A transmission infrared cell design for temperature-controlled adsorption and reactivity studies on heterogeneous catalysts

A heterogeneous catalyst accelerates a chemical reaction by facilitating the adsorption and reaction of molecules and intermediates on active sites at the fluid-solid interface. These active sites provide access to new reaction pathways that may otherwise be inaccessible via thermal or homogeneous routes. The rate at which a reaction occurs in heterogeneously catalyzed processes is influenced by several parameters including the number and types of active sites available to stabilize different molecules and intermediates. Therefore, an important challenge to address in fundamental studies of heterogeneous catalysts is to accurately identify and count the number of catalytic sites, required in order to rigorously normalize the number of reaction turnovers per site per unit time and quantify catalytic performance.¹ One method to distinguish and quantify the types of active sites available on a catalyst surface is titration with a probe molecule that shows characteristic vibrational bands within the infrared (IR) region. Basic titrants such as pyridine, NH₃, CD₃CN, CO, and C₃H₆ have been used widely together with IR spectroscopy to discriminate between Brønsted and Lewis acid sites, as well as surface hydroxyl groups, within solid acid catalysts such as zeolites and mixed metal oxides.^2–10 Acidic titrants such as CO₂, CHCl₃, and acetylene have been used analogously to probe base sites present in these surfaces.^2,8,11

IR spectra collected at sub-ambient temperatures are especially useful to characterize the surfaces of heterogeneous catalysts when surface structures or intermediates are not observable or stable at higher temperatures.^2,12,13 Under cryogenic temperatures (<150 K), CO coordinates with surface hydroxyl groups and Lewis acid centers via hydrogen-bonding and σ-bonding interactions that cause IR frequency shifts relative to gas-phase CO. At ambient temperatures (303 K), nitrile stretches can be distinguished of CD₃CN bound at partially hydrolyzed “open” ((SiO)₃M(OH); 2316 cm⁻¹) and fully coordinated closed ((SiO)₄M; 2308 cm⁻¹) Lewis acidic Sn⁴⁺ sites located within beta zeolite frameworks, but not of CD₃CN bound at these site coordinations for Ti⁴⁺ or Zr⁴⁺.^14,15 A recent study by Sushkevich et al.,¹⁴ however, has reported that CO can differentiate between open and closed framework Lewis acid sites within Zr-beta at cryogenic temperatures (∼100 K). Similarly, transmission IR spectra of Brønsted acidic zeolites, such as H-mordenite, after cryogenic CO adsorption (T ∼ 123 K) have been used previously to distinguish between Lewis acidic Al present as extra-framework Al³⁺ and penta-coordinated Al³⁺ sites.^16,17

Many designs have been proposed for transmission IR gas cells with various capabilities used to study surface chemistries that occur on heterogeneous catalysts. Examples of IR cell design features include operation under vacuum (P < 10⁻² Torr),^2,18–20 operation at sub-ambient temperatures,^2,18,19 operation at elevated temperatures (T > 770 K),¹⁸ temperature control and programmability,^18,21–23 rapid cooling rates (∼−0.6 K s⁻¹),^18,19 and the ability to perform in situ gas treatments.^{18,20,24–26} While several reported designs^18,19 incorporate many of the these features, they often introduce additional experimental difficulties, such as cumbersome sample preparation techniques¹⁸ and the use of flanges or sealants to secure IR-transparent windows onto the cell. The transmission IR gas cell described herein incorporates the aforementioned experimental capabilities along with improved design features to simplify cell operation over a wide experimental range. A key design feature is the modified, hand-tightened Cajon fittings to provide leak-tight operation and enable rapid removal and loading of catalyst samples with minimal IR cell disassembly. Additionally, standard catalyst powder samples are utilized as pressed, self-supporting wafers. This IR cell has been used with titrants such as pyridine and CD₃CN to quantify framework Lewis acidic sites within Sn-beta zeolites.⁶ A complete list of design characteristics and capabilities is shown in Table I.

Itemized lists of materials to construct the transmission IR gas cell and ancillary temperature controller can be found in Tables SI and SII, respectively, of the supplementary material. A schematic of the gas cell along the path of the IR beam and an assembled view of the cell within the heating-cooling chamber are shown in Fig. 1. The gas cell is a fused quartz tube (25.4 mm O.D. × 152.4 mm L), as shown by item ii in Fig. 1(a) and in Fig. S1 of the supplementary material. The gas cell has indentations located at fixed positions along the interior surface which function as back-stops to properly align the catalyst sample holder within the quartz tube. The catalyst sample holder is a smaller quartz tube (20 mm O.D. × 25.4 mm L) with a recessed slit at its mid-point to hold the catalyst wafer, and it is placed within the outer quartz tube as indicated by item v in Fig. 1(a). A detailed view of the catalyst sample holder is shown in Fig. S2. Modified 1 in. (25.4 mm). Cajon fittings that contain integrated thermocouples for temperature indication and control, along with inlet and outlet connections to enable gas flow through the cell, are used to seal both ends of the gas cell as indicated by item ix in Fig. 1. Additional details for the modified Cajon fittings are shown in Fig. S3. The outer nut on each 1 in. Cajon fitting is used to hold a standard 25.4 mm diameter × 10 mm thick IR-grade, CaF₂ window. A custom 316 stainless steel retainer ring (Fig. S4) is placed between the CaF₂ window and the Cajon fitting body in order to secure the window in place. Elastomeric O-rings within the Cajon fittings provide leak-tight seals around the CaF₂ windows, outer quartz tube, and thermocouples. These O-rings can be easily exchanged to ensure thermal and chemical compatibility with the various probe molecules of interest at relevant temperatures.

The assembled IR gas cell is placed inside of a temperature-controlled chamber as shown in Fig. 1(b). The modified Cajon fittings and approximately 6 mm (1/4 in.) of the 25.4 mm O.D. quartz tube are exposed on either side of the temperature-controlled chamber in order to provide ambient cooling and maintain the temperature of the elastomeric O-rings within their operating range. The insulated chamber is comprised of a brass block (Fig. 1(b), item xi) that is housed within an alumina silicate cover (Fig. 1(b), item x). Brass was chosen as the material of construction due to its superior thermal conductivity (k = 97–104 W m⁻¹ K⁻¹) relative to stainless steel (k = 16 W m⁻¹ K⁻¹) across the operating temperature range of interest (100–1000 K).²⁷ Heat is supplied to the IR cell by six cartridge heaters (Chromalox, 1/4 in. D × 2-1/2 in. L, 200 W) that are oriented perpendicular to the quartz tube inside of the brass block: three on the top section and three on the bottom section (Fig. 1(b), item xvii). Cooling is achieved via 4.8 mm diameter (3/16 in.) internal channels on the top and bottom sections of the brass block to direct the flow of liquid nitrogen (LN₂) through the cell chamber (Fig. 1(b), items xiv and xv) and out to the ventilation system. These cooling channels encompass the interior perimeters of the top and bottom segments of the brass block, independently. Comprehensive design drawings of the brass block, ceramic cover, and height-adjustable base can be found in Figs. S5–S8 of the supplementary material. A 28 mm (1.1 in.) diameter hole was bored through the brass block and ceramic cover to enable insertion and removal of the quartz IR gas cell without requiring disassembly of the temperature-controlled chamber (Fig. S7). Furthermore, the 0.25 mm (0.01 in.) tolerance between the quartz tube and the brass block maximizes the surface area for heat transfer. Images of the assembled and disassembled IR cell temperature-controlled chamber are shown in Fig. 2, with additional images shown in Figs. S9 and S10 of the supplementary material.

The cartridge heaters are connected to a process temperature control module (Eurotherm, 2404) that also operates a cryogenic solenoid valve (Asco, 2-position, normally closed) to control the flow of LN₂ to the temperature-controlled chamber, based on a temperature reading from inside of the quartz cell measured by a K-type thermocouple positioned within 1 cm of the front face of the catalyst wafer. A second thermocouple, also located inside the quartz cell, provides temperature indication at the back face of the catalyst wafer. These interior thermocouples (Figs. 1(a) and 1(b), item i) indicate the bulk fluid temperature near the catalyst surface, which is assumed to be in thermal equilibrium. Quantification of reaction rates or other kinetic parameters for highly exothermic or endothermic reactions during in situ gas treatments may be complicated by temperature gradients between the catalyst surface and interior thermocouple, and by concentration gradients caused by impaired diffusion of reactants to the wafer surface resulting from gas bypassing through the annular region between the sample wafer holder and interior wall of the quartz cell. For a specific reaction, the severity of potential temperature and concentration gradients external to the wafer can be estimated by using Mears’ criterion.²⁸ 

An independent thermocouple is inserted into the brass block and connected to a secondary temperature control module (Omega, CN9000) with an interlock to provide over-temperature protection (T = 1000 K) for the brass block and IR cell. The process and over-temperature control modules, high-temperature interlock, and heating-cooling switches and indication lights are housed within a temperature controller as shown in Figs. S11 and S12. A complete electrical wiring diagram for the temperature controller is shown in Fig. S13 of the supplementary material.

The IR cell chamber assembly is equipped with additional safety features that include cryogenic brass relief valves (P = 100 psig) both upstream and downstream of the LN₂ control valve to provide over-pressure protection for the brass block, solenoid valve, and process tubing (Fig. S14). The relief valves provide an outlet for trapped LN₂ in the unlikely events that (i) the cooling exhaust becomes obstructed by ice formed from condensation of ambient water vapor during cryogenic experiments, or (ii) the LN₂ cooling system is inadvertently isolated from the LN₂ tank, which is equipped with a pressure relief valve. Additionally, the cold N₂ exhaust from the cell chamber is vented through a closed vapor collection system that eliminates the potential for personnel exposure as well as oxygen displacement in the laboratory.

Approximately 35–75 mg of a powder catalyst sample is pressed into a self-supporting wafer (0.9–1 cm radius) and placed into the recessed slit of the quartz sample holder (Fig. 1(a), item v). Details regarding the preparation of the sample wafer can be found in Section 3 of the supplementary material. The quartz sample holder is carefully inserted into the outer quartz sample tube and positioned by the interior indentations within the outer tube as shown in Fig. 1(a). Next, the quartz tube and sample holder assembly are inserted through the 28 mm bore-through hole of the temperature-controlled chamber. The modified Cajon fittings with integral thermocouples and CaF₂ windows are fastened onto the ends of the outer quartz tube. Lastly, the cell chamber is covered with a layer of fiber-blanket refractory and wrapped in a high-temperature woven insulation tape in order to reduce thermal losses from the IR cell (Fig. S10).

The assembled transmission IR cell is placed inside of the IR spectrometer (Nicolet 4700) sample compartment and connected to a glass dosing system via a transfer line with $14$ in. Cajon unions as shown by the schematic in Fig. 3. Purified air (CO₂, H₂O free) is supplied continuously to the IR beam path in order to maintain a clean background. Exhaust vapors from the IR cell are vented through a laboratory fume hood. LN₂ delivery and exhaust lines are insulated with low-temperature foam insulation and connected to the brass block via 1/4 in. VCR fittings as shown in Fig. 2. Prior to the start of the experiment, the IR cell is leak tested under vacuum (rotary vane pump, Alcatel 2008A, ∼10⁻² Torr) by using a pressure transducer (MKS, type 660). First, the outlet vent needle valve is closed and the cell is evacuated so that the pressure transducer reads 0.0 Torr. Next, the cell and dosing manifolds are isolated from the vacuum manifold by closing a glass plug valve. Lastly, the pressure inside of the cell is monitored by using the pressure transducer shown in Fig. 3. If the pressure inside of the cell does not increase by more than 0.1 Torr during a 15 min period, then the cell is assumed to be leak-tight.

The pressure transducer on the dosing manifold in Fig. 3 is also used to determine the initial and final amounts of titrant present in the IR cell and manifold. A calibration vial (V_c = 23.99 cm³) connected to the glass manifold is used to measure the total volume of the IR cell, transfer line, and dosing manifold as shown in Fig. S14. The calibration vial and dosing manifold are filled with He to an initial pressure (P_i), measured by the pressure transducer. Then, the calibration vial is isolated from the dosing manifold by closing a glass needle valve, and the dosing manifold, transfer line, and IR are evacuated to ∼10⁻² Torr. The calibration vial isolation valve is then opened and the pressurized He is allowed to equilibrate throughout the dosing manifold, IR cell, and transfer line. The final pressure (P_f) of the system is recorded and the total system volume (V_t), including dosing manifold, IR cell, and transfer line, can be calculated by using the ideal gas equation,

where n_i and n_f are the initial and final number of moles of gas present, respectively; R is the universal gas constant (8.314 J mol⁻¹ K⁻¹); T is the system temperature (K); and V_t is the total volume (m³). Thus, by knowing V_t along with P_i and P_f for each dose, assuming that residual titrant does not adhere to the quartz cell or tubing throughout the system, the total moles of titrant adsorbed on the sample wafer can be determined.

Liquid-phase titrants, such as pyridine and CD₃CN, are purified via three freeze-pump-thaw cycles prior to use, as described elsewhere,⁶ while gaseous titrants, such as CO, are admitted directly into the IR cell from the dosing manifold. The pressure transducer on the dosing manifold in Fig. 3 is used to sequentially introduce controlled aliquots of the titrant into the IR cell. After each dose, the sample is allowed to equilibrate until the pressure stabilizes inside of the cell and dosing manifold (typically 180 s). The final pressure is used in conjunction with Equations (1) and (2) to determine the moles of titrant adsorbed on the sample. In cases of sub-saturation coverage where the final pressure is 0.0 Torr, all of the titrant is assumed to be adsorbed on the catalyst wafer.⁶ IR spectra are subsequently collected at 2 cm⁻¹ resolution by averaging 64 scans between 400 cm⁻¹ and 4000 cm⁻¹ relative to an empty cell background reference under vacuum. Titrant dosing cycles are repeated until saturation coverage is observed in the IR spectra.

Zeolites beta (^*BEA) and mordenite (MOR) are large-pore aluminosilicates with ordered microporous frameworks that contain 12-membered ring (12-MR) pores that are ∼7 Å in diameter. ^*BEA consists of a 3-dimensional 12-MR channel network while MOR consists of a 1-dimensional 12-MR system with smaller, 8-MR side pockets located along the inner surface of the 12-MR channels. Isomorphic substitution of a lower valence heteroatom, such as Al³⁺, for a Si⁴⁺ within the lattice results in the formation of a Brønsted acid site (H⁺) to maintain charge neutrality.^29–31 Additionally, metal heteroatoms (e.g., Al³⁺, Ti⁴⁺, Zr⁴⁺, and Sn⁴⁺) may be deposited on the zeolite surface as extra-framework species or incorporated into the framework structure to create Lewis acid sites of varying strength and reactivity.^32–34 As described previously, the interaction of CO at cryogenic temperatures (T < 150 K) with H⁺ sites, cationic metals, and surface hydroxyl (OH) groups enables detecting and quantifying these sites within zeolites such as ^*BEA and MOR, by monitoring characteristic shifts in the ν(C–O) IR vibrational frequency relative to gas-phase CO (2143 cm⁻¹) as shown in Fig. 4.

As an example application of the cell, transmission IR spectra were collected during cryogenic CO adsorption onto commercial samples of ^*BEA and MOR in order to characterize the Brønsted and Lewis acid sites present. Prior to characterization in the IR cell, NH₄⁺-exchanged zeolite samples were converted to the H⁺ form by calcination in 100 cm³ min⁻¹ (g zeolite)⁻¹ dry air (Matheson, Ultra Zero Grade) at 823 K for 4 h (14.4 ks) in a muffle furnace. For H-^*BEA (Zeolyst, Si/Al = 18), a 41.4 mg sample of the calcined material was pressed into a self-supporting wafer and placed within the quartz sample holder. The sample was then pre-treated under a flow of dry air (20 cm³ min⁻¹) inside of the assembled IR cell at 823 K for 1 h (3.6 ks) in order to remove residual moisture and physisorbed compounds from the catalyst wafer and quartz cell. The catalyst wafer was held under dynamic vacuum (∼10⁻² Torr) at 823 K for 1 h (3.6 ks) and then cooled under He (Matheson, 99.999%) flow (20 cm³ min⁻¹) to 128 K for adsorption experiments with CO as shown in Fig. 5. The catalyst wafer was cooled to below 120 K from 800 K in approximately 1.3 ks (∼22 min). As shown in Fig. 5, the temperatures of the quartz cell interior (T_cell) and brass block (T_block) remained within 33 ± 8 K of one another during cooling from 800 K to 303 K and within 20 ± 6 K from 300 K to 100 K. For CD₃CN adsorption studies at 303 K, the cell can be cooled from 800 K to 303 K in approximately 0.5 ks (∼9 min).

CO (Matheson, 99.998%) was subsequently introduced into the IR cell from the dosing manifold (Fig. 3) in sequential doses (∼0.2–0.5 × 10⁻⁶ mol) as shown in Fig. 6(a). As shown in Fig. 6(a) inset, bands associated with extra-framework Al³⁺ Lewis sites at 2230 cm⁻¹ and 2221 cm⁻¹ along with penta-coordinated Al³⁺ Lewis sites at 2190 cm⁻¹ are titrated first at the initial CO doses (0.2–1.6 × 10⁻⁶ mol).^16,17 After higher doses of CO are introduced to the sample, the band attributed to CO adsorbed on Brønsted sites (H⁺)³⁵ appears at 2174 cm⁻¹ followed by a band attributed to silanol (SiOH) groups³⁶ at 2155 cm⁻¹. The lower frequency bands at 2143 cm⁻¹ and 2132 cm⁻¹ reflect gas-phase CO. A small, unknown feature appeared at 2215 cm⁻¹; however, the intensity of this band did not increase upon subsequent doses of CO and may be due to a surface or gas-phase impurity. Integrated band areas for each type of surface feature were determined by curve-fitting the spectra, as shown in Fig. 6(b) for H-^*BEA after adsorption of 5.6 × 10⁻⁶ mol of CO (i.e., 5.6 μmol CO) at 128 K. A 30-70 mixed Gaussian-Lorentzian curve shape was found to produce the most accurate fit of the experimental data.

The integrated IR band areas were multiplied by the wafer cross sectional area (A_C = 2.54 cm²) and plotted against the cumulative amount of CO adsorbed on the H-^*BEA sample as shown in Fig. 7. Only H⁺ sites identified by the IR band at 2174 cm⁻¹ are selectively titrated during the region of cumulative adsorption between 5.6 and 17.2 μmol of CO adsorbed. After subsequent doses, CO also titrated SiOH sites as shown by the increase in the IR band area at 2155 cm⁻¹. Since the product of the 2174 cm⁻¹ IR band area and A_C varies linearly with the moles of CO adsorbed in the region of 5.6–17.2 μmol, the slope of this line represents the integrated molar extinction coefficient (E) of 0.95 ± 0.10 cm μmol⁻¹. This experimentally determined E value corresponds to the ν (C–O) vibrational mode of CO adsorbed at H⁺ sites. During the initial CO doses on H-^*BEA between 0.2 and 3.3 μmol, only the extra-framework (2230 cm⁻¹ and 2221 cm⁻¹) and penta-coordinated (2190 cm⁻¹) Al³⁺ Lewis sites were titrated as shown in Fig. 7(b). However, individual E for these two sites could not be determined because all Lewis acid sites were simultaneously populated by CO.

In transmission IR spectroscopy, the absorbance of IR light is directly proportional to the concentration of adsorbed CO on the catalyst surface. Thus, the concentration of H⁺ sites per gram can be determined based on the integrated area of the 2174 cm⁻¹ band at saturation according to³ 

where C (H⁺) is the concentration of Brønsted acid sites (μmol g⁻¹); r is the radius of the catalyst wafer (cm); W is the catalyst mass (g); and IA(H⁺) is the integrated area of the H⁺ IR band (cm⁻¹). Based on the E (H⁺) of 0.95 ± 0.10 cm μmol⁻¹ measured at initial CO coverages, and the IA(H⁺) of 11.97 cm⁻¹ for the 2174 cm⁻¹ band at saturation CO coverages, the concentration of Brønsted acid sites was determined to be 387 ± 20 μmol H⁺ g⁻¹ or 0.61 ± 0.20 H⁺/Al. By comparison, the concentration of Brønsted acid sites on this sample was measured to be 0.82 ± 0.04 H⁺/Al according to an NH₃ temperature-programmed desorption (TPD) protocol that has been described by Di Iorio et al.³⁷ Thus, the results obtained from the IR cell agree to within the uncertainty of those obtained from NH₃-TPD.

For H-MOR (Clariant, Si/Al = 12), 72.6 mg of the calcined material was pressed into a self-supporting wafer, pre-treated in the assembled IR cell in the same manner as H-^*BEA, and then cooled under He flow prior to dosing CO at T = 128 K. As shown in Fig. 8(a), extra-framework and penta-coordinated Al³⁺ Lewis acid sites were observed by bands at 2230 cm⁻¹ and 2195 cm⁻¹, respectively, along with Brønsted acid sites as shown by the band at 2176 cm⁻¹. These observations agree with those reported by Gounder and Iglesia¹⁶ for a series of H-MOR samples. Upon further doses of CO up to saturation coverage, bands indicative of CO at SiOH groups (2158 cm⁻¹) and gas-phase CO (2143 cm⁻¹ and 2132 cm⁻¹) were also observed (Fig. 8(b)). At saturation, the IA(H⁺) of the IR band at 2178 cm⁻¹, as shown in Fig. 8(b), was 25.02 cm⁻¹, which provided a Brønsted acid site concentration of 462 ± 24 μmol H⁺ g⁻¹, or 0.70 ± 0.20 H⁺/Al based on the E(H⁺) of 0.95 ± 0.10 cm μmol⁻¹. By comparison, the concentration of Brønsted acid sites on this sample was measured to be 0.59 ± 0.03 H⁺/Al according to NH₃-TPD. As with H-^*BEA, the quantification of H⁺ sites obtained from CO IR and NH₃-TPD for H-MOR agrees to within the measurement uncertainty.

A versatile, temperature-controlled transmission IR gas cell was developed in order to identify and quantify various surface sites present on heterogeneous catalysts. This cell was designed to accommodate a variety of probe molecules across a wide temperature range in order to broaden spectral measurement capabilities and perform temperature-controlled adsorption and reactivity studies on heterogeneous catalytic sites. Specific capabilities that distinguish this IR cell from previously reported designs include the following:

1. Use of custom, hand-tightened Cajon fittings to provide leak-tight operation across the cell operating range. These fittings reduce experimental complexity by enabling rapid replacement of catalyst samples without requiring disassembly of the temperature-controlled chamber for cell removal. Furthermore, the Cajon fittings do not require flanges and gaskets or sealants to fasten the CaF₂ IR windows onto the cell.

2. An integral temperature controller with adjustable heating and cooling rates that can be specifically tuned for different temperature regions. The temperature controller accommodates cell operation from 100 to 1000 K, in turn, enabling a wide range of temperature-programmed capabilities.

3. Incorporation of standard powder catalysts as self-supporting wafers for ease of sample preparation, compatibility with a variety of probe molecules, and examination under in situ gas treatment conditions.

The application of this transmission IR gas cell for quantification of acid sites within zeolites by cryogenic adsorption of CO has been demonstrated for H-^*BEA and H-MOR samples and independently verified by using an NH₃-TPD protocol. The IR cell can be easily adapted to examine the kinetic properties of various probe molecules on the surfaces of additional zeolites, molecular sieves, metal-oxides, and supported metal catalysts.

See supplementary material online for itemized lists of materials to construct the transmission IR gas cell and ancillary temperature controller, additional design drawings and photos of the cell, temperature-controlled chamber, temperature controller, an electrical wiring diagram of the temperature controller, a process flow diagram of the transmission IR cell and chemisorption unit, and details for the IR sample wafer preparation.

Support for this research was provided by the Purdue Process Safety and Assurance Center (P2SAC), and the U.S. Department of Energy, Office of Basic Energy Sciences, Catalysis Science Grant No. DE-FG02-03ER15408. We would like to thank David Taylor for design and fabrication of the temperature controller for the IR cell. We would also like to thank James R. Zimmerman in the Department of Chemistry at Purdue University for developing schematics of the temperature controller. Additionally, we thank Kris L. Davis and Verlin D. Lindley of Research Machining Services at Purdue University for design and fabrication of the temperature-controlled chamber for the IR cell.