Oxidation of ethylene glycol: Unity of chemical and electrochemical catalysis
Graphical Abstract
Is the chemical oxidation of ethylene glycol over a heterogeneous catalyst a manifestation of electrochemical reactions? The products observed indicate a compelling case.
Abstract
The catalytic oxidation of ethylene glycol, both chemically (reaction with oxygen) and electrochemically (application of potential), have been studied in a batch reactor on three different catalysts. The products have been identified and their concentrations measured. For experiments on chemical oxidation, in addition to the analysis of the products, the variation of the potential developed on the catalyst with time was measured. The electrochemical experiments were designed such that the potential of the catalyst-coated electrode varies with time in a manner identical to variation of the potential with time observed for chemical experiments. The products and their respective concentrations obtained for both the chemical oxidation of ethylene glycol with molecular oxygen as well as the electrochemical oxidation were found to be in excellent agreement. The data presented here suggest that at molecular level, catalytic chemical oxidation on heterogeneous catalyst involves transfer of electrons from ethylene glycol to oxygen via two or more electrochemical reactions through the catalyst.
1 INTRODUCTION
In the past few decades, the catalytic chemical oxidations of polyols have been studied extensively. As with most alcohols,[1-6] oxidative dehydrogenation mechanism has been generally accepted for the oxidation of most polyols including ethylene glycol (EG).[7-10] Quite often, the fragmentation of the carbon chain in polyols leads to aldehydes, ketones, and carboxylic acids with fewer carbon atoms.[7] A more aggressive condition will eventually lead to formation of carbon dioxide and water. It may be mentioned that most of these reactions are activated catalytically and comparatively more literature exists for catalytic chemical oxidation of ethylene glycol than normal oxidation. Although ethylene glycol is fairly resistant to total oxidation,[7] depending on the reactions conditions, and the catalyst employed, a wide range of products have been reported. For example, oxidation of ethylene glycol produces glyoxal on CuO[8] or promoted Ag/SiC catalyst[9, 10] but glycolates/glycolic acid on Au,[9] Pd,[10] and Pt.[1-6, 11, 12] Interestingly, oxidized fractions involving only one carbon atom (formic acid, formaldehyde, CO, and CO2) were only sporadically observed.[13-15] Meanwhile, the electrochemical studies have traditionally been more focused on the potential-current behavior on different catalysts as well as products upon oxidation.[16-22] Since the pioneering studies by Weber et al.[23] on oxidation of ethylene glycol over platinum, many products such as glyoxal, glycolaldehyde, glycolic acid, glyoxalic acid, oxalic acid,[13-15, 24, 25] and so on have been identified by chromatography.[25-29] Here also, primarily two pathways have been identified: direct oxidation leading to CO2 and indirect pathways leading to oxidized products which may eventually oxidize to CO2. However, the direct oxidation would imply transfer of 10 electrons and CO2 as product has seldom been observed. Thus, as far as the detection of CO2 is concerned both the chemical oxidation and electrochemical oxidation bear a degree of similarity. Interestingly, the C1 oxidized products such as formic acid/formaldehyde are also rarely observed in regular electrochemical experiments. Though, formic acid has been observed in some non-traditional configurations such as a channel flow reactor.[30]
Other than a superficial semblance in the above two oxidation schemes, there are hardly any systematic studies to compare the products upon oxidation. The disparity stems primarily from the contrasting experimental requirements. In chemical oxidation, typically both ethylene glycol and oxygen are fed to a reactor while in electrochemical studies special care is taken to avoid any interference by oxygen. This contrast can be resolved to a certain extent if it is argued that during chemical oxidation, the only role of oxygen is to accept electrons released upon oxidation of ethylene glycol and in the process the catalyst acquires a unique potential. In absence of oxygen as in electrochemical experiments, the whole process of managing electrons and the potential of the catalyst are delegated to the potentiostat. Thus, a comparison of products formed upon oxidation of ethylene glycol (both chemical and electrochemical) can be made if the electrochemical oxidation is performed at the same potential that develops on the catalyst when both ethylene glycol and oxygen are present. This potential though would vary with time and so the nature and concentration of different products. Of course, this would imply that at molecular level chemical catalysis and electrochemical catalysis are the same. This hypothesis though speculated by earlier authors, has recently been experimentally proven rigorously by us for a very simple system which reacts to form a single product in solution.[31]
This article tries to compare the products obtained from both chemical and electrochemical oxidation of ethylene glycol. For chemical catalysis, oxygen saturated solution of ethylene glycol was used and both the open circuit potential developed (OCP) on the catalyst and the product concentrations were monitored with time. On the other hand, for electrochemical experiments, argon saturated solution was used and the potential was varied with time matching the variation of OCP with time obtained during chemical experiments. For both these experiments, catalyst coated glassy carbon electrode in a rotating disk configuration was used. Initially, the catalyst used was Pt at three different rotational rates. Later, a comparison was also made with Pt-Ir and Pt-Ru catalysts at a single rotation rate. The identification and quantification of the products which were performed by HPLC (high performance liquid chromatography) revealed an excellent match between these two diverse oxidation pathways (chemical using molecular oxygen and electrochemical by application of potential). This gives further credence to the hypothesis that underlying a catalytic chemical reaction, an electrochemical mechanism involving electron transfers operates. The results presented here suggest that molecular oxygen does not participate in the reaction and its role is limited to accepting electrons and maintaining suitable potential. These results establish the correspondence and basic unity between chemical catalysis and electrochemical catalysis which in turn corroborate the results obtained for the chemical oxidation studies on glycerol using 18O2-labeled isotopes.[32]
2 RESULTS AND DISCUSSION
2.1 Chemical catalysis experiment
Figure 1 shows the variation of the OCP with time (8 h) obtained during the chemical experiments. It can be clearly observed that the OCP gradually increases with time for all the three scan rates. Interestingly, for approximately 3 h, the OCP for all the rotational rates were similar. However, at longer times, the OCP for 1600 rpm was lower while it remained similar for 2000 and 2400 rpm. Analyses of the products reveal that the major C2 constituents obtained upon oxidation of ethylene glycol are glycolic acid and glycolaldehyde. Other than these, glyoxylic acid and glyoxal were obtained as minor products.
Figure 2 shows the variation of the concentrations of these products with time as measured by HPLC. As expected with passage of time, the concentration of the both the major and minor products increased as shown in Figure 2a,b, respectively. However, a subtle difference was noted for the major and minor products. While, the concentration profile of the major products were invariant with respect to the rotational rate shown in Figure 2a, an increase in the concentration of glyoxylic acid and a decrease in the concentration of glyoxal with time on increasing the rotation rate were observed from Figure 2b. This suggests that the reactions that result in these minor C2 products are mass transfer controlled and probably glyoxal is a precursor to glyoxylic acid. Of course, without identification of reaction pathways and modeling the reaction rates in detail, the reasons can only be speculated. Other than these four fractions, formic acid (C1 product) was also identified at 1600 and 2000 rpm but no clear trend in the concentration was observed (see Supporting Information S1). Curiously, at higher rotation rate of 2400 rpm, no formic acid was detected.
2.2 Electrochemical catalysis experiment
As earlier mentioned, the electrochemical experiments were conducted at near identical conditions except the solution was Ar saturated and an Au counter electrode was used. It may also be re-emphasized that a combination of several LSVs was used to mimic the variation of the OCP with respect to time (as obtained during chemical catalysis experiments). The sample was collected at the end of every 30 min and analyzed by HPLC. Similar to the products obtained for the electrochemical catalysis experiments, two major products (glycolaldehyde and glycolic acid) and two minor products (glyoxal and glyoxylic acid) were obtained as shown in Figure 3a,b. Furthermore, the variation in concentration of different products with time also echoed that obtained for the chemical experiments. Indeed, it would be instructive to compare the concentrations for individual products more closely.
2.3 Comparison of chemical and electrochemical experiments results
As pointed out earlier, one of the key aspects of the investigations is to ascertain if the products obtained for chemical oxidation in presence of oxygen and electrochemical oxidation in absence of oxygen but application of potential are identical. This section compares the concentration of various products obtained from catalytic oxidation of ethylene glycol reaction via chemical and electrochemical pathway for both the major and minor products. Indeed, as can be seen in Figure 4a,b, the concentrations of the major products (glycolaldehyde and glycolic acid) as well as the minor products (glyoxylic acid and glyoxal) are nearly the same for every rotational rate (comparison of formic acid is shown in Supporting Information S2). The excellent match obtained corroborates the contention that even in catalytic chemical oxidation of ethylene glycol, electrons are released which are accepted by molecular oxygen which gets reduced to water. This broadly reinforces the assertions made by us earlier for a very simple system involving oxidation of ferrous sulfate by molecular oxygen.[31, 32]
Nevertheless, it still needs to be investigated if the phenomenon is unique to Pt. The following section provides a comparison of the products obtained on the two catalysts (Pt-Ir and Pt-Ru) at a rotational rate of 1600 rpm.
2.4 Influence of catalyst on electrochemical and chemical oxidation of ethylene glycol
The experimental procedure followed for both the Pt-Ir and Pt-Ru remained identical to that employed for Pt (see supporting information S3 for the variation of OCP with respect to time). For both these catalysts, the major products were glycolaldehyde as shown in Figure 5a and glycolic acid as shown in Figure 5b, and glyoxylic acid and glyoxal as minor products (see Supporting Information S4). Figure 5a,b shows the comparison of the major products obtained for both the chemical and electrochemical oxidation at 1600 rpm on Pt, Pt-Ir, and Pt-Ru catalyst. It can be clearly seen that as far as the concentrations of the products are concerned, the match between the chemical oxidation experiments and the electrochemical oxidation experiments are fairly good. This further proves that the phenomenon is general and explains the results observed Zope et al.[33] It may be pointed out that after nearly 3 h a sharp rise in the OCP for Pt-Ru was observed that manifested as a plateau in the concentration versus time graph. The reason presently are not clear and no efforts were made to investigate it, though we suspect it may be due to dissolution of Ru.
3 CONCLUSION
The present experiments demonstrate that underlying chemical oxidation, there exist an electron transfer mechanism. As an example, experiments were conducted both on chemical oxidation of ethylene glycol by molecular oxygen and electrochemical oxidation in absence of oxygen but by application of potential. Not only the chemical identity of the products obtained by both the chemical oxidation of ethylene glycol and the electrochemical oxidation are same; but also, their concentration trends were in excellent agreement. This has been verified for platinum at three rotational rates and for two other catalysts (Pt-Ir and Pt-Ru) at a single rotation rate.
4 EXPERIMENTAL SECTION
4.1 Chemicals
Ethylene glycol (assay ≥ 99.8%, Sigma–Aldrich, USA) was used as reactant. Sulfuric acid (H2SO4, Emsure grade, assay > 95–97%, Merck, Germany) was used both as supporting electrolyte and the mobile phase for HPLC. Unsupported platinum black (Fuel cell store; USA), platinum ruthenium (50:50 at%, Fuel cell store, USA), platinum iridium (50:50 at%, Fuel Cell Store; USA) were used as catalyst. Nafion (5 wt%, Sigma–Aldrich) was used for making all the catalyst dispersions. High purity argon gas was obtained from Mars Gas of India. Deionized water (18.2 MΩ.cm; Millipore, USA) was used for preparing aqueous solutions as well as mobile phase for HPLC. Expected products of partial oxidation of ethylene glycol such as glyoxylic acid, glyoxal, glycolaldehyde, glycolic acid, and formic acid (Sigma–Aldrich, USA) were used to generate the calibration curve by HPLC and later used for quantification.
4.2 Electrode preparation and instrumentation
A standard three electrode cell setup along with a computer connected GAMRY Interface 1000 potentiostat (USA) was used to carry out all the experiments. Catalyst coated glassy carbon electrode (GCE) was used as working electrode (WE). To prepare the catalyst ink, 5 mg of 100 wt% Pt black was taken in 1 ml of deionized water along with 20 μL Nafion solution. The mixture was then sonicated for 1 h to obtain a homogeneous black ink. Before drop casting the ink on the GC (5 mm diameter, Pine instruments, USA) electrode, it was polished to a mirror finish with successive grades of 1, 0.3, 0.05 μm alumina (Buehler, USA). Afterward, 120 μl of the catalyst dispersion was pipetted onto the polished GCE and kept under an infrared lamp to evaporate the solvent. The catalyst loading was 3057 μg/cm2 Pt, based on the electrode area. Finally, the WE was fixed on the Pine rotor (Model AFMSRCE RDE, Pine instruments). The Hg/Hg2SO4 (Sensor technologies, India) was used as a reference electrode (RE). The RE was calibrated against an in-house fabricated reversible hydrogen electrode (RHE) in 0.5 M H2SO4 and the potential of the Hg/Hg2SO4 electrode was measured to be 0.690 V. In this manuscript, all the potential values mentioned are against the RHE. Platinum mesh (99.9%, 100 mesh gauge (2 cm × 2 cm nominal)) was used as a counter electrode (CE) for the experiments involving cleaning the catalysts electrochemically by cyclic voltammetry. Other than these, a high purity gold foil (0.25 mm thick, 2 cm × 2 cm) was used. The cleaning of the catalysts was performed for all the experiments by cycling the potential of the drop caste catalysts in Ar saturated 0.5 M H2SO4 at 50 mV/s between 0.025 V to 1.2 V (both vs RHE) for 20 cycles. A check was performed vis-a-vis the electrochemical surface area of the catalysts and the reproducibility of the cyclic voltammograms (see Supporting Information S5). The average surface area of the drop cast Pt nanoparticles was found to be 24 m2/g, which is in good agreement with reported literature. This is also important that before start of any experiment, all the glassware was washed with aqua regia and then rinsed with deionized water to avoid any contamination and impurities.
4.3 Methodology
In the present work, the experiments were separated into two major parts (i) chemical experiment where the catalytic oxidation of ethylene glycol was performed in presence of oxygen and the open circuit potential (OCP) was measured with time, and (ii) electrochemical experiments, where ethylene glycol were oxidized electrochemically in Ar saturated solution through an extended linear sweep voltammetry (LSV) technique matching the variation of the OCP with time obtained during the chemical oxidation experiments. For both the chemical and electrochemical experiments, the samples were collected at different time period and analyzed by HPLC.
4.4 Chemical catalysis experiment
The chemical catalysis experiments were performed in the electrochemical cell as shown in Supporting Information S6. The working electrode was the thin film type that was prepared exactly as described earlier in electrode preparation and instrumentation section. The platinum loading was kept the same. The reaction cell containing 60 ml of 0.1 M ethylene glycol in 0.5 M H2SO4 solution was initially saturated with oxygen for an hour, afterward the working electrode was immersed in the solution. The rotational rate of the working electrode was set at a particular value viz 1600, 2000, 2400 rpm. For these chemical oxidation experiments, oxygen was bubbled continuously for the whole duration of the experiment (8 h). The potentiostat was programmed for open circuit potential (OCP) technique where the working electrode potential was measured against the reference electrode. The counter electrode was absent to exclude any possibility of electron transfer and thus the assembly constitute a batch reactor. In order to measure the oxidized products and their concentration, 100 μl of the solutions were collected at different time intervals and analyzed by HPLC.
4.5 Electrochemical catalysis experiment
Meanwhile, for the electrochemical experiment, the potentiostat was programmed to follow the potential versus time profile obtained during the chemical oxidation experiments. This was achieved by fragmenting the OCP versus time data (chemical oxidation experiments) in small time intervals (1 h) where the potential can be approximated to be varying linearly with time. Obviously, for these experiments, the technique used was linear sweep voltammetry where the scan rate was decided based on the slope of OCP versus time data for each fragment. The procedure was repeated for each catalysts and rotational rates. Furthermore, the experimental setup including the volume of solution remained same except (i) the solution was Ar purged and (ii) a gold counter electrode was used. For the experimental setup, see Supporting Information S7.
4.6 Product analysis
HPLC (1200 infinity series, Agilent technologies, USA) was used to analyze the different intermediate products/fractions and their concentration. The mobile phase was 13 mM H2SO4 and the flow rate was kept at 0.3 mL/min. The pH of the samples was maintained at 3.0. Aminex HPX – 87H (Biorad) column was used for separation of various fractions. Ultraviolet (UV) along with refractive index (RI) detectors were used for product identification and the respective concentrations. For precise quantification, calibration curves were obtained for possible different products along with the ethylene glycol (EG) at different concentration.
ACKNOWLEDGMENTS
Financial support by Department of Science and Technology Government of India, (SB/S3/CE/037/2013) and Ramanujan Fellowship is gratefully acknowledged.
CONFLICT OF INTEREST
The authors declare no conflict of interest.
Open Research
DATA AVAILABILITY STATEMENT
Data included in the article supporting information, or on request.
