bar could be analyzed in some detail and, as expected, relatively low coverages were reached. The θ/(1 - θ) versus p Langmuir plots obtained from the isotherms of Figure 8 are reported in Figure 9. The high scattering of the data obtained with 1-butene uptake in the temperature range explored does not allow
3938
Langmuir, Vol. 21, No. 9, 2005
Magnacca and Morterra
Figure 9. Langmuir plot applied to the experimental adsorption data of Figure 8 (square, circle, triangle symbols). The straight lines are linear fits of the experimental data (including the point θ ) 0 for p ) 0) obtained for the adsorption temperatures indicated in the figure.
Figure 10. Plot of ln p versus 1/T for the adsorption of 1-butene on activated A200 silica. The straight lines drawn in the figure are linear fits of the experimental data obtained for the coverages indicated in the figure. (Experimental data were obtained from Figure 8 by reading the (p-T) values along the isosteric dotted lines there indicated.) Table 3. Equilibrium Data Obtained, According to Langmuir’s Model, for the Adsorption of 1-Butene on Activated A200 Silica T (K)
Keq
∆G° (kJ/mol)
299 303 316 323 343
3.61 3.20 2.14 1.64 1.16
-3.20 -2.93 -2.00 -1.33 -0.42
it to be as positive as in the case of 1-butyne uptake on the applicability of Langmuir’s model. Still, assuming that the model is actually obeyed, the (Keq)T and ∆G° figures reported in Table 3 are obtained. Definitely lower thermodynamic equilibrium constants are observed, as expected of a weaker adsorption interaction, and this leads to (∆G°)T data still negative (as expected of a spontaneous process) but numerically indicative of a very slightly exergonic process. The average ∆adsh, calculated by application of the Van’t Hoff equation, turns out to be -25.5 kJ mol-1 and indicates a process only slightly less exothermic with respect to 1-butyne adsorption. Figure 10 reports the ln p versus 1/T plots for four isosteric conditions in the 0.05-0.12 coverage range, whereas the isosteric heats obtained from the slope of the graphs yield the data reported in Table 4. Despite the rather large data scattering already mentioned, a reasonable parallelism is observed for the four straight lines of Figure 10, yielding an average ∆adsh ≈ -26 kJ mol-1, that is, a standard molar adsorption enthalpy (obtained through the isosteric heats method)
Figure 11. Molar adsorption heats versus coverage for 1-butene adsorbed at 303 K on activated A200 silica. Table 4. Molar Adsorption Enthalpies in Isosteric Conditions for the Adsorption of 1-Butene on Activated A200 Silica in the Temperature Range 299-343 K θ
∆adsh (kJ/mol)
0.120 0.100 0.065 0.050
-28.27 -30.21 -25.19 -19.72
that is only some 10-12% lower than that obtained for the 1-butyne/silica system. Again, to check the accuracy of the noncalorimetric determination, a direct gas-volumetric/microcalorimetric measurement was carried out, and the differential molar interaction energies reported in Figure 11 were obtained. The molar adsorption enthalpies are fairly constant in the range 30-35 kJ/mol, with only a minor decrease in the first adsorption steps. This confirms that Langmuir’s hypothesis of adsorption energies largely independent of surface coverage is reasonably obeyed and indicates that 1-butene can reveal only to a minimal extent the existence of a fraction of more energetic sites acting in the earliest stages of the adsorption process. The average datum, ∼ -31 kJ mol-1, is in good agreement with that obtained with the isosteric heats method and confirms that the average interaction energy of 1-butene with surface silanols is (only) some 10% lower than that of 1-butyne. As for the standard molar entropy change, for Tads ) 303 K, ∆adsG° ) -2.93 kJ mol-1, and ∆adsH° ≈ -30 kJ/mol (averaging the data from calorimetric and noncalorimetric methods), we obtain ∆adsS° ≈ -89.3 J mol-1 K-1. This entropic term, again well consistent with Trouton’s rule and again slightly in excess over the standard entropy change for condensation, for 1-butyne (∆condS° ) 78.4 J mol-1 K-1), is only some 10-15% lower than that obtained. It is therefore deduced that the definitely less exergonic character of the silica/1-butene interaction (with respect to the silica/1-butyne interaction) is ascribable to a (moderate) difference of both enthalpic and entropic terms. This, in turn, means that the two interactions are only moderately different in terms both of interaction energy (as demonstrated by the different ∆νOH observed in the two cases: (∆νOH)1-butyne ≈ -174 cm-1, and (∆νOH)1-butene ≈ -153 cm-1) and of the extent of disorder-order achieved on passing from the gas to the adsorbed phase. Conclusions It has been confirmed that a highly dehydroxylated nonporous pyrogenic silica is a system for which the adsorption of light unsaturated hydrocarbons, yielding weak H-bonding interactions with isolated surface silanols, can be interpreted in terms of Langmuir’s model, despite the heavy conditions/approximations imposed by the model.
Adsorption of C4 Unsaturated Hydrocarbons
Even if the accuracy with which we could obtain adsorption data for the silica/1-butyne and silica/1-butene systems is rather different, reasonable and consistent thermodynamic information could be obtained for both systems of interest. The application of Langmuir’s model and equations, based on the use of the sole IR spectroscopic data, leads to the determination of important thermodynamic features of the adsorption process, such as equilibrium constants and standard free energy changes. Yet, the use of the sole IR spectroscopic data turns out to be inadequate for the determination of the standard enthalpy changes, as the empiric correlations elaborated in the past between experimental ∆νOH shifts and ∆adsH° are not sufficiently accurate nor selective, at least for the systems here investigated. Enthalpy changes connected to the adsorption process must be determined independently, either by the use of spectroscopic data obtained at different temperatures (and resorting to the van’t Hoff equation or, better, to the isosteric heats method), or by the use of direct micro-
Langmuir, Vol. 21, No. 9, 2005 3939
calorimetric experiments. ∆adsH° data obtained through the two methods are in good agreement and correlate nicely with condensation enthalpy changes, for which differences within some 15-20% exist between C4 alkenes and alkines. The simultaneous and independent knowledge of ∆adsH° and ∆adsG° leads to the determination of standard entropy changes connected to the adsorption process. ∆adsS° figures so calculated also nicely correlate to the molar condensation changes of entropy and present small differences between the two families of unsaturated hydrocarbons. This confirms that ∆adsS° (∆adss) is mainly ascribable to disorder/order transition in a process closely comparable to the gas/liquid-phase transition, even if the interaction dealt with here is a specific (chemisorptive) one and the adsorbing medium is a typically half-space-filling one. Acknowledgment. This research was partially financed with funds from the Consortium INSTM (Florence, Italy), project PRISMA2003. LA0473761
