Skip to main content
NCBI home page
JACS Au logoLink to JACS Au
. 2025 Oct 23;5(11):5665–5675. doi: 10.1021/jacsau.5c01157

Hydrophilic or Hydrophobic? How Byproducts Change the Water Affinity of Fischer–Tropsch Catalysts

Sanghamitra Sengupta †, Alexander J Klaver †, Qingyuan Zheng ‡, Marco de Jong †, Robert LC Voeten †, Lynn F Gladden ‡, G Leendert Bezemer †,*
PMCID: PMC12648334  PMID: 41311936

Abstract

The impact of reaction byproducts on the water wettability of Fischer–Tropsch catalysts has been investigated by using supported cobalt and ruthenium catalysts. Water contact angle measurements showed an evolution from 75° to 140° for a supported cobalt on titania catalyst due to catalytic operation, indicating that the catalyst changed from hydrophilic to almost superhydrophobic. MALDI-FT-ICR-MS studies revealed that the dewaxed spent catalyst contained long-chain carboxylates, in line with the presence of carboxylic acids with carbon numbers up to C100 in the heavy wax. Using dedicated experiments with alumina, silica, and titania support materials, it was found that alcohols, carboxylic acids, and amines adhered so strongly to the surface that they could not be removed by Soxhlet extraction with xylene. The resulting surface loadings varied from 0.3 to 3 molecules per square nanometer. For amines and carboxylic acids on alumina and titania, it resulted in a similar increase in water contact angle as observed for the spent catalyst. The molecular fingerprint of the organic-support interaction was obtained with ATR-IR, and using TGA data, it was shown that the presence of only 1–2 wt % of carboxylates on titania was sufficient to halve the water uptake capacity. Finally, operando NMR data using a Ru/TiO2 catalyst provided evidence that during the first weeks of catalytic operation, oxygenates build up with a concomitant decrease in liquid water on the catalyst surface. The profound influence of minor amounts of adsorbed products is of strong relevance for both catalyst design and catalytic evaluation.

Keywords: hydrophobicity, Fischer−Tropsch, operando, magnetic resonance, FT-ICR-MS, ATR-IR, contact angle

graphic file with name au5c01157_0009.jpg

graphic file with name au5c01157_0007.jpg

Introduction

Fischer–Tropsch synthesis (FTS) involves the catalytic conversion of CO and H2 into hydrocarbons and water over supported catalysts. Clean chemicals and transportation fuels can be made from noncrude oil feedstocks using this process. Shell started research on FTS more than 50 years ago and operates the world’s largest gas-to-liquids (GTL) plant in Qatar. However, the technology can also be used for the production of sustainable aviation fuel from renewable feedstocks involving green hydrogen combined with biomass (BTL) or CO2 (PTL) , In both GTL and BTL/PTL schemes, cobalt-based catalysts are preferred over iron catalysts due to their high liquid selectivity , and the low production of CO2.

During FT catalysis using cobalt catalysts, water production is inevitable. Water is both in weight and as a molar fraction (far) more abundant than the hydrocarbon products. Unfortunately, the presence of water is associated with higher production of undesired CO2 and catalyst deactivation. Key deactivation mechanisms like carbide formation, oxidation, sintering, and the formation of irreducible metal–support compounds can be accelerated by water. , It has been established in our earlier work that catalyst deactivation proceeds faster at high water partial pressures. , Hence, a mitigation strategy to reduce deactivation could be to limit the catalytic conversion. Commercially, this may not be attractive, and hence, alternative directions have been to add promoters or influence the catalyst’s susceptibility to water. Pioneering work was done by Sun and Rytter, who showed the impact of silylation on alumina and silica-supported catalysts. , Recent studies suggest that hydrophobicity can also be introduced in the bed through the loading of comaterials and that there can be benefits of confining the active phase in hydrophobic capsules to impact the catalytic performance of FT catalysts. − In another important development, it was shown that the water diffusion over a hydrophobic catalyst was unidirectional, whereas it was bidirectional over hydrophilic catalysts. Together, this shows the high relevance of changes from hydrophilic to hydrophobic character for the FT process. ,

Although the on-purpose modification of the catalyst’s affinity to water has shown a lot of promise, the available literature has not addressed whether the properties of the catalyst change after reduction and during catalytic testing. Especially during catalytic operation, the potential interaction of products with the catalyst needs attention. Besides olefins and paraffins as the main products, the FT reaction also produces a range of oxygenates, which could potentially interact with the catalyst support. Although the presence of oxygenates on spent catalysts has been reported earlier, , the understanding of the type and strength of interactions between FT byproducts and catalyst supports is quite limited. Hence, in this work, we investigate how the water affinity of Co/TiO2 catalysts changes during catalytic operation and perform an in-depth analysis of the distribution of byproducts, which is then complemented with matrix-assisted laser desorption/ionization Fourier transform ion cyclotron resonance mass spectrometry (MALDI FT-ICR-MS) on dewaxed spent catalysts. Next, we investigated a few common FT support materialstitania, alumina, and silicaand impregnated them with paraffins, alcohols, amines, and carboxylic acids. To obtain molecular understanding, attenuated total reflection infrared spectroscopy (ATR-IR) was used and combined with elemental analysis. For a macroscopic understanding, we deployed water contact angle (CA) measurements and combined them with gas-phase water adsorption measurements. Finally, operando nuclear magnetic resonance (NMR) studies were done to measure the buildup of oxygenates and the resulting impact on water contents during catalyst testing.

Experimental Methods and Materials

Materials

Catalyst support materials involved titania (TiO2), alumina (Al2O3), and silica (SiO2) extrudates. Additional information is provided in the Supporting Information. The solvents used, along with their respective supplier and purity/grade, are tetrahydrofuran (Superclo, MQ-100), n-heptane (Superclo, ≥99.3%), toluene (Emsure, ≥99%), m-xylene (Thermo Scientific, ≥99%), n-nonane (Sigma-Aldrich, MQ-200), and acetone (VWR Chemicals, ≥99%). All solvents were used as received without further purification. The SX-70 wax used in this work is a commercial product from Shell MDS Bintulu. The acids (hexanoic acid (C6), lauric acid (C12), and stearic acid (C18)) used were all purchased from Sigma-Aldrich as synthesis-grade chemicals. The alcohols involved 1-hexanol (Sigma-Aldrich, MQ-200) and 1-nonanol (Sigma-Aldrich, ≥98%). Hexylamine was acquired from Sigma-Aldrich (for synthesis, MQ-200).

Below, the most important procedures are discussed, while the reader is referred to the Supporting Information section for additional details.

Support Impregnation Method

For impregnation, 15 g of SX-70 wax or solid acids (i.e., lauric acid (C12) and stearic acid (C18)) was melted in an oven at temperatures 5–10 °C above the melting point, poured over 3 g of support material, and put back in the oven for 1 h. Subsequently, the excess liquid was removed, and the support was transferred to a bowl with a tissue and filter paper. Finally, it was placed back in the oven to soak up the remaining wax on the outside of the support (Figure S1A for images of the process). After removing the support from the oven, it was weighed again. Liquid impregnation was done using hexanoic acid, hexylamine, and 1-hexanol. Similar to the wax impregnation methodology, 3 g of support and 15 g of liquid were weighed, and the liquid was poured over the support and left in the fume hood for 1 h. Other steps to remove the excess liquid were done as described above.

Soxhlet Extraction Method

Soxhlet extraction was used to empty the pores of species without a strong interaction with the support. Solvents with different boiling points and solvency were used: n-pentane, n-heptane, n-nonane, tetrahydrofuran, toluene, m-xylene, and water. Most experiments used a combination of 4 h Soxhlet extraction with m-xylene, followed by 2 h Soxhlet extraction with n-pentane and 2 h drying in an oven at 70 °C. A detailed description of the procedure is given in SI-3.

Catalyst Preparation and Testing

A 20 wt % Co/TiO2 (P25) catalyst, referred to as Co20/TiO2, was prepared following the literature procedure and was reduced and subsequently passivated. Catalytic performance was evaluated in a fixed-bed unit where 10 g of catalyst material was diluted with 17 g of SiC 0.2 mm. After an in situ reduction treatment, the Fischer–Tropsch experiments were conducted at 210 °C, 30 bar syngas pressure with H2/CO = 1.2 and 45% CO conversion. Analysis of water, light, and heavy hydrocarbons was carried out according to protocols reported earlier.

ATR-IR Experiments

Infrared spectroscopy spectra were acquired using an attenuated total reflectance infrared (ATR-IR) spectrometer from Bruker (Pike MIRacle universal ATR Tensor-II). Extrudates were measured between 600 and 4000 cm–1 with 4 cm–1 resolution. However, for Al2O3 and TiO2, a strong background below 1000 cm–1 was seen, obscuring any features below 1000 cm–1. Hence, spectra between 1000 and 4000 cm–1 were reported. One to multiple extrudates were used for each measurement, depending on the support used. Each sample was measured at least in triplicate and averaged prior to processing.

CA

Sample preparation for CA measurements is discussed in detail in SI 5. The slide containing the catalyst was placed under a drop shape analyzer (DSA25 from Kruss) with a needle filled with distilled water. A ∼30 μL drop of water was dropped onto the surface, and the CA was calculated. Reported values are averaged from three measurements.

Temperature-Programmed Studies

TPR studies were done using a Netzch STA 449F3 Jupiter thermal analyzer. Approximately 40 mg of sample was loaded into an 85 μL aluminum crucible and heated at 2 °C/min under a total gas flow of 70 mL/min (5% H2/Ar). The TGA signal was corrected with a blank run using an inert powder in the same gas atmosphere. Gases evolved during the reaction were monitored online with a QMS 403D Aeolos mass spectrometer.

Water uptake measurements were executed on a TGA-550 from TA Instruments, which was customized to carry saturated vapors using a bubbler/condenser type setup. Sample intake was 65–75 mg with a 60 mL/min gas flow. After initial heating to 120 °C for 1 h under dry nitrogen gas flow, the sample was cooled to 60 °C, where the flow was switched to 1.1 vol % water. After stabilization for an hour, the temperature was lowered to subsequently 40 and 25 °C, and mass uptake due to water adsorption was determined. Finally, the flow was switched back to dry conditions, and the sample was heated again to 120 °C.

FT-ICR-MS

The measurements were performed based on the solvent-free preparation developed by Trimpin et al. The heavy wax and catalyst samples were ground to a fine powder using a mortar and pestle as previously reported. A small scoop of the wax was directly applied and smeared on an MTP 384 target plate. A piece of double-sided carbon tape was taped on the plate, on which a small scoop of ground catalyst sample was smeared. The ‘super sponge’ 1,8-bis­(tetramethylguanidino)­naphthalene (TMGN) was used for negative mode analysis, ensuring deprotonation. Approximately 15 mg of TMGN was dissolved in 1 mL of tetrahydrofuran (THF), yielding a 15 mg/mL TMGN solution. Four μL of TMGN solution was pipetted per area, equaling a single MTP-plate spot size. Measurements were conducted with a 12T Bruker solariX XR FT-ICR-MS instrument (Bruker Daltonics) equipped with a ParaCell and a MALDI source. Data collection was conducted over a 1 min period. The FT-ICR-MS spectra were processed using Data Analysis 5.3 software (Bruker Daltonics), exported as csv files, and imported into Sierra Analytics’ Composer software, which is well accepted in the FT-ICR community for molecular formula annotation. The annotations were transferred to Excel for the visualization and analysis of carboxylate-acid distributions by plotting relative intensities or concentrations against carbon numbers.

Carbon Content

The carbon content was measured with an Eltra CS2000 instrument at 1350 °C with IR detection. During the analysis, pure oxygen (99.999% pure) is used for combustion, and a calibrated IR detector is used to quantify the amount of CO2 generated. Reported results are the average of three measurements.

Operando NMR Measurements

A 1 wt % Ru/TiO2 catalyst was used for the reaction. The catalyst is a cylindrical extrudate 1 mm in diameter and 5 mm in length. The BET surface area, BJH average pore diameter, and pore volume of the catalyst were measured by Ar sorption as 52.3 m2 g–1, 27.5 nm, and 0.38 cm3 g–1, respectively. The metal surface area, particle size, and dispersion of the catalyst were measured as 0.43 m2 g–1, 11.3 nm, and 7.9%, respectively, using oxygen chemisorption. The FT reaction was carried out in a tubular fixed-bed reactor, of which the detailed description and the loading procedure have been reported elsewhere. Before the reaction, the catalyst (3.3 g) was reduced in situ at 250 °C and 6 bar for 40 h using a 4.8 vol % H2/N2 gas mixture. The reaction was operated at 220 °C and 37 bar with a feed H2/CO ratio of 2 and 17 vol % N2. The total gas flow rate was 10.8 NL h–1, resulting in a weight hourly space velocity (WHSV) of 4 NL h–1 gcat –1. Details of the reaction operation were reported earlier. The effluent gas composition was analyzed to an accuracy of ± 0.5 mol % using an online gas chromatograph (GC) (Agilent Refinery Gas Analyzer, Fast RGA 7890B-0378) equipped with two thermal conductivity detectors and one flame ionization detector. The operando NMR measurements were performed on a Bruker Avance III HD spectrometer. The instrumental details, experimental parameters used in this work, and methods for data analysis are discussed in SI 6.

Experimental Results

The water CA of Co20/TiO2 was measured at various stages of its catalyst life. In Figure A, a water droplet placed on a glass slide covered with fresh catalyst is shown. The calculated CA is found to be 75 ± 2°, making it hydrophilic. This is a lower angle than that of pure titania reported in literature and also experimentally obtained in this article, shown in Figure . This showcases the role of cobalt oxide in the catalyst wettability. Results of Co20/TiO2 after reduction and passivation showed a higher CA at 103 ± 4° (Figure B), which could point to loss of titanol (Ti–O–H) groups during heating in hydrogen. The catalyst, after catalytic testing (termed as ‘spent’ in the article text), followed by Soxhlet extraction, showed distinctively different characteristics. The CA increased to 140 ± 4°, showing significant water repulsion approaching superhydrophobic behavior.

1.

1

Open in a new tab

(A) Droplet of water is placed on glass slides covered with (A) Co20/TiO2 as prepared, (B) Co20/TiO2 as reduced, and (C) Co20/TiO2 spent after Soxhlet extraction. The calculated average CA of these three systems is 75 ± 2° (A), 103 ± 4° (B), and 140 ± 4° (C), respectively.

4.

4

Open in a new tab

(A) Surface loading and (B) water contact angle of titania, alumina, and silica modified with C6 alcohol, acid, and amine. Impact of acid chain length (C6, C12, C18) on surface loading (C) and water contact angle (D). Surface loading of acids on alumina is shown in Figure S14.

The catalytic test was performed in a fixed-bed reactor after reduction of the cobalt oxide phase with hydrogen. The performance was evaluated at a syngas pressure of 30 bar. To ensure sufficient catalyst and product slate stabilization (i.e., reaching a steady state), the experiment was run for 15 days in total. The catalyst activity was 0.16g·Lcat –1·h–1 with a selectivity to liquid products of 92 wt %. The gas, water, light wax, hydrocarbon, and oxygenate compositions were combined with the respective flow rates and reconciled. The resulting Anderson–Schultz–Flory (ASF) distributions of hydrocarbons, alcohols, acids, and aldehydes are plotted in Figure A. The hydrocarbon distribution shows the typical behavior of metallic cobalt with methane above and ethane/ethene below the line. Above C10, a slight downward bend can be observed in hydrocarbon ASF, which is attributed to an increasing fraction of molecules recovered in the heavy wax. The acid and alcohol product distribution shows higher selectivity toward C1 and C2 products but otherwise moves roughly parallel to the main hydrocarbon product, which reflects the similarity in chain growth probability. The carboxylic acid distribution also shows a deviation between C3 and C5, which is solely based on the lower concentration in water as obtained with liquid chromatography (LC). We checked with an alternative method (GC-MS using SIM of m/z 60) and found higher concentrations, and hence we attribute this deviation to an underestimation of the concentration by LC. Our observation that acids and alcohols have similar chain growth probability values compared to the hydrocarbon products agrees with work done by Wilson et al. using a Co/TiO2 system. Finally, aldehydes could be quantified in the water phase between acetaldehyde and butyraldehyde, while in the wax, only traces were seen.

2.

2

Open in a new tab

Hydrocarbon (black), alcohol (red), acid (blue), and aldehyde (orange) ASF distributions (A) from a combination of gas, light wax, and water compositions. The relative (B) and ASF (C) hydrocarbon (black) and acid (blue) distributions were determined by GC-FID and MALDI-FT-ICR-MS from the heavy wax sample. Carboxylate distributions were observed with MALDI-FT-ICR-MS on the spent catalyst (D) sample from the top (black), middle (red), and bottom (blue) bed.

The heavy wax was analyzed up to C100, and the results are provided in Figure B as the relative contribution and in Figure C as the ASF distribution. The hydrocarbon distribution (black) starts at C10 and increases toward around C25. This mirrors the light wax composition, and it reflects the gas–liquid equilibrium in the high-temperature separator as mentioned above. Part of the wax product had chain lengths beyond C100. In both the distribution and the ASF plot, a bend can be seen, which reflects the presence of multiple alpha values. Deviations from single alpha product distributions have been reported earlier and have been attributed to chain length-dependent olefin readsorption/desorption, gradients in process conditions, or superpositions of multiple product distributions. −

To optimize MALDI analysis in the negative ion mode to achieve detection of acids within the wax sample, a series of matrices were evaluated, including 9-aminoacridine (9-AA), 2,5-dihydroxybenzoic acid (DHB), α-cyano-4-hydroxycinnamic acid (CHCA), and also the strong organic bases 1,8-bis­(dimethylamino)­naphthalene (proton sponge) and TMGN. Among the tested matrices, only TMGN yielded satisfactory ionization efficiency, attributed to its high proton affinity and spectral quality in the present study. The observed acid distribution and corresponding ASF-plot are presented in Figure B,C (blue), respectively, and range from C12 up to C100, similar to the main hydrocarbon product. However, the increase above C15, up to the maximum observed at C30, deviates from that of the hydrocarbons, i.e., rises more slowly and declines faster above C30, revealing an overall narrower distribution for the acids. The lower measured contribution of acids up to C30 is revealed in the ASF-plot, but a similar alpha is observed for larger chains (i.e., C30 up to C70). The signal spread above C70 is attributed to the low intensity signals attained and, hence, deviates more abruptly.

The spent catalyst was also analyzed for residual carbon and hydrogen after the Soxhlet extraction of hydrocarbons in the catalyst pores. For the sample from the middle of the bed, this amounted to 4.3 wt % carbon and 0.7 wt % hydrogen and a molar H/C ratio of 2.0, which could point to the presence of carboxylic acids or olefins. In Figure D, results are shown from the spent catalyst samples that revealed the adsorbed carboxylate distribution. The carboxylate distribution is shown to be independent of the catalyst extrudate’s spatial location within the reactor bed (top, middle, or bottom section) as indicated by the overlapping distributions. The distributions span C15 up to as high as C70 with maximization around C22. Using the average chain length (C29) and the surface capacity of titania for acids (vide infra Figure A), a carbon content of 3 wt % due to acids is obtained, which shows that most of the residual carbon originates from carboxylic acids. The initial increase in relative intensity from C15 reflects the changing hydrocarbon distribution due to the increasing presence of molecules in the liquid phase. In contrast, the decline above C22 reflects the acid distribution observed from the wax samples. An analysis with thermal desorption (TD)-GC-MS was conducted and showed olefin, aldehyde, and ketone signals with only minor amounts of carboxylic acids (Figure S4). However, using a reference sample containing surface-modified lauric acid on a reduced and passivated Co/TiO2 sample, large signals of the decomposition products of the C12 acid were observed, and only a trace signal of the acid was found (Figure S3). The decomposition and conversion of carboxylates during heating are in line with literature data, but they limit the use of TD. , Another confirmation of the nature of the deposit on the catalyst was obtained by submerging a spent catalyst particle in a water mixture for a few days and subsequent analysis of the liquid phase with GC-MS. This experiment yielded only features related to acids and not aldehydes/ketones, in agreement with the results in Figure D obtained with MALDI. The corresponding ATR-IR spectra of the spent catalyst, showing the carboxylate bands, are shown in Figure S6, again confirming the MALDI findings.

From the previous results, it appeared that during the catalytic run, the water wettability of a catalyst changed from hydrophilic to hydrophobic. Moreover, in the Fischer–Tropsch product slate, byproducts were identified that could adhere to the catalyst surface, and from elemental analysis, it appeared that dewaxed spent catalyst still contained 5 wt % of organics, of which carboxylates were directly identified with MALDI-FT-ICR-MS. To test the hypothesis that byproducts could modify the water wettability of catalysts, three common base supports (alumina, silica, and titania) were investigated in the interaction with several byproducts. We selected primary alcohols, as these form the most prevalent byproduct class, primary carboxylic acids, because of their confirmed accumulation on spent catalysts and the proof that the product distribution extends into the heavy wax product. Furthermore, we selected amines, as our previous study showed the presence of amines in heavy wax beyond C100. To capture the molecular footprints on functionalized base supports, an impregnation step was performed with hexanol, hexylamine, or hexanoic acid. After a Soxhlet extraction to remove the material not strongly interacting with the support material, vibrational ATR-IR spectroscopy was deployed. To verify the Soxhlet extraction protocol for spent catalysts containing heavy wax, impregnations were performed with hydrogenated FT wax (SX-70) on the selected base supports. The ATR-IR spectra depicted in Figure S7 indicate that xylene is the most efficient solvent for wax removal, independent of the base supports. Consequently, xylene was selected for all subsequent Soxhlet extraction procedures throughout this study. The ATR-IR spectra of the FT wax, the bare crystal with the tip on top (without any sample), and the bare base supports are shown in Figures S8, S9, and S10, respectively.

In Figure , the three supports impregnated with hexanol, hexylamine, or hexanoic acid are shown before and after Soxhlet extraction. In terms of the vibrational signatures, first we discuss the common features for all three supports, subsequently the impregnated samples, and finally the samples after Soxhlet extraction. A clear signal of the C6 hydrocarbon tail is seen for all three functional groups. The bands around ∼2860 and 2930 cm–1 have contributions from both CH2 and CH3 symmetric stretches, and the band at 2960 cm–1 is assigned to the CH3 asymmetric stretch, assignments which are in line with previously published work for long hydrocarbons. The functional-group-specific signatures appear below 2000 cm–1, and the fingerprint region is quite crowded. For hexanoic acid after the impregnation (top rows Figure ), the free acid can be identified for all three supports at 1710 cm–1 (CO stretch), an assignment in agreement with previous literature findings. The O–H stretch associated with the free acid is identified as a broad band between 3000 and 3500 cm–1, whereas around 1630 cm–1, a weak bending mode is observed. For the samples impregnated with hexylamine (middle rows, Figure ), the signature C–N bands appear at 1570 cm–1 and, to a lesser extent, at 1470 cm–1. The band around 3315 cm–1 is attributed to an N–H stretch for the primary amine, which is strongest for titania and barely observed in the case of alumina. The assignments of the functional groups for amine impregnation are in line with previous work done with titania nanoparticles. To the best of our knowledge, this is the first report of amine impregnation within silica and alumina.

3.

3

Open in a new tab

Overview of ATR-IR spectra of the hexanoic acid, hexylamine, and hexanol impregnation of all three base supports TiO2 (A), SiO2 (B), and Al2O3 (C) before and after Soxhlet extraction with xylene.

In the samples impregnated with hexanol (bottom rows, Figure ), the broad O–H stretching peak around 3350 cm–1 is seen for all substrates, which can be assigned to the O–H stretching vibration attached to the substrate (both chemisorbed and physisorbed). The C–O stretching band appears around 1120 cm–1, which is very clear for both TiO2 and Al2O3. The band around 1060 cm–1 can potentially be assigned to the C–O stretch of the surface-anchored alkoxy species. However, for SiO2, the Si–O–Si band appears at the same frequency and is much stronger, which obscures the detection of the C–O band on silica.

The spectra after Soxhlet extraction (black lines in Figure ) show a clear reduction in vibrations assigned to the organics, but no full removal. Moreover, the support OH group density is reduced (Figure S11). This suggests that strongly bonded molecules are still present on these samples, for which the IR already provided evidence with the assignment of carboxylate and alkoxy moieties on the surface. We next moved on to quantify the amounts of residual carbon on the samples after the Soxhlet extraction. Indeed, on each of the supports, carbon was found and increased for hexanol and hexylamine in line with the support surface area from 1.0 to 3.1 wt % and from 1.2 to 6.6 wt %, respectively. With hexanoic acid, lower amounts were found for titania and silica at 0.7 and 1.5 wt %, whereas on alumina, a surprisingly high amount of 30 wt % was measured (Table S2). Such unusual changes specific to alumina are also visualized through transmission electron microscopy (TEM) images. The images show a clear contrast between the bare y-Al2O3 (Figure S15) and y-Al2O3 subjected to hexanoic acid impregnation and successive xylene Soxhlet extraction (Figure S16). The corresponding images and the description can be found in SI 17. A further investigation of the carbon content at various stages of the preparation is provided in Figure S12 for all three supports and chemicals. On average, Soxhlet extraction with xylene removes 85% of the chemicals, and a subsequent water Soxhlet extraction removes 98%. Contrary to this, the amount of residual carbon of hexanoic acid on alumina is not reduced after Soxhlet extraction in xylene and water. This, combined with the observation that the alumina had lost its structural integrity after contact with acid, points toward alumina degradation, which has been reported to take place below pH 4. Using the molar mass and the surface area from argon physisorption, the results, excluding the carboxylic acid on alumina, were converted to molecules/nm2 and are provided in Figure A. Among the three supports, a higher surface loading is obtained on titania (1.6 to 3.0) versus alumina (1.0–2.2) and silica (0.8–1.7 molecules/nm2). Between the different chemicals, an increasing trend is seen from acid, via alcohol, to amine.

The impact of adsorbed organics on water wettability was interrogated with CA measurements, and the results are presented in Figure B. The untreated alumina and silica supports were hydrophilic with CAs around 60 °, while titania was assessed as an intermediate between hydrophilic and hydrophobic properties. The samples with alcohols had similar or slightly lower water CAs. However, significant changes emerged for the amine- and acid-modified samples. With silica, the CA increased by around 10°, but it remained hydrophilic. In contrast, both titania and alumina exhibited pronounced hydrophobicity, with CAs ranging between 100° and 140°. Overall, the acid-treated samples showed the highest CAs, and hence, we moved on to investigate the impact of the carboxylic acid chain length. From the results presented in Figure D, a modest impact on wettability is seen for TiO2 with an increase in CA from 125° to 140° with no effective change for Al2O3 and SiO2. The CA measurements were further supported by determining the molecular surface concentration for the various supports, as shown in Figure C. Shorter-chain acids are present at higher surface concentrations, which may be attributed to steric effects or a greater propensity for the dissolution of longer acids during Soxhlet extraction. Notably, this variation does not appear to influence the overall degree of hydrophobicity.

Next, the impact of modification on water adsorption from both gas and liquid was investigated. This is critically relevant for FT synthesis, as water can play a key role in different deactivation mechanisms. The titania systems modified with acids were studied for water uptake at various temperatures with a controlled water atmosphere. The experimental procedure is shown in Figure A. Initially, the samples are dried at 120 °C, after which the temperature is reduced to 60 °C, where a gas stream containing 1.1% vol % water is passed over the sample. After weight stabilization, the temperature is reduced to 40 °C and finally to 25 °C. The experiment concludes with a final drying step. Mass developments for the nonmodified sample show a 0.4 wt % increase at 60 °C, increasing to 1.1 wt % at 25 °C. These weight increases are equivalent to a 0.3–0.7 ML water layer, respectively. The acid-modified samples showed a significantly lower water uptake of at most 0.6 wt %. The weight increases versus the relative humidity are also plotted in Figure B, showing an increasingly lower water uptake of the acid-modified samples at higher relative humidity levels. Combined with the water CA results, these findings do indicate that the titania surface properties regarding water wettability and water uptake are modified due to the presence of small amounts of strongly bonded carboxylic acids. Both techniques show a small outperformance of lauric and stearic acids compared to hexanoic acid, highlighting the importance of chain length in this regard.

5.

5

Open in a new tab

(A) Normalized weight and temperature over time during water adsorption and desorption measurements on titania and modified titania. During treatment at 120 °C dry gas was used, whereas uptake measurements were done with a 1.1 vol % water stream. (B) Water uptake changes as a function of the relative humidity.

Until now, it has been established that after catalytic operation, FT catalysts have become hydrophobic and that long carboxylates have accumulated within the catalyst pores. Moreover, it was shown that strong adsorption of various organics on catalyst supports can impact water wettability in the same way as on the unloaded spent catalyst (Figure ). Next, making use of operando NMR spectroscopy, it will be shown that oxygenate buildup and its impact on water adsorption can be monitored in time on a working Ru/TiO2 catalyst. Please note that the choice of the metal is related to the magnetic properties of cobalt, which interfere with the NMR spectroscopy. The elementary steps on cobalt and ruthenium surfaces are the same, allowing direct comparison between the two systems. Figure A shows the 2D δ-T 1 spectrum acquired at the reaction steady state at time on stream (TOS) of 510 h, and the corresponding 1D T 1 distribution is shown in Figure S2a. Three components are identified in the spectrum. The main component with the highest intensity is located at a chemical shift of 1.27 ppm and has an average NMR T 1 relaxation time of 4 s. This component is identified as a hydrocarbon species. , The component located at a similar chemical shift but having an average T 1 value of 0.1–1 s is associated with hydrocarbon-containing species, which have a stronger affinity for the catalyst surface than hydrocarbons. On the basis of our previous work, these spectral signals are assigned to oxygenate species inside the catalyst pores (Figure S2b). , The lack of chemical resolution of the operando NMR measurement does not allow us to distinguish between different types of oxygenate species, such as alcohols or acids. The spent catalyst was analyzed, and long carboxylic acids were found (Figure S13), whereas there were no signals related to alcohols. A third component located at a chemical shift ∼5 ppm and at a T 1 range of 10–3–10–2 s is identified as the liquid phase water present on the catalyst pore surface. , The time evolution of the intensities of oxygenates and water during the reaction startup and at the steady state is presented in Figure B. It is observed that the intensity of oxygenates increases until it stabilizes at TOS ∼ 300 h, while the intensity of water shows an opposite trend and stabilizes at a similar TOS. It is seen that the water weight fraction roughly halves over time from 0.37 to 0.20 wt % (approximately 0.1 to 0.2 ML water) during 300 h.

6.

6

Open in a new tab

Operando NMR characterization of water and oxygenates in catalyst pores. (A) 2D δ-T1 spectrum acquired at the steady state of the FT reaction at a TOS = 510 h, where the signals associated with hydrocarbon, oxygenates, and water are identified. (B) Time evolution of the signal intensities of oxygenates and water inside the catalyst pores during the reaction startup. (C) Variation of the weight fraction of water in pores during the reaction startup. The data were obtained during FTS at 220 °C, 37 bar, H2/CO feed ratio = 2, X CO = 27%, and S C5+ = 93%.

Discussion

In this article, we demonstrated that the Fischer–Tropsch catalyst becomes hydrophobic during catalytic operation. This is not the consequence of the residual wax but rather the result of oxygenates in strong interaction with the support. Our detailed study at a molecular level using three common catalyst support materials revealed which functional moieties enable such drastic changes. Water CA measurements evidenced that the acid groups have the strongest affinity to the support surface. Among the other base supports, TiO2 showed the highest dependence on the acid chain length and wettability. The dependence of the chain length on the wettability also indicates that the hydrophobic interaction between the hydrocarbon tails impacts the surface properties. The analyses confirmed the presence of long-chain carboxylates on the used catalyst extrudates, which facilitates the transition from hydrophilic to hydrophobic properties. Some earlier reports have shown that spent FT catalysts can contain carboxylic acids up to C25, which might be related to the dissolution limitations of longer oxygenates during the extraction. In our work, dewaxed spent catalysts are directly analyzed with MALDI FT-ICR-MS and reveal carboxylates ranging from C10 up to C70. This technique was earlier limited in its applicability for analyzing such oxygenates on catalysts due to, among others, poor ionization efficiency. We have overcome this problem by using the TMGN matrix to facilitate laser energy adsorption and enable the desorption of analytes from the catalyst surface by disrupting surface interactions, which are especially important for these strongly adsorbed species. The finding that the adsorbing carboxylates are much longer than previously seen is relevant for the design and operation of FT catalysts. The remaining amount of carbon of the hexanoic acid-treated sample after xylene amounted to 3 wt % of the titania pore volume, whereas in the used Co20/TiO2 catalyst, 15% of the pore volume of the catalyst was taken up by the longer carboxylates. Especially in systems with smaller pores, diffusion constraints could arise due to the presence of carboxylates with dimensions of up to 10 nm.

The spectroscopic studies brought two main findings to this research: (i) Molecular-level evidence of tightly bound surface species for each type of base support impregnation. (ii) Soxhlet extraction with apolar solvents is only effective for the removal of molecules not surface-anchored to the support. An analysis of impregnated samples before Soxhlet extraction showed that, despite the same chain length, acid groups were more strongly retained on the surface than alcohols and amines. It is also clear that bonding is strongest between titania and the respective anchoring groups and that the hydroxyl group density is reduced after surface modification. These results are in line with CA data, which showed that the water CA is highest for acid-functionalized TiO2. With the set of acid-modified titania samples, we confirmed with water adsorption measurements the trend, as seen with water CA measurements, that the loading of 1 molecule/nm2 is sufficient to significantly alter the affinity for water. The limited yet consistent difference between C6 and C12/C18 acids signifies the importance of the hydrophobic tail.

The magnetic properties of cobalt cause a lack of quantitation in the analysis of 1H NMR spectra of species interacting with the metal sites, and therefore, the more expensive ruthenium was used. Both systems are, however, comparable in elementary steps, and the selection of a catalyst with relatively large ruthenium particle sizes of 11 nm ensured a comparable selectivity to liquid products as industrial cobalt catalysts (>90%). The NMR data connected critical insights between the findings from spent catalyst analysis, model studies, and catalytic operation. The increased oxygenate intensity with time-on-stream (Figure B) suggests the accumulation of oxygenate species in the catalyst pores during the first few weeks of operation. Our previous study has shown that long-chain products accumulate over a startup period of hundreds of hours inside the catalyst pores due to the slow diffusion of hydrocarbon species within the pore. Figure B shows that accumulation of oxygenates occurs on the same time scale as the decline of the water signal. This observation is consistent with the accumulated oxygenates, leading to desorption of surface-bound water. This surface modification occurs over ∼300 h for our system. This indicates that for the catalyst surface to be fully modified by oxygenates, sufficient reaction time is required to allow oxygenates to accumulate in the pores. A relevant avenue for future studies would be to assess how byproduct-induced hydrophobicity impacts the adsorption of reactants, intermediates, and products and how this, in turn, impacts the catalytic mechanism. A strong impact on the water gas selectivity has been seen by multiple groups for hydrophobic iron catalysts. Although on cobalt and ruthenium, far lower selectivity toward CO2 is obtained, directionally, an impact through hydrophobicity is expected. In our experiment with Co20/TiO2, the CO2 selectivity dropped more than 10% between day 2 and day 15, and for Ru/TiO2, the CO2 signals decreased in time, although the absolute low amounts were close to the GC detection limit of 0.5 mol %. These results suggest a potential impact of change toward hydrophobic character on the water–gas-shift selectivity, but this needs further substantiation. A more detailed follow-up study would be to obtain catalyst samples after various TOS, determine the oxygenate content and hydrophobicity, and retest them in a dedicated setup with steady-state transient kinetic analysis capability to determine the surface coverage of CO, H2, and intermediates while measuring the selectivity toward CO2 and the other products. Water concentration is not only relevant for the CO2 selectivity but also plays a key role in catalyst deactivation mechanisms, contributing to particle sintering, oxidation, carbide formation, and mixed oxide formation. The results presented are relevant in that context in three different ways. First, the adsorption of carboxylates switches the surface from hydrophilic to hydrophobic during catalytic operation. This change needs to be considered in studies where silylation and other methods are used to make catalysts hydrophobic prior to the start of the reaction. ,, The question remains, however, whether the impact of hydrophobicity pertains to catalyst activation and initial operation or extends throughout the catalytic lifespan, which can be multiple years in commercial operations. Second, operando characterization revealed that around 2 weeks of operation were required before the oxygenates and water levels stabilized. , This highlights the importance of allowing sufficient time to reach a steady state prior to assessing the catalyst stability and catalytic performance. Finally, the water content in the catalyst particles during catalytic operation decreased from 0.37 to 0.20 wt %. Especially with cobalt and iron catalysts, a water-rich environment can have an impact on the long-term catalytic performance. With the current developed tools, research can be guided toward developing improved start-up protocols that would minimize high water levels during early operation. Potential avenues can be to start with liquid-filled pores, lower the initial conversion rates, or incorporate hydrophobic agents.

Conclusions

We showed that FT catalysts change from hydrophilic to hydrophobic during catalytic operation under the FT process conditions due to the accumulation of oxygenate species at the support surface. The affinity for water of commonly used Fischer–Tropsch support materials was evaluated after surface modification. Alcohols, carboxylic acids, and amines all strongly adhere to the support materials, as they could not be removed by Soxhlet extraction using xylene and pentane. The resulting surface loadings varied from 0.3 to 3 molecules per square nanometer. The lowest surface loadings were seen with silica, and the highest were seen with titania. Both alumina and titania showed a strong increase in the water CA to around 140°. This indicates that even at low organic loading, a complete switch from hydrophilic to hydrophobic can be achieved. For titania, it was shown that the water uptake from the gas phase is reduced by the presence of minor amounts of carboxylic acids and depends on the hydrophobic tail length of the acid. This is particularly seen between the C6 and C12.

Analysis of a Co20/TiO2 catalyst showed that the water CA evolved from 75° in the freshly prepared state to as much as 140° for the dewaxed spent catalyst. It was shown with MALDI-FT-ICR-MS that the spent catalyst contained carboxylates with an average chain length of C29, in line with the residual carbon content of 4.3 wt %. A detailed product analysis of the gas phase, water, and light wax was performed. This was done to obtain integrated distributions of aldehydes, alcohols, carboxylic acids, and hydrocarbons. Our results suggest similar chain growth probabilities of alcohols and acids compared to the main hydrocarbons. Further evidence for the presence of heavy functionalized molecules was obtained by MALDI-FT-ICR-MS on heavy wax, showing carboxylic acids with carbon ranges from C12 to C100, matching the hydrocarbon distribution. Finally, it was shown with operando NMR data using a Ru/TiO2 catalyst that the buildup of oxygenates, which were separately proven to be carboxylates, takes around 300 h of operation to complete. In the same period, the signal of liquid water on the catalyst surface drops, suggesting that the accumulation of oxygenates results in the reduction of water, fully in line with observations of the various modified support materials.

The data and deduced inference can have a large consequence for the study on the impact of hydrophobicity in the field of catalysis, and especially Fischer–Tropsch Synthesis. Studies on the impact of hydrophobicity on FTS potentially could aim to evaluate hydrophobicity not only before but also during and after catalytic testing. Moreover, it shows that stabilization of various weeks is required to allow separation between the initial and long-term performance differences.

Supplementary Material

au5c01157_si_001.pdf (1.7MB, pdf)

Acknowledgments

The authors would like to acknowledge funding by Shell Global Solutions International B.V. to enable the internship and support for preparing this manuscript. We would also like to thank Ahmed Ali, Boaz de Letter, Alp Korkmaz, Sharon Stephen Garcia, and Lasse Bremmers for sample preparation and collecting CA data. We thank Abid Biran and Ravi Agrawal for collecting the water adsorption data, Aram Yoon for collecting the TEM images, and Sander van Bavel for reviewing the manuscript.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.5c01157.

  • Information on support materials, images of the stepwise impregnation method, Soxhlet extraction procedure, determining detailed product distributions, sample preparation methods for contact angle measurements, NMR experimental details and data analysis methods, TD-GC-MS results and instrument settings, ATR-IR spectra of the spent catalyst, ATR-IR study of impact of solvent on SX-70 wax removal, ATR-IR spectra of the SX-70 wax, ATR-IR spectra of the tip on the diamond crystal, full-scale ATR-IR spectra of the bare base supports (TiO2, SiO2, and Al2O3), semiquantitative comparison using ATR-IR on the reduction of O–H intensity for different base supports, carbon elemental composition comparison of different base supports upon impregnation, sample destruction and acid analysis after extraction, carbon elemental composition of base supports modified with various carboxylic acids, and TEM images of the bare alumina and the impregnated alumina after Soxhlet extraction (PDF)

The research work was planned by G.L.B. The manuscript was conceptualized and designed by G.L.B. and S.S. The figures are designed and created by G.L.B. and S.S. The samples investigated in this work are prepared mostly by A.J.K. R.L.C.V. performed all the MALDI experiments and data analysis. S.S. has done all the ATR-IR work. The CA measurements were done by A.J.K., and M.D.J. performed TD-GC-MS measurements. Q.Z. led the experimental design, the data acquisition and analysis of the operando NMR measurement, and wrote the relevant part of the manuscript. L.F.G. oversaw the project of operando NMR characterization of FTS. All authors participated in writing and discussion. CRediT: Sanghamitra Sengupta data curation, investigation, visualization, writing - original draft, writing - review & editing; Alexander J. Klaver data curation, investigation, writing - review & editing; Qingyuan Zheng data curation, formal analysis, investigation, methodology, writing - review & editing; Marco de Jong data curation, formal analysis, investigation, writing - original draft; Robert L. C. Voeten conceptualization, data curation, formal analysis, investigation, methodology, writing - original draft, writing - review & editing; Lynn F. Gladden conceptualization, investigation, methodology, supervision, writing - review & editing; Gerrit Leendert Bezemer conceptualization, formal analysis, funding acquisition, investigation, methodology, resources, supervision, writing - original draft, writing - review & editing.

The authors declare no competing financial interest.

References

  1. Dancuart, L. ; Steynberg, A. . Fischer–Tropsch based GTL technology: A new process? In Studies in surface science and catalysis; Elsevier, 2007, vol 163; pp 379–399. [Google Scholar]
  2. Martinelli M., Gnanamani M. K., LeViness S., Jacobs G., Shafer W. D.. An overview of Fischer–Tropsch Synthesis: XtL processes, catalysts and reactors. Applied Catalysis A: General. 2020;608:117740. doi: 10.1016/j.apcata.2020.117740. [DOI] [Google Scholar]
  3. Vogt E. T. C., Weckhuysen B. M.. The refinery of the future. Nature. 2024;629(8011):295–306. doi: 10.1038/s41586-024-07322-2. [DOI] [PubMed] [Google Scholar]
  4. van Bavel S., Verma S., Negro E., Bracht M.. Integrating CO2 electrolysis into the gas-to-liquids–power-to-liquids process. ACS Energy Letters. 2020;5(8):2597–2601. doi: 10.1021/acsenergylett.0c01418. [DOI] [Google Scholar]
  5. Teimouri Z., Abatzoglou N., Dalai A. K.. Kinetics and Selectivity Study of Fischer–Tropsch Synthesis to C5+ Hydrocarbons: A Review. Catalysts. 2021;11(3):330. doi: 10.3390/catal11030330. [DOI] [Google Scholar]
  6. Rytter E., Tsakoumis N. E., Holmen A.. On the selectivity to higher hydrocarbons in Co-based Fischer–Tropsch synthesis. Catal. Today. 2016;261:3–16. doi: 10.1016/j.cattod.2015.09.020. [DOI] [Google Scholar]
  7. Wolf M., Fischer N., Claeys M.. Water-induced deactivation of cobalt-based Fischer–Tropsch catalysts. Nature Catalysis. 2020;3(12):962–965. doi: 10.1038/s41929-020-00534-5. [DOI] [Google Scholar]
  8. van Koppen L. M., Iulian Dugulan A., Leendert Bezemer G., Hensen E. J. M.. Sintering and carbidization under simulated high conversion on a cobalt-based Fischer–Tropsch catalyst; manganese oxide as a structural promotor. J. Catal. 2022;413:106–118. doi: 10.1016/j.jcat.2022.06.020. [DOI] [Google Scholar]
  9. van Koppen L. M., Iulian Dugulan A., Leendert Bezemer G., Hensen E. J. M.. Elucidating deactivation of titania-supported cobalt Fischer–Tropsch catalysts under simulated high conversion conditions. J. Catal. 2023;420:44–57. doi: 10.1016/j.jcat.2023.02.019. [DOI] [Google Scholar]
  10. Bezemer G. L., Remans T. J., van Bavel A. P., Dugulan A. I.. Direct Evidence of Water-Assisted Sintering of Cobalt on Carbon Nanofiber Catalysts during Simulated Fischer–Tropsch Conditions Revealed with in Situ Mössbauer Spectroscopy. J. Am. Chem. Soc. 2010;132(25):8540–8541. doi: 10.1021/ja103002k. [DOI] [PubMed] [Google Scholar]
  11. Shi L., Li D., Hou B., Sun Y.. Organic Modification of SiO2 and Its Influence on the Properties of Co-Based Catalysts for Fischer–Tropsch Synthesis. Chinese Journal of Catalysis. 2007;28(11):999–1002. doi: 10.1016/S1872-2067(07)60084-9. [DOI] [Google Scholar]
  12. Rytter E., Salman A. U. R., Tsakoumis N. E., Myrstad R., Yang J., Lögdberg S., Holmen A., Ro̷nning M.. Hydrophobic catalyst support surfaces by silylation of Γ-alumina for Co/Re Fischer–Tropsch synthesis. Catal. Today. 2018;299:20–27. doi: 10.1016/j.cattod.2017.04.031. [DOI] [Google Scholar]
  13. Xu Y., Li X., Gao J., Wang J., Ma G., Wen X., Yang Y., Li Y., Ding M.. A hydrophobic FeMn@Si catalyst increases olefins from syngas by suppressing C1 by-products. Science. 2021;371(6529):610–613. doi: 10.1126/science.abb3649. [DOI] [PubMed] [Google Scholar]
  14. Fang W., Wang C., Liu Z., Wang L., Liu L., Li H., Xu S., Zheng A., Qin X., Liu L., Xiao F.-S.. Physical mixing of a catalyst and a hydrophobic polymer promotes CO hydrogenation through dehydration. Science. 2022;377(6604):406–410. doi: 10.1126/science.abo0356. [DOI] [PubMed] [Google Scholar]
  15. Xu Y., Zhang Z., Wu K., Wang J., Hou B., Shan R., Li L., Ding M.. Effects of surface hydrophobization on the phase evolution behavior of iron-based catalyst during Fischer–Tropsch synthesis. Nat. Commun. 2024;15(1):7099. doi: 10.1038/s41467-024-51472-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Pinard L., Bichon P., Popov A., Lemberton J. L., Canaff C., Maugé F., Bazin P. S., Aguiar E. F., Magnoux P.. Identification of the carbonaceous compounds present on a deactivated cobalt based Fischer–Trospch catalyst resistant to “rejuvenation treatment”. Appl. Catal., A. 2011;406(1):73–80. doi: 10.1016/j.apcata.2011.08.012. [DOI] [Google Scholar]
  17. Xu Y., Liang H., Li R., Zhang Z., Qin C., Xu D., Fan H., Hou B., Wang J., Gu X. K., Ding M.. Insights into the Diffusion Behaviors of Water over Hydrophilic/Hydrophobic Catalysts During the Conversion of Syngas to High-Quality Gasoline. Angew. Chem., Int. Ed. 2023;62(37):e202306786. doi: 10.1002/anie.202306786. [DOI] [PubMed] [Google Scholar]
  18. Javed M., Zhang G., Gao W., Cao Y., Dai P., Ji X., Lu C., Yang R., Xing C., Sun J.. From hydrophilic to hydrophobic: A promising approach to tackle high CO2 selectivity of Fe-based Fischer–Tropsch microcapsule catalysts. Catal. Today. 2019;330:39–45. doi: 10.1016/j.cattod.2018.08.010. [DOI] [Google Scholar]
  19. Peña D., Griboval-Constant A., Lancelot C., Quijada M., Visez N., Stéphan O., Lecocq V., Diehl F., Khodakov A.. 9. Catal. Today. 2014;228:65–76. doi: 10.1016/j.cattod.2013.10.005. [DOI] [Google Scholar]
  20. Bessell S.. Support effects in cobalt-based fischer-tropsch catalysis. Applied Catalysis A: General. 1993;96(2):253–268. doi: 10.1016/0926-860X(90)80014-6. [DOI] [Google Scholar]
  21. Goes, M. F. ; Rek, P. J. M. ; Schaddenhorst, D. ; Lange, J. P. ; Geerlings, J. J. C. ; Huisman, H. M. ; Oosterbeek, H. . Catalyst and process for the preparation of hydrocarbons U.S. Patent 5981608, 1999.
  22. Voeten R. L. C., Hendriks F., Bezemer G. L.. Fischer–Tropsch Synthesis for the Production of Sustainable Aviation Fuel: Formation of Tertiary Amines from Ammonia Contaminants. ACS Omega. 2024;9(29):31974–31985. doi: 10.1021/acsomega.4c03734. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Trimpin, S. Solvent-Free MALDI Sample Preparation. In Maldi Mass Spectrometry for Synthetic Polymer Analysis; Wiley, 2009; pp 159–186. [Google Scholar]
  24. Cao D., Hu M., Han C., Yu J., Cui L., Liu Y., Wang H., Cai Y., Kang Y., Zhou Y.. Proton sponge-functionalized silica as high performance adsorbents for solid-phase extraction of trace perfluoroalkyl sulfonates in the environmental water samples and their direct analysis by MALDI-TOF-MS. Analyst. 2012;137(9):2218–2225. doi: 10.1039/c2an16190g. [DOI] [PubMed] [Google Scholar]
  25. Pailler L., Renard P., Nicol E., Deguillaume L., Bianco A.. How Well Do We Handle the Sample Preparation, FT-ICR Mass Spectrometry Analysis, and Data Treatment of Atmospheric Waters? Molecules. 2022;27(22):7796. doi: 10.3390/molecules27227796. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Taylor K. C.. Determination of ruthenium surface areas by hydrogen and oxygen chemisorption. J. Catal. 1975;38(1):299–306. doi: 10.1016/0021-9517(75)90091-3. [DOI] [Google Scholar]
  27. Zheng Q., Williams J., van Thiel L. R., Elgersma S. V., Mantle M. D., Sederman A. J., Baart T. A., Bezemer G. L., Guédon C. M., Gladden L. F.. Operando magnetic resonance imaging of product distributions within the pores of catalyst pellets during Fischer–Tropsch synthesis. Nature Catalysis. 2023;6(2):185–195. doi: 10.1038/s41929-023-00913-8. [DOI] [Google Scholar]
  28. Cheever G. D., Ulicny U. C.. Interrelationships between pigment surface energies and pigment dispersions in polymer solutions. J. Coat. Technol. 1983;55:53–63. [Google Scholar]
  29. Partington R., Clarkson J., Paterson J., Sullivan K., Wilson J.. Quantitative carbon distribution analysis of hydrocarbons, alcohols and carboxylic acids in a Fischer–Tropsch product from a Co/TiO2 catalyst during gas phase pilot plant operation. J. Anal. Sci. Technol. 2020;11(1):42. doi: 10.1186/s40543-020-00235-5. [DOI] [Google Scholar]
  30. Todic B., Ma W., Jacobs G., Davis B. H., Bukur D. B.. CO-insertion mechanism based kinetic model of the Fischer–Tropsch synthesis reaction over Re-promoted Co catalyst. Catal. Today. 2014;228:32–39. doi: 10.1016/j.cattod.2013.08.008. [DOI] [Google Scholar]
  31. Donnelly T. J., Yates I. C., Satterfield C. N.. Analysis and prediction of product distributions of the Fischer–Tropsch synthesis. Energy Fuels. 1988;2(6):734–739. doi: 10.1021/ef00012a003. [DOI] [Google Scholar]
  32. Madon R. J., Iglesia E.. Hydrogen and CO intrapellet diffusion effects in ruthenium-catalyzed hydrocarbon synthesis. Journal of catalysis. 1994;149(2):428–437. doi: 10.1006/jcat.1994.1309. [DOI] [Google Scholar]
  33. Mansir N., Mohd Sidek H., Teo S. H., Mijan N.-A., Ghassan Alsultan A., Ng C. H., Shamsuddin M. R., Taufiq-Yap Y. H.. Catalytically active metal oxides studies for the conversion technology of carboxylic acids and bioresource based fatty acids to ketones: A review. Bioresource Technology Reports. 2022;17:100988. doi: 10.1016/j.biteb.2022.100988. [DOI] [Google Scholar]
  34. Kim K. S., Barteau M. A.. Pathways for carboxylic acid decomposition on titania. Langmuir. 1988;4(4):945–953. doi: 10.1021/la00082a028. [DOI] [Google Scholar]
  35. Sengupta S., Gera R., Egan C., Morzan U. N., Versluis J., Hassanali A., Bakker H. J.. Observation of Strong Synergy in the Interfacial Water Response of Binary Ionic and Nonionic Surfactant Mixtures. J. Phys. Chem. Lett. 2022;13(49):11391–11397. doi: 10.1021/acs.jpclett.2c02750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Nakayama N., Hayashi T.. Preparation of TiO2 nanoparticles surface-modified by both carboxylic acid and amine: Dispersibility and stabilization in organic solvents. Colloids Surf., A. 2008;317(1–3):543–550. doi: 10.1016/j.colsurfa.2007.11.036. [DOI] [Google Scholar]
  37. Jackson P., Parfitt G.. Infra-red study of the surface properties of rutile. Adsorption of ethanol, n-butanol and n-hexanol. Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases. 1972;68:1443–1450. doi: 10.1039/f19726801443. [DOI] [Google Scholar]
  38. Brownson J. R., Tejedor-Tejedor M. I., Anderson M. A.. FTIR spectroscopy of alcohol and formate interactions with mesoporous TiO2 surfaces. J. Phys. Chem. B. 2006;110(25):12494–12499. doi: 10.1021/jp0614547. [DOI] [PubMed] [Google Scholar]
  39. Averill, B. ; Moulijn, J. ; van Santen, R. ; van Leeuwen, P. . Catalysis: An integrated approach; Elsevier, 1999. [Google Scholar]
  40. Shetty S., van Santen R. A.. CO dissociation on Ru and Co surfaces: The initial step in the Fischer–Tropsch synthesis. Catal. Today. 2011;171(1):168–173. doi: 10.1016/j.cattod.2011.04.006. [DOI] [Google Scholar]
  41. Zheng Q., Williams J. H., Elgersma S. V., Mantle M. D., Sederman A. J., Bezemer G. L., Guédon C. M., Gladden L. F.. Operando characterisation of the products of Fischer–Tropsch synthesis in a fixed-bed reactor studied by magnetic resonance. Catal. Today. 2024;428:114416. doi: 10.1016/j.cattod.2023.114416. [DOI] [Google Scholar]
  42. Ward-Williams J. A., Karsten V., Guédon C. M., Baart T. A., Munnik P., Sederman A. J., Mantle M. D., Zheng Q., Gladden L. F.. Extending NMR Tortuosity Measurements to Paramagnetic Catalyst Materials Through the Use of Low Field NMR. Chem.–Methods. 2022;2(8):e202200025. doi: 10.1002/cmtd.202200025. [DOI] [Google Scholar]
  43. Wang C., Liu L., Li H., Wang L., Xiao F.-S.. Hydrophobic catalysts for syngas conversion. Matter. 2023;6(9):2697–2710. doi: 10.1016/j.matt.2023.05.010. [DOI] [Google Scholar]
  44. Gonugunta P., Dugulan A. I., Bezemer G. L., Brück E.. Role of surface carboxylate deposition on the deactivation of cobalt on titania Fischer–Tropsch catalysts. Catal. Today. 2021;369:144–149. doi: 10.1016/j.cattod.2020.04.037. [DOI] [Google Scholar]
  45. Kistamurthy D., Saib A. M., Moodley D. J., Preston H., Ciobîcă I. M., van Rensburg W. J., Niemantsverdriet J. W., Weststrate C. J.. The role of carboxylic acid in cobalt Fischer–Tropsch synthesis catalyst deactivation. Catal. Today. 2016;275:127–134. doi: 10.1016/j.cattod.2015.11.012. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

au5c01157_si_001.pdf (1.7MB, pdf)

Articles from JACS Au are provided here courtesy of American Chemical Society

RESOURCES