Interrogation of the Plasma-Catalyst Interface via In Situ/Operando Transmission Infrared Spectroscopy

Abstract

Plasma-surface coupling has emerged as a promising approach to perform chemical transformations under mild conditions that are otherwise difficult or impossible thermally. However, a few examples of inexpensive and accessible in situ/operando techniques exist for observing plasma-solid interactions, which has prevented a thorough understanding of underlying surface mechanisms. Here, we provide a simple and adaptable design for a dielectric barrier discharge (DBD) plasma cell capable of interfacing with Fourier transform infrared spectroscopy (FTIR), optical emission spectroscopy (OES), and mass spectrometry (MS) to simultaneously characterize the surface, the plasma phase, and the gas phase, respectively. The system was demonstrated using two example applications: (1) plasma oxidation of primary amine functionalized SBA-15 and (2) catalytic low temperature nitrogen oxidation. The results from application (1) provided direct evidence of a 1% O₂/He plasma interacting with the aminosilica surface by selective oxidation of the amino groups to nitro groups without altering the alkyl tether. Application (2) was used to detect the evolution of NO_X species bound to both platinum and silica surfaces under plasma stimulation. Together, the experimental results showcase the breadth of possible applications for this device and confirm its potential as an essential tool for conducting research on plasma-surface coupling.

1. Introduction

Non-thermal plasmas (NTPs) are partially ionized gases characterized by a non-equilibrium in temperature between high energy electrons (>10,000 K) and low energy ions and neutrals (∼ambient temperature).¹ The energetic electrons are the primary driver for promoting gas-phase and/or surface reactions through collisions with neutral gas-phase molecules to form electronically, vibrationally, and rotationally excited species as well as radicals that react with increased activity over ground-state species² and surfaces to induce morphological or chemical changes (e.g., sputtering or etching).³ These excited species can further collide with other gas-phase species to form high energy mixtures of electrons, ions, radicals, excited species, metastables, and photons while maintaining a bulk-gas temperature near ambient.²

Low-temperature plasma-induced interactions have been exploited for a wide range of applications including medical treatments,⁴⁻⁷ pollutant removal,^8,9 electronics manufacturing,^5,10 material synthesis,^2,5,11−16 and heterogeneous catalysis.^2,17−24 Plasma-assisted catalysis has seen a significant growth in research activity over the past decade, primarily motivated by the ability to activate stable molecules (e.g., CH₄,^23,25−29 CO₂,^{23,25−27,29−32} and N₂^22,33−38) at relatively low temperature through electrical energy input. However, the complexity of plasma-phase interactions with reactive surfaces has resulted in many empirical studies on the performance of different catalytic materials with limited information on the fundamental elementary processes occurring at the plasma–catalyst interface.^2,21 Therefore, a transition toward operando and in situ characterization approaches can advance the field by directly observing how plasma stimulation affects the evolution of surface speciation. Of additional importance is the need for simple and inexpensive plasma modules capable of interfacing with existing and commonly utilized in situ/operando characterization techniques. Wide access to these systems provides researchers with a tool to correlate product output with time-resolved characterization of both the catalyst surface and the plasma phase.

Recent progress in this area has largely focused on identifying adsorbates on a metal surface under plasma stimulation and correlating plasma input conditions with observations. For plasma-assisted catalytic ammonia synthesis, Barboun et al. used inelastic neutron scattering (INS) combined with sequential plasma and thermal treatments of N₂ and H₂ to identify NH_X species formed on Ni/γ-Al₂O₃ under plasma exposure, which were not observed when solely thermal treatments were applied.³⁹ Alternatively, Gibson et al. probed the local temperature, oxidation state, and overall structure of Pd nanoparticles for methane oxidation with in situ X-ray absorption spectroscopy under plasma stimulation.⁴⁰ While these studies represent significant contributions toward understanding specific plasma-catalyst systems, the techniques employed require specialized synchrotron radiation facilities, which can impede research advancement due to limited access to instrumentation for data collection.

Other approaches to observe the plasma-catalyst interface have exploited the ubiquity and versatility of Fourier transform infrared (FTIR) spectroscopy to detect functional groups on a solid surface. Stere et al. combined a He plasma jet with a modified diffuse reflectance (DRIFTS) cell to successfully identify isocyanate as the reactive surface intermediate in the low-temperature, Ag-catalyzed selective catalytic reduction of NO_x.⁴¹ Other in situ/operando plasma studies have also used DRIFTS cells,^42,43 including a design⁴⁴ for a small-scale, helical surface discharge (HSD) developed by Turan et al. for millimeter-scale applications. While the HSD design and others based on DRIFTS are relatively simple to use and require minimal construction/modification, the cost of a new DRIFTS cell coupled with permanently modifying it may suppress interest and prevent widespread adoption. Additionally, from our experience with the HSD design, HSD plasma ignition can alter the sample packing through ejection of powder from the sample cup causing difficulties when correctly subtracting the background spectrum from the collected data. More sophisticated FTIR spectroscopy techniques have also been used, such as polarization-modulation infrared reflection-absorption spectroscopy (PM-IRAS), which uses both p- and s-polarized light to continually discriminate between gas-phase signals and surface-bound signals. Recently, an Ar-plasma jet and an optical emission spectroscopy (OES) probe were incorporated in a PM-IRAS cell to successfully observe surface-, gas-, and plasma-phase species simultaneously for the NTP-activated, non-oxidative coupling of methane on a model surface.⁴⁵

While the infrared-based designs mentioned above have been largely successful at generating useful information about the plasma-surface interface, the sample configuration and metal housing of DRIFTS and PM-IRAS cells present additional engineering challenges when combining the high-voltage discharge with a partially exposed sample. These challenges have previously been addressed by generating a plasma jet in an inert gas with a low breakdown voltage (e.g., He or Ar) and then bringing the plume close enough to interact with the sample.^41,45 While this approach does expose the sample to a plasma, it is unclear how the plasma characteristics change relative to a discharge generated directly in the reactive gas.⁴⁶ Unlike DRIFTS and PM-IRAS, transmission IR requires the sample to be suspended directly in the path of the IR light source, which passes through the sample to the detector. This configuration allows electrodes to be placed on either side of the sample, generating a plasma discharge directly in the reactive gas encompassing the surface of interest. Several designs modifying traditional transmission cells have been implemented in recent years,⁴⁷⁻⁵² including a glow-discharge cell designed by Rivallan et al.,⁵³ which has already been used in several studies on NTP-activated CO₂ hydrogenation at pressures around 10^–3 atm.⁵⁴⁻⁵⁶ However, molecular gases have higher breakdown voltages and thus require designs that either (1) increase the applied voltage, (2) decrease the electrode gap, or (3) apply vacuum pressure to achieve a stable discharge according to Paschen’s Law.⁵⁷ Because lab-scale modules are limited by how much voltage can safely be applied, a balance must be established between (2) and (3): reducing the electrode gap will restrict spectroscopic access to the sample while decreasing pressure causes the in situ conditions to deviate from real atmospheric-pressure plasma reactors, resulting in the “pressure gap” problem commonly faced in catalysis research.⁵⁸

Here, we present an alternative, facile design for a plasma-transmission IR cell capable of surface-sensitive, in situ/operando measurements under a wide range of plasma conditions. The design and function of the cell are comparable to commonly used dielectric barrier discharge (DBD) plasma-catalytic reactors, and its inexpensive construction, quick sample preparation, and ability to simultaneously interface with commonplace analytical techniques such as FTIR, OES, and mass spectrometry (MS) promote its adoption as a widespread tool for rapidly obtaining large volumes of in situ plasma-catalysis data. The effectiveness of the cell was demonstrated using (1) plasma-assisted oxidation of 3-aminopropyl SBA-15 (AP-SiO₂) to establish plasma-surface contact and (2) low-temperature plasma-assisted nitrogen oxidation over a Pt/SiO₂ catalyst to provide an example of plasma–catalyst interactions.

2. Experimental Section

2.1. Design Overview

The design of the transmission plasma-IR cell was initially inspired by the transmission cell constructed and reported by Cybulskis et al.⁵⁹ Therefore, it can serve both plasma and thermal purposes once constructed (Figure 1). Additionally, the overall form of the plasma generation component can be easily adapted to other similar transmission cell designs that incorporate long quartz tubes to house the sample. Key adaptations to the original design are discussed, including (1) the addition of adjustable, non-invasive copper electrodes with an off-center sample placement, (2) the tapered cell diameter, (3) the temperature control system, and (4) the implementation of an OES probe.

Figure 1.

Plasma transmission cell (a) during plasma operation (b) and during thermal operation (c).

2.2. Dimensions

A detailed schematic of the cell and sample holder with dimensional labels is provided in Figure 2. The cell is a fused quartz tube of length 157.2 mm with a tapered center (14.0 mm OD; 12 mm ID) and two electrodes positioned coaxially outside the tube. The cell is sealed by fittings identical to those used by Cybulskis et al. The tube length is specific to the sample compartment of a Bruker Vertex 70 FTIR spectrometer equipped with a liquid nitrogen-cooled mercury-cadmium-telluride (MCT) detector. The materials of construction for the sample holder are similar (fused quartz tube; 10.7 mm OD, 6.0 mm ID) with a recessed slit to hold a 7.0 mm diameter sample wafer. The sample can be heated in situ by heating tape tightly wrapped in a spiral around the full length of the cell (Figure 1). The electrodes are made from 8.5 mm wide copper tape with an electrically conductive adhesive (Kraftex) and wrapped ∼1.1–1.5 times around the outer diameter with leading edge positions at 28.8 and 68.8 mm from the inlet of the tube. The electrodes are covered by insulating Kapton tape (Equity Baymers) to prevent unwanted discharging outside the cell. Additionally, all nearby metals are adequately grounded to ensure that a hazardous buildup of charge does not occur. The vacuum system features a needle valve placed downstream of the cell to control the applied pressure from an in-line vacuum pump (Welch Vacuum 2014B-01), allowing the operator to adjust the absolute pressure between 0.1 and 1.0 atm. The cell is suspended within the sample compartment of the spectrometer by two metal supports placed on either end of the tube. Each support is capable of elongating via an adjustable screw to position the sample wafer at the correct height within the incident beam.

Figure 2.

Orthographic (a) and isometric (b) views of the tube and sample holder (not to scale). Outside edges are solid lines, and inside edges are dashed lines. The copper electrodes are shaded light gray, the sample wafer is shaded black, and the quartz body is left unshaded.

2.3. Configuration

This unique electrode geometry was inspired by the design first reported by Teschke et al. for generating plasma jets,⁶⁰ and its non-invasive arrangement eliminates concern of possible catalytic activity originating from the electrode material. Additionally, the impermanence and adjustability of the electrodes provide significant freedom when choosing where to generate a NTP within the cell (e.g., post-plasma catalysis versus in-plasma catalysis studies). We found that the maximum achievable power and the absence of unwanted discharging to the metal fittings on the ends of the cell occurs with the high-voltage electrode at the center of the cell with the ground electrode near the inlet. This arranges the high voltage electrode equidistant from the grounded metal fittings and ensures the ground electrode offers the path of least resistance for plasma generation. To prevent the sample holder from moving axially along the cell during setup, optional indentations located 66 mm from the inlet of the tube were added to serve as backstops (omitted from Figure 2 for clarity). This location was chosen specifically so that the sample holder can either be centered between the electrodes (depicted in Figures 1 and 2) or at the midpoint of the cell. Additionally, the diameter of the cell was decreased by 44% compared to the Cybulskis et al. design⁵⁹ to facilitate gas breakdown at vacuum pressures closer to atmospheric by reducing the effective discharge gap, d_eff (eq 1). This increases the maximum operating pressure capable of achieving a stable discharge at the same breakdown voltage in the collision dominated regime of Paschen’s Law.⁵⁷

1

Here, r is the radius of the ring electrodes and d is the linear distance between electrodes. Equation 1 shows that the effective discharge gap, d_eff, scales with r^3/2 and d, indicating that the radius of the ring electrode, r, has a more significant contribution toward decreasing the effective discharge gap and increasing the maximum operating pressure compared to the actual gap distance, d, between the electrodes. The smaller diameter tube was also chosen to fit a smaller sample wafer to minimize potential temperature and concentration gradients within the sample. As noted by Cybulskis et al., studies seeking quantitative determination of reaction rates and kinetic parameters can estimate the influence of potential temperature and concentration gradients for a specific reaction using Mears’ criteria.^59,61 Additionally, the compact size of the Specac Mini Pellet Press used in this study combined with shut-off valves on both ends of the transmission cell allow each to be easily moved into and out of a N₂ atmospheric glovebox for preparation of highly air-sensitive samples.

2.4. Heating System

In our experience, internal temperature control poses a significant engineering challenge in both in situ spectroscopic systems and plasma-based systems. For most in situ spectroscopic systems, the limited space available in the instrument sample compartment prohibits the use of full-size tube furnaces that are commonly used in lab-scale reactors to control the temperature. The parent design overcame this issue by controlling the temperature of a bore-through brass block, which was able to rapidly heat and cool the sample thanks to a feedback temperature controller combined with the high thermal conductivity of brass.⁵⁹ However, we have found that for both this plasma system and others, the high voltage electrode tends to discharge to metal heating elements and thermocouples placed near the plasma zone, making it difficult to measure the bulk gas temperature and preventing feedback temperature controllers from being used. Additionally, the chosen heating element should be easily obtainable for any researcher and quickly removeable/replaceable between thermal and plasma treatments to conduct sequential studies similar to those performed by AlQahtani et al.²⁴ and the INS studies mentioned previously.³⁹ The solution chosen involves the use of standard heating tape commonly found in any laboratory setting wrapped in a co-axial spiral configuration around the quartz tube (Figure 1). The bulk gas temperature near the sample wafer is then calibrated with a thermocouple against the output of a variable transformer (Variac) that controls the input power to the heating tape. Two different heating tape models were used: model A (BriskHeat BIH051020: 105 W, 0.5″ width × 24″ length) and model B (BriskHeat FGH051-040 L: 210 W, 0.5″ width × 48″ length). To prolong the lifetimes of both heating tape models, model A is predominantly used for thermal applications requiring higher temperatures up to 400 °C, where model B is only used for applications below 300 °C. If additional temperature control during thermal operation is desired (e.g., for heating at a controlled ramp rate), thermocouples can be inserted into the sample volume to allow operation with feedback control systems.

2.5. Integration with OES and MS

The transmission cell can be easily integrated with additional probes to analyze the plasma phase and effluent gases exiting the reactor zone. In this study, an OES probe (Ocean Optics USB2000 + XR1 spectrometer with THORLABS F280SMA-A collimator and Ocean Insight QP300–1-SR fiber optic cable) was positioned external to the transparent quartz cell body via a ring stand clamp to observe the photons emitted from plasma-stimulated species within the cell. Additionally, a Pfeiffer OmniStar process MS (200 amu) was affixed to the effluent side of the reaction chamber to analyze the production and evolution of stable, gas-phase products.

2.6. AP-SiO₂ Synthesis

SBA-15 was synthesized according to previous reports.⁶² SBA-15 was dried under vacuum at 200 °C and stored in a N₂ drybox. The SBA-15 was post-synthetically modified with primary amine functionalities through reaction with 3-aminopropyltrimethoxylsilane (Sigma-Aldrich 97%).⁶³ In a typical procedure, 2.0 g of 3-aminopropyltrimethoxysilane was added to 2 g of SBA-15 dispersed in dry toluene (∼50 mL). The mixture was then stirred in a nitrogen atmosphere for 24 h at room temperature to homogenize the aminopropyl functionalization on the silica surface. A white solid was filtered from the mixture, washed with dry toluene, dried under vacuum at 70 °C for 3 h, and stored in the N₂ drybox. The amine loading of the resulting AP-SiO₂ sample was determined as 1.9 mmol/g by thermogravimetric analysis of a temperature programmed oxidation.

2.7. Pt/SiO₂ Catalyst Synthesis

The 5 wt % Pt/SiO₂ catalyst was synthesized onto a fumed silica support (Spectrum Chemical: Aerosil 200, 175–225 m²/g) by an incipient wetness impregnation procedure. The fumed silica (2.00 g) was first dried and calcined at 200 °C for 12 h. Next, 0.30 g of dihydrogen hexachloroplatinate(IV) hexahydrate (Beantown Chemical, 99.95%) was dissolved in 2.52 g of deionized water and added dropwise to the fumed silica while stirring. The material was dried at 120 °C for 12 h and reduced in pure hydrogen at 200 °C for 5 h (ramped at 1 °C/min). The site density of the Pt/SiO₂ was 44.1 μmol/g measured by CO pulse chemisorption. The surface areas of the fumed silica support (206.8 m²/g) and the Pt/SiO₂ (194.6 m²/g) were measured by N₂ physisorption analysis at 77 K (Quantichrome NOVA 2200e).

2.8. Procedures for Application 1: Plasma Oxidation of Primary Amine-Functionalized SBA-15

The AP-SiO₂ sample was first prepared into a self-supporting wafer using a hydraulic pellet press (Specac Mini Pellet Press). The 7 mm diameter wafer was pressed with 1.0 ton of force for 5 min. The sample was subsequently weighed and placed into the sample holder inside the transmission cell. The system was checked for leaks prior to operation. The sample was then pretreated at 120 °C for 1.5 h under 100 mL/min of 0.1 atm pure He flow (Airgas 99.997% purity). Pretreatment at 120 °C was chosen based on CO₂ desorption experiments performed on similar materials.⁶³ For the thermal only experiment, a background spectrum was taken under the pretreatment conditions at the end of 1.5 h. The gas feed was then adjusted to 100 mL/min of 1% O₂/He (Airgas mixed tank, 0.9999 mol % O₂, balance He) with all other conditions maintained. Scans were taken every 5 min. For the plasma experiment, the sample was cooled after pretreatment to room temperature under 100 mL/min of 0.1 atm pure He. The plasma was then switched on under the same He flow and adjusted to the desired setting. A 2 W plasma was used, which corresponds to a specific energy input (SEI) of 1200 J/L. Here, SEI is defined as the plasma power divided by the total gas flow rate. The plasma was kept relatively low to minimize the resistive heating (outer surface temperature was maintained at 60 °C and recorded by thermal imaging camera: RIGID tools, model RT-9x) and to ensure the bulk gas temperature remained well below 120 °C. Because the background spectrum is dependent on the temperature, the plasma was allowed to reach its autogenous temperature under the inert gas flow for 10–15 min before the background spectrum was recorded. Then, the gas feed was adjusted to 100 mL/min of 1% O₂/He with all other conditions maintained. The plasma power was measured by integration of charge-voltage plots,⁶⁴ and IR scans were recorded every 5 min.

2.9. Procedures for Application 2: Catalytic Low-Temperature Nitrogen Oxidation

The Pt/SiO₂ and SiO₂ samples were first prepared into 7 mm-diameter self-supporting wafers using the same hydraulic press (Specac Mini Pellet Press). Both samples were undiluted and pressed with 1.75 tons of force for 5 min. The sample wafers were then weighed and loaded into the cell, and the system was checked for leaks. Both Pt/SiO₂ and SiO₂ were pretreated at 300 °C for at least 1 h under 30 mL/min of 0.3 atm pure H₂ flow (Airgas 99.999% purity). The higher pretreat pressure was chosen to increase the gas collisions with the surface and shorten the time required for pretreatment. For the thermal only experiment, the Pt/SiO₂ pellet was allowed to cool after pretreatment to 150 °C under 0.1 atm of 100 mL/min He and 30 mL/min N₂ (Airgas 99.999% purity) flow, and a background spectrum was recorded. The gas feed was then switched to 100 mL/min of 1% O₂/He and 30 mL/min N₂ with all other conditions maintained. IR scans were recorded every 5 min.

For the plasma experiments, the sample wafer was allowed to cool to room temperature under 0.1 atm of 100 mL/min He and 30 mL/min N₂ flow. The plasma was then switched on under the same He/N₂ flow and adjusted to the desired setting. A 0.3 W plasma was used, which corresponds to an SEI of 138 J/L. The sample was allowed to reach autogenous temperature for 10–15 min before a background spectrum was recorded. The gas feed was then switched to 100 mL/min of 1% O₂/He and 30 mL/min N₂ with all other conditions maintained. The plasma power was subsequently measured by integration of charge-voltage plots,⁶⁴ and IR scans were recorded every 5 min.

3. Results and Discussion

3.1. Thermal Control Validation

Before the in situ/operando experiments were conducted, we first sought to statistically quantify the accuracy and reproducibility of the heating system in the plasma cell for gases of varying thermal conductivity (He and N₂) at both high (0.3 atm) and low (0.1 atm) pressures. The internal temperature was measured by a thermocouple placed near the sample holder, and the heating tape was removed and replaced after each trial for a total of three trials per data set. The calibrations given in Figure 3 show temperature control to within ±10 °C with 95% confidence. The calibrations also show that temperature control is insensitive to both the thermal conductivity of the gas and the absolute pressure over the normal operating range of the system (0.1–0.3 atm).

Figure 3.

Variac temperature calibration for use with heating tape models A (black) and B (red). Total flowrate is 100 mL/min. Temperature is measured by an internally placed thermocouple. Heating tape was removed and replaced after each trial. Error bars on model A represent one standard deviation after three trials per data set.

For temperature comparisons between thermal and plasma experiments, an estimation of the autogenous temperature of the plasma is also required. We varied both the pressure and applied power of a pure N₂ discharge over the full operating range of the cell and measured the resulting maximum surface temperature (located on the surface of the high voltage electrode) with a thermal imaging camera. The results are given in Figure 4, which serves to both estimate the resistive heating of the plasma and show how the pressure can be adjusted to shift the range of accessible plasma powers.

Figure 4.

Resistive heating as a function of plasma power and overall pressure. The plasma was generated at a constant frequency (23.5 kHz) with an empty sample holder and 100 mL/min pure N₂ flow.

3.2. Application 1: Plasma Oxidation of Primary Amine-Functionalized SBA-15

The primary function of the transmission plasma-IR cell is the observation of plasma interactions with different types of solid materials. Therefore, we begin by providing evidence that the system can both induce and detect surface changes in response to plasma stimulation. To demonstrate this ability, plasma-induced oxidation of an organic/inorganic hybrid aminosilica material (AP-SiO₂) was conducted. The AP-SiO₂ (Scheme 1) used here is part of a widely studied class of functionalized materials with applications in CO₂ capture and heterogeneous catalysis.^63,65−69 This class of materials was chosen for several reasons, including (1) the distinct IR features of the organic functionality and oxidation products (e.g., NO_X), (2) their straightforward synthesis onto silica supports, (3) the high achievable loading of the 3-aminopropyl groups onto the SBA-15, and (4) the existing literature characterizing the behavior and stability of primary amines in oxygen environments. For example, several previous works have reported that primary amines will oxidize at low temperatures in the presence of reactive oxygen derivatives such as ozone (O₃).⁷⁰⁻⁷³ This process was shown to selectively form nitroalkanes in high yields without oxidizing the carbon framework. Therefore, we hypothesized the oxygen plasma would similarly generate nitroalkane derivatives from the primary amines on AP-SiO₂.

Scheme 1. Chemical Structure of Unreacted AP-SiO₂.

Bollini et al. demonstrated using solid-state ¹³C nuclear magnetic resonance spectroscopy that 3-aminopropyl groups on silica remain unchanged in pure O₂ up to at least 135 °C.⁷⁴ We used this finding as a thermal benchmark for comparison to plasma stimulation experiments. To verify this outcome in our own system, we conducted thermal control experiments with 100 mL/min of 1% O₂/He at 120 °C and 200 °C and obtained the same result. The data at 120 °C are provided in Figure 5a (black solid curve), which indeed shows no spectral changes were observed after 30 min of thermal oxygen treatment.

Figure 5.

FTIR (a), MS (b), and OES (c) data with observed surface species (d) from plasma (2 W; 60 °C; 8.6 mg; purple) and thermal (120 °C; 4.3 mg; black) oxidation of AP-SiO₂ in a 100 mL/min of 1% O₂/He feed at 0.1 atm. FTIR spectra are baseline corrected and background subtracted. OES spectrum is background subtracted to correct for stray light.

Next, an equivalent experiment was performed with plasma stimulation of the 100 mL/min of 1% O₂/He feed at 60 °C autogenous temperature. The background-subtracted spectra of the AP-SiO₂ surface after 30 min are shown in Figure 5a. Several changes were observed at the higher frequencies of the fingerprint region (1600–1300 cm^–1), indicating plasma-surface reactions are taking place that were not observed thermally. The absence of negative peaks in the C–H stretching region (2850–3000 cm^–1) suggests that the propyl backbones of the amino groups are not oxidized by the plasma and remain intact (inset of Figure 5a). This observation is supported by the MS trace for CO₂ (44 m/z) given in Figure 5b, which shows that no CO₂ was formed upon introducing O₂ (32 m/z) to the system. Additionally, no evidence for either NO (30 m/z) or NO₂ (46 m/z) was present from these experiments, further suggesting that the nitrogen remains tethered to the surface by the propyl backbone.

In Figure 5a, the negative peak in the N-H bending range (1590–1627 cm^–1) represents the disappearance of amino groups upon reaction with an oxygen plasma.⁷⁵ This combined with the generation of H₂O (18 m/z) shown in Figure 5b indicates oxidation of the surface amines. The likely products of this oxidation are R-NO₂ species,⁷⁰⁻⁷³ which are expected to have strong IR signals in the 1650–1300 cm^–1 region.⁷⁶ Looking at the FTIR spectra in Figure 5a, the sharp peaks at 1556 and 1381 cm^–1 are characteristic of the symmetric and asymmetric stretching of nitro (-NO₂) groups, respectively.⁷⁶ This result is in agreement with previous works that show high yields of nitroalkanes from reacting primary amines with ozone.⁷⁰⁻⁷³ The other prominent peak at 1435 cm^–1 (Figure 5a) is associated with the C–H bending in the alkyl backbone,⁷⁷ which is observed in both nitroalkanes and amines.⁷⁸ The increase in the intensity of the 1435 cm^–1 band with time is related to changes in the local chemical environment surrounding the alkyl backbone rather than the formation of C-H bonds.⁷⁹ Other smaller peaks are observed in Figure 5a, including carbonyl and imine stretches between 1650 and 1750 cm^–1.^80,81 However, the extinction coefficients for these species are relatively large,^80,81 indicating low surface concentration. The other peaks in this region likely arise from trapped water vapor, which has many strong rovibrational peaks in this region.⁸²

The plasma phase was characterized using OES, which detects photons emitted by various electronic relaxations of both atomic and molecular species. Figure 5c shows a representative OES spectrum of the 1% O₂/He plasma taken after 15 min. The spectrum has the He I feature at 706 nm (3³S → 2³P) as well as clear triplet O I features at 777 nm (3⁵P → 3⁵S) and 845 nm (3³P → 3³S).⁸³ The presence of electronically excited atomic oxygen is evidence that the dioxygen in the feed is dissociated by the plasma, resulting in the formation of O₃.⁸⁴ Due to the half-life of O₃ (10³ s) compared to the half-life of atomic oxygen (10^–6 s), O₃ is likely interacting with internal amines by diffusing within the pores of the SBA-15, whereas atomic oxygen can only interact with the species outside the pores.⁸⁴ Additionally, the OES spectrum shows strong features associated with the N₂(B³Π_g → A³Σ_u⁺) first positive system (FPS), N₂(C³Π_u → B³Π_g) second positive system (SPS), and N₂⁺(B²Σ_u⁺ → X²Σ_g⁺) first negative system (FNS),^83,85 which are presumed to be from air contamination. Spectral lines due to foreign gases are commonly seen in OES for even the purest He feed streams due to the ability of excited He atoms to efficiently ionize impurities through Penning ionization, causing the impurities to have disproportionately intense OES signals.^83,85 Additionally, the MS trace detected no nitrogen oxidation products when oxygen was introduced, and the vacuum system was checked for leaks before each experiment. Therefore, the concentration of air in the system is considered negligible. While any degree of contamination is generally undesirable, the N₂ rovibrational bands can be used to quickly estimate the rotational temperature of N₂ as a proxy for the bulk gas temperature.^85,86 Unfortunately, this technique requires a spectral resolution of at least 0.2 nm, which is outside the capability of the spectrometer used here (∼2 nm resolution).⁸⁶ Nevertheless, additional opportunities exist to couple this system with more specialized OES techniques that further characterize the plasma by calculating the electronic/vibrational temperature, the electron density, the degree of ionization, and the species mole fractions.

The FTIR, MS, and OES results clearly suggest that oxygen plasma selectively oxidized primary amines to nitro groups without altering the hydrocarbon tether (Figure 5d). The use of NTP to perform this transformation obviates the need for harmful reactants that would typically be used (e.g., hypofluorous acid, Caro’s acid, and organic peroxides), and the resulting nitro groups can be readily converted to other desired functional groups.⁷³ This has clear applications in the synthesis of biomaterials, where surface hydrophobicity needs to be carefully controlled to either prevent or promote the adhesion of cells and the growth of biofilms.⁸⁷ The plasma transmission cell was capable of monitoring the evolution of surface-bound species over time, demonstrating its usefulness as an in situ characterization technique.

3.3. Application 2: Catalytic Low-Temperature Nitrogen Oxidation

After establishing the capability of the transmission plasma-IR setup to directly observe plasma-surface interactions of organic/inorganic materials, we further demonstrate the application of this system to observing adsorbates on metal surfaces during plasma-assisted nitrogen oxidation. Rather than attempting to determine plasma-catalyst interactions as has been done previously,⁸⁸ these experiments are aimed at characterizing the catalyst surface in situ under realistic conditions.

Plasma technologies for chemical synthesis date back to the early 1900s when the Birkeland–Eyde process for nitric acid synthesis became the first ever plasma system to be commercially implemented.⁸⁹ A thermal, high temperature, equilibrium plasma reactor was used to produce nitric oxide (NO) by activating nitrogen and oxygen in the ambient air and absorbing the effluent stream in water to form nitric acid.⁸⁹ However, the process was ultimately abandoned by industry due to poor energy efficiency and low conversion.⁸⁹ More recently, there has been renewed interest in plasma-activated nitrogen activation, especially in combination with a catalyst to reduce the energy barrier for nitrogen dissociation.^{18,33−39,52,88} The nitrogen oxidation reaction is an ideal candidate to be studied by this design primarily because its large endotherm (N₂ + O₂ → 2NO; ΔH = 90 kJ/mol⁹⁰) makes it severely thermodynamically limited. Thus, its light-off temperature is beyond the capability of standard thermal IR cells (equilibrium concentration of NO in synthetic air is 100 ppm at 1000 K⁸⁹), requiring a non-thermal plasma to achieve high enough conversions to study the reaction in situ. Additionally, the products (NO_X) are highly IR-active, whereas the reactants (N₂ and O₂) are IR-inactive, which ensures that any IR signal observed is exclusively from the products.

To demonstrate the lack of thermal reactivity between N₂ and O₂, we performed a thermal control experiment on Pt/SiO₂ at 150 °C with 30 mL/min of N₂ and 100 mL/min of 1% O₂/He. As expected, the IR spectrum showed no detectable changes at the catalyst surface after 30 min of thermal exposure (Figure 6a). This result also eliminates concern that observed peaks in subsequent experiments are due to impurities from the gas tanks adsorbing onto the surface.

Figure 6.

FTIR (a,b), MS (c), and OES (d) data of 0.3–0.4 W plasma oxidation of N₂ with Pt/SiO₂ (red; 4.6 mg pellet) and SiO₂ (blue; 2.1 mg pellet) in a 30 mL/min N₂, 1 mL/min O₂, and 99 mL/min He feed at 0.1 atm. Thermal treatment (black) of Pt/SiO₂ was at 150 °C. Plasma autogenous temperature was 39 °C. FTIR spectra are baseline corrected and background subtracted. OES spectrum is background subtracted to correct for stray light.

Next, a second control experiment was performed under the same conditions on SiO₂ under a 0.3 W plasma treatment and 39 °C autogenous temperature. The FTIR results given in Figure 6b show a single peak in the N–O stretching region at 1648 cm^–1, which was assigned to nitrite species⁸¹ covalently bound to silicon as Si–O–N=O.⁹⁰ Although covalently bonded silicon nitrite is generally not observed in thermal NO_X-adsorption studies on silica materials,⁹¹ Chao and Lunsford first reported its presence after NO adsorption on decationated zeolite Y up to 200 °C.⁹⁰ They compared the results of several Y-type zeolites and only observed the silicon nitrite on the decationated Y, which contains oxygen vacancies as trigonal Si⁺ centers.⁹² This led to the conclusion that the Si⁺ oxygen vacancy was a required precursor to the formation of silicon nitrite species.⁹⁰ Because NTP is known to induce oxygen vacancy defects in silica,⁹³⁻⁹⁵ we can conclude from the IR data in Figure 6b that the plasma generated by the transmission cell is creating trigonal Si⁺ centers, which interact with NO produced in the plasma phase to form silicon nitrite groups.

The FTIR results for Pt/SiO₂ under the same plasma conditions as the SiO₂ control experiment are given in Figure 6a. The IR spectra have three main features at 1700-1800, 1653, and 1580 cm^–1, with the middle peak being previously assigned to silicon nitrite.⁹⁰ All peaks remained unchanged after the plasma was turned off and the cell was purged with He. Based on other literature reports, the features at 1700–1800 and 1580 cm^–1 are linear and bent NO species adsorbed on Pt, respectively.^96,97 The linear region is further divided into two distinct peaks around 1770 and 1720 cm^–1, which correspond to linearly bound NO on different Pt sites. From literature FTIR studies on single crystal Pt surfaces, we can assign the 1770 cm^–1 peak to the Pt(110) surface and the 1720 cm^–1 peak to the Pt(111) surface; no linearly bound NO is expected on Pt(100).⁹⁷⁻¹⁰⁰ These reports also suggest that the 1580 cm^–1 peak assigned to bent NO is broad enough to account for adsorption on all three low-index planes of Pt. As time-on-stream increases, Figure 6a shows (1) the 1770 cm^–1 peak intensity increases until reaching a steady state after 15 min, (2) the 1720 cm^–1 peak intensity reaches a maximum after 15 min and begins decreasing, and (3) the 1580 cm^–1 peak intensity steadily increases over the full 30 min. Therefore, the coverages of the corresponding species follow the same progression with time.

The MS traces of the expected nitrogen oxide products (NO 30 m/z, N₂O 44 m/z, and NO₂ 46 m/z) for both the SiO₂ and Pt/SiO₂ are given in Figure 6c. The production rates of all three products were below the detection limits for the MS used here, and no clear improvement of Pt/SiO₂ over SiO₂ was observed. This is not unexpected considering Ma et al. has shown that for O₂/N₂ molar ratios above 0.01 (here we use 0.03), the oxygen species dominate the Pt sites and slow the production of NO to near plasma only background levels with negligible amounts of N₂O and NO₂ formed.⁸⁸ The lack of detectable products is also likely a consequence of the low SEI used here (138 J/L), which has been shown to scale linearly with reaction rates due to decreased generation of discharge filaments containing excited species.³³ This was supported by the OES spectrum given in Figure 6d, which is representative of the plasma under these conditions because no changes in the OES spectra were observed over time or between materials. Figure 6d shows only features associated with the SPS (C³Π_u → B³Π_g) and FNS (B²Σ_u⁺ → X²Σ_g⁺) of N₂, which was likely excited through Penning collisions with He.^83,85 Additionally, the absence of atomic O/N lines and NO systems suggests that these species are either absent altogether or only present in low concentrations. However, the IR data showing adsorption of nitrogen oxides on the surface of the Pt/SiO₂ catalyst suggests that atomic O/N and molecular NO_X are likely present in the plasma-phase.

The goal of this example application was to demonstrate the ability of the cell to collect information on the system under realistic reactor conditions. A lack of information from both OES and MS highlighted the importance of studying plasma-catalysis in situ and with multiple characterization techniques to obtain the full picture. Here, the IR data was key in drawing conclusions about the catalyst under plasma exposure, giving merit to the design and demonstrating its functionality for gathering information about the plasma-catalyst interface.

Conclusions

This work presents a simple design for a plasma-capable IR transmission cell for probing a surface under plasma stimulation. The materials for constructing the IR cell are ubiquitous and inexpensive, ensuring that the technology is available to any researcher studying plasma-surface coupling. Additionally, the design is easily customizable to suit specific needs (e.g., use of vacuum manifold for controlled dosing,⁵⁹ in-plasma vs post-plasma configurations,² integration with step-scan technique for microsecond resolution,^53,55 etc.) and can be paired with additional analytical and/or characterization techniques such as MS and OES. The thermal control of the cell was tested at different pressures (0.1 and 0.3 atm) and feed gases (He and N₂), and the setpoint temperature was found to be reproducible to ±10 °C with 95% confidence across all conditions.

To verify the ability to both promote and recognize plasma effects on a surface, the oxidation of primary amine-functionalized SBA-15 was performed as an example application of the cell. The combined results from FTIR, MS, and OES lead to the conclusion that plasma generated reactive oxygen derivatives selectively oxidized the primary amines to nitro groups without further oxidation of the hydrocarbon linker to the support. This result was supported by literature, showcasing the ability to generate reliable conclusions about plasma-surface interactions. As a second example application, the cell was used to observe surface species on a Pt/SiO₂ catalyst for plasma-assisted nitrogen oxidation. IR results from the support-only experiment identified silicon nitrates formed by reaction with NO_X and plasma-generated oxygen vacancies on the silica surface. On Pt/SiO₂, both linear and bent NO species adsorbed on several Pt facets, and the cell was able to track the waxing and waning of these species on the catalyst surface over time. These results highlight the utility of this plasma cell and establish its capacity to study plasma-catalysis in situ.

The authors declare no competing financial interest.