Platinized Titanium as Alternative Cost‐Effective Anode for Efficient Kolbe Electrolysis in Aqueous Electrolyte Solutions

Abstract

Five commercial materials were assessed for electrochemical conversion of n‐hexanoic acid by Kolbe electrolysis. Platinized titanium performed best, achieving a coulombic efficiency (CE) of 93.1±6.7 % (n=6) for the degradation of n‐hexanoic acid and 48.3±3.2 % (n=6) for the production of n‐decane, which is close to the performance of pure platinum (89.7±14.4 and 55.5±3.5 %; n=6). 56.7 mL liquid fuel was produced per mole n‐hexanoic acid, converting to an energy demand of 6.66 kWh and 1.22 € per L. Using optical profilometry and scanning electron microscopy coupled with energy‐dispersive X‐ray spectroscopy, it was shown that the degree of coverage of the titanium surface with platinum played the most important role. An uncovered surface of as little as 1–3 % already led to a deterioration of the CE of approximately 50 %. Using platinized titanium requires >36 times less capital expenditure at only <10 % increased operational expenditure; an electrode lifetime of 10000 h can be expected.

All that glitters is (not) platinum: Platinum‐covered titanium, available at scale and at only 3 % of the cost of Pt, can replace monolithic platinum as anode for the Kolbe electrolysis. Yet, this only holds true for material having a continuous and equal surface, as already small irritations lead to a dramatic performance decrease.

Introduction

The future of our economy has to be circular and based on renewable resources. Achieving sustainability for the material basis (e. g. the feedstock of carbon, nitrogen, etc.) is as important as it is for the energetic basis that is mainly electric power and heat.[¹, ²] Thereby interweaving of the different sectors of a bio‐based economy is of utmost importance to achieve the ambitious goals, for instance, formulated in the Green New Deal of the European Union.^[3] Hence, it is particularly important to link electric power production and storage with the chemical sector. Such links can be created by Power‐to‐Chemicals, Power‐to‐Fuels, or more generally Power‐to‐X that have gained special attention in the last decade.^[4] Power‐to‐X aims at utilizing (surplus) electric energy for electrochemical synthesis including electrochemical upgrading of compounds that underwent preceding process steps as well as the electrochemical supply of intermediates for follow‐up conversions. This allows combining electrochemical synthesis with chemical, physical, and biological process steps.[⁵, ⁶, ⁷] When electrochemical syntheses are integrated into process lines comprising microbial conversions in a Power‐to‐X approach, electrobiorefineries are set into place.^[8]

Among the plethora of electrochemical synthesis are, for instance, the chlorine‐alkali electrolysis,^[9] the conversion of muconic acid to diacid monomers,^[10] and the CO₂ reduction to formate, CO, or hydrocarbons.^[11] An electrochemical synthesis that was already discovered by Faraday in 1834^[12] is the Kolbe electrolysis. Of special recent interest is the Kolbe electrolysis of medium‐chain carboxylic acids (MCCA), which are carboxylic acids with a carbon chain length of 5 to 8 C‐atoms (C₅−C₈). Urban et al. showed the conversion of corn beer and corn silage to a hydrocarbon mixture with fuel properties using the Kolbe electrolysis.^[13] This was achieved by combining a biological step for converting the complex substrate into a MCCA mixture with the extraction of the MCCA and subsequent Kolbe electrolysis. The process line possessed an efficiency of 0.48 g chemical oxygen demand (COD) of hydrocarbons per 1 g COD of corn beer, where COD is a sum‐parameter of oxidizable carbon equivalents, and the Kolbe electrolysis showed a coulombic efficiency (CE) >80 % for MCCA oxidation.^[13]

During Kolbe electrolysis MCCAs are either converted into longer‐chain n‐alkanes (Kolbe products) via an intermediate radical step or a carbocation is formed from the radical intermediate, which further reacts to form n‐alkenes, alcohols, or esters (non‐Kolbe products, see Scheme 1). Thereby, the pathways, yields, and side reactions strongly depend on the reaction conditions.

Scheme 1.

Kolbe electrolysis of n‐hexanoic acid to the Kolbe product n‐decane and non‐Kolbe products such as alkenes, esters, and alcohols. For a detailed reaction pathway see Figure S1.

Detailed studies on the Kolbe electrolysis of MCCA are scarce (see Table 1), and different MCCA in aqueous solution have been used. Among others, the influence of the electrode material, the electrode potential, and the type and concentration of the supporting electrolyte were examined. For the latter, we previously demonstrated that Na₂SO₄, as supporting electrolyte different to KNO₃, has no effect on the CE of the Kolbe electrolysis of n‐valeric acid, but a higher a concentration of Na₂SO₄ increases the rate of acid degradation.^[14] Additionally, we also proved that due to local pH shifts n‐octanoic acid/n‐octanoate forms agglomerates in aqueous solutions during Kolbe electrolysis, leading to a deterioration of the electrolysis.^[15] In summary, supporting electrolyte species and concentration, pH, as well as the nature of the MCCA have an influence on the Kolbe electrolysis performance and therefore have to be well balanced. Concerning the electrode material most work was performed using (monolithic) pure platinum (Pt). Using pure Pt anodes, it is the current density (j) rather than the potential that determines the efficiency of the Kolbe electrolysis as shown in Table 1. To date only few other mostly self‐made electrode materials such as thin film Pt, RuO₂, IrO₂, or boron‐doped diamond have been examined.[¹⁶, ¹⁷] However, these are not commercially available, hardly scalable, and do not have a CE and selectivity comparable to that of pure Pt. Furthermore, most of the studies only investigate acid degradation and thus CE values are only reported for this process. However, when considering the diversity of products that can be gained (Scheme 1), this only provides an insufficient picture. Yet, comparable data like CE, yield, and selectivity for the production of Kolbe and specific non‐Kolbe products resulting from Kolbe electrolysis of MCCAs are almost completely missing.

Table 1.

Literature overview for the performance of Kolbe electrolysis with MCCA in aqueous media.^[a]

c Substrate	Anode	pH; supporting electrolyte	CEsubstrate [%]	CEdimer [%]	Yield [%]	Side products	r [mol cm−2 L−1 h−1]	Specifications	Ref.
0.5 m C5	Pt foil	pH 5 0.14 m Na2SO4	88.7	–	–	butene butanol butanoic acid esters aldehyde ethanol methanol acetic acid formic acid levulinic acid	$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0158	two‐chamber electrochemical cell galvanostatic (50 mA cm−2) 15–20 °C 1 h 40 mL	[16]
	RuO2 thin film on Ti foil		49.7	–	–		$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0095
	IrO2 thin film on Ti foil		7.1	–	–		$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0021
	Pt thin film on Ti foil		not stated; acid conversion <0.2 %	–	–		not active for decarboxylation
	Pt foil		91.2 (3 V) 85.3 (4 V) 83.5 (5 V)	–	–		$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e};3\mathrm{V}}:$ 0.0019 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e},4\mathrm{V}}:$ 0.0079 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e},5\mathrm{V}}:$ 0.0114	two‐chamber electrochemical cell potentiostatic (3.0, 4.0, 5.0 V vs. RHE) 15–20 °C 2 h 40 mL
	RuO2 thin film on Ti foil		46.1 (3 V) 56.0 (4 V) 39.4 (5 V)	–	–		$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e};3\mathrm{V}}:$ 0.0026 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e},4\mathrm{V}}:$ 0.0061 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e},5\mathrm{V}}:$ 0.0081
	IrO2 thin film on Ti foil		6.5 (3 V) 8.5 (4 V) 9.3 (5 V)	–	–		$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e};3\mathrm{V}}:$ 0.0006 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e},4\mathrm{V}}:$ 0.0012 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e},5\mathrm{V}}:$ 0.0021
	Pt thin film on Ti foil		not stated; acid conversion <0.2 %	–	–		not active for decarboxylation
0.5 or 1 m C5	pure Pt	pH 5.5 K2CO3	66.5	–	–	n‐butyl valerate n‐butanol	–	one‐chamber electrochemical cell potentiostatic (3.5 V vs. Ag/AgCl sat. KCl) 18 °C 100 mL	[19]
0.58 m C5	pure Pt	pH 5 0.0 m KNO3 0.10 m KNO3 0.20 m KNO3 0.99 m KNO3	(all n=3) 95.7±3.8 66.7±7.2 59.8±5.0 39.0±1.2	–	–	–	$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0056±0.0006 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0050±0.0009 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0057±0.0004 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0080±0.0009	one‐chamber electrochemical cell potentiostatic (3 V vs. Ag/AgCl sat. KCl) 2 h 200 mL	[14]
		pH 5 0.0 m Na2SO4 0.07 m Na2SO4 0.14 m Na2SO4 0.70 m Na2SO4	(all n=3) 93.1±9.5 92.0±7.3 91.6±10.3 92.0±1.1	–	–	–	$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0065±0.0005 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0082±0.0004 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0105±0.0009 $r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0161±0.0002
0.74–0.78 m C6	pure Pt	pH 6.1 1 m NaOH	–	–	$Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r};0.08\mathrm{A}{\mathrm{c}\mathrm{m}}^{-2}}$ : 0 $Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r};0.13\mathrm{A}{\mathrm{c}\mathrm{m}}^{-2}}$ : 24±3 $Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r};0.18\mathrm{A}{\mathrm{c}\mathrm{m}}^{-2}}$ : 45±5 $Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r};0.35\mathrm{A}{\mathrm{c}\mathrm{m}}^{-2}}$ : 40±5	–	–	one‐chamber electrochemical cell galvanostatic (0.08–0.35 A cm−2) 20±3 °C 20 mL sono‐emulsion (ultrasound)	[17]
0.77 m C7		pH 6.5 1 m NaOH	–	–	$Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r};0.18\mathrm{A}{\mathrm{c}\mathrm{m}}^{-2}}$ : 15±2 $Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r};0.35\mathrm{A}{\mathrm{c}\mathrm{m}}^{-2}}$ : 6±2	–	–	one‐chamber electrochemical cell galvanostatic (0.18–0.35 A cm−2) 20±3 °C 20 mL sono‐emulsion (ultrasound)
0.69 m C7	poly‐crystal‐line boron‐doped diamond	pH 6.5 1 m NaOH	–	–	$Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r};0.35\mathrm{A}{\mathrm{c}\mathrm{m}}^{-2}}$ : 40±5 $Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r};0.7\mathrm{A}{\mathrm{c}\mathrm{m}}^{-2}}$ : 14±5	–	–	one‐chamber electrochemical cell galvanostatic (0.35–0.70 A cm−2) 20±3 °C 20 mL sono‐emulsion (ultrasound)
0.1 or 0.5 m C8	pure Pt	pH 7.1–12.7 –	52±1 to 90±33 (n=2)	–	–	–	–	one‐chamber electrochemical cell potentiostatic (3 V vs. Ag/AgCl sat. KCl) 3 h 50 mL	[15]
		pH 6.5–12.6 0.06–0.5 m Na2SO4	25±1 to 70±2 (n=2)	–	–	–	–
		pH 6.6–7.7 0.05/0.5 m phosphate buffer	22±1 to 75±5 (n=2)	–	–	–	–
0.5 m C6	pure Pt	pH 7 0.25 m Na2SO4	94.1±17.4	(n=6) 41.9±3.4	$Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r}}$ : 46.2±9.6	C3–C12 alkanes C3–C5 alkenes C5 alcohols	$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0263±0.0050 $r_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r}}:$ 0.0059±0.0005	one‐chamber electrochemical cell galvanostatic (150 mA cm−2) room temperature 4 h 200 mL	this study (Figure S2)
0.5 m C6	pure Pt	pH 7 0.25 m Na2SO4	89.7±14.4	(n=6) 55.5±3.5	$Y_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r}}$ : 63.7±12.4	C3–C12 alkanes C3–C5 alkenes C5 alcohols esters	$r_{\mathrm{s}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}}:$ 0.0251±0.0040 $r_{\mathrm{d}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{r}}:$ 0.0078±0.0005	two‐chamber electrochemical cell galvanostatic (150 mA cm−2) room temperature 4 h 200 mL	this study (Figure S2)

[a] CE_substrate means the CE based on the substrate consumption and CE_dimer means the CE based on the formation of the dimerized Kolbe product. For the substrate C_n means the respective carboxylic acid with n C‐atoms. All values are either stated in the references or derived from primary data thereof as described in the Experimental section.

To propel the development of Kolbe electrolysis of MCCA, in this study we screened different commercially available electrode materials using pure Pt as benchmark. Therefore, the conversion of n‐hexanoic acid (C₆) to n‐decane (C₁₀) in aqueous solution served as model reaction (see Scheme 1). In addition to the CE for the acid degradation, we also report the CE for the produced by‐products as well as the selectivity and the yield based on the Kolbe product n‐decane.

Results and Discussion

Validation of the electrochemical cell

In order to validate the experimental setup, Kolbe electrolysis was performed in one‐ and two‐chamber electrochemical cells using a Pt anode. As Table 1 shows, Pt is the most common anode material used for Kolbe electrolysis of MCCA. CE values based on the degradation of the MCCA used as substrate of around 90 % are reported. This was confirmed for both one‐ and two‐chamber setups, with a CE_{hexanoic acid} of 94.1±17.4 % (n=6) and 89.7±14.4 % (n=6), respectively. Noteworthy, when considering the CE_decane based on the formation of the Kolbe product n‐decane (see Figure S1), a significant difference for the two configurations of the electrochemical cell is observed. CE_decane is significantly higher in the two‐chamber setup (CE_decane=55.5±3.5 %, n=6) than in the one‐chamber system (CE_decane=41.9±3.4 %, n=6; see also Figure S2). This might be explained by the prevention of side reactions by separation of anode and cathode chamber. Since the two‐chamber system has an increased internal resistance due to the presence of the membrane, the cell voltage (E _cell) was 25.4±2.1 V at the start of the experiment and decreased during the 4 h of electrolysis to 12.5±1.7 V due to the pH‐shift (see below). In contrast to this, E _cell was only 6.5±0.1 to 6.2±0.2 V in the one‐chamber setup. Also, the pH of the reaction solutions differed. In the two‐chamber setup the pH in the anode chamber dropped from 7.05±0.06 to 6.09±0.20 during the duration of the experiment due to the fact that H⁺ from the converted n‐hexanoic acid remained in the reaction solution, while the pH in the cathode chamber increased due to the accumulation of OH⁻ ions from water splitting and H₂ evolution. This can be expected as for pH neutral conditions other ions than H⁺ and OH⁻ are responsible for the charge‐balancing ion transfer.^[18] In the one‐chamber system both processes necessarily had to take place in the same reaction solution, so the pH increased from 7.00±0.08 to 8.33±0.07. Noteworthy, esters were only found in the product spectrum for the two‐chamber setup. In conclusion, based on the higher CE_decane as well as the more confined experimental space, all follow‐up experiments were performed using two‐chamber electrochemical cells.

Electrochemical characterization

As mentioned above, the most common material used as anode for Kolbe electrolysis is pure monolithic Pt. Especially for the Kolbe electrolysis of MCCA only a few other materials were investigated. For instance, RuO₂ thin film, IrO₂ thin film and Pt thin film based on Ti were assessed in comparison to pure Pt by Qiu et al. for their decarboxylation efficiency of valeric acid, showing the highest CE of 91.2 % using Pt foil as anode material while RuO₂ thin film as anode only showed a maximum CE of 49.7 %.^[16] However, all suitable materials other than pure Pt are by far not commercially available and their production is hardly scalable. In the aforementioned study, for example, the RuO₂ thin film electrode that showed the highest efficiency for electrolysis following pure Pt was prepared by pre‐treatment of the Ti serving as current collector followed by stepwise dropping RuCl₃ precursor solution on Ti, thermal decomposition, and calcination at 470 °C for 3 h.^[16] This procedure is time consuming and can only be applied to small electrode surfaces. However, cheaper electrode materials than pure Pt available at scale are needed to allow implementation of Kolbe electrolysis of MCCA at a technical scale (see also “Consequences for implementation”). Therefore, different already commercially available electrode materials were tested for conversion of n‐hexanoic acid to n‐decane by Kolbe electrolysis using the validated two‐chamber electrochemical cell.

Figure 1 summarizes the achieved CE. It is remarkable that when using platinized Ti Type B as anode a CE_{hexanoic acid} of 93.1±6.7 % is reached that is not significantly different than the CE_{hexanoic acid} for pure Pt (89.7±14.4 %). All other electrode materials show a CE_{hexanoic acid} in the range from 28.3 to 67.5 %, including the second material that is based on platinized Ti (Type A, CE_{hexanoic acid}=46.6±9.9 %). When considering the formation of the Kolbe product n‐decane the CE also strongly differs. The highest CE_decane is reached using pure Pt as anode material with 55.5±3.5 %, followed by platinized Ti Type B with CE_decane=48.3±3.2 % and the platinized Ti Type A with CE_decane=27.4±15.2 %. Using anodes based on Ru MMO and Ir MMO on Ti did not lead to the formation of n‐decane, at all. The latter is of special interest when considering that often only the substrate consumption (here the degradation of n‐hexanoic acid), but not the formation of the Kolbe product as well as of side products is analyzed (see Table 1).

Figure 1.

Overview of different anode materials for Kolbe electrolysis of n‐hexanoic acid in two‐chamber electrochemical cells with 150 mA cm⁻². CE_i for substrate consumption and the formation of different products as well as CE_overall of the electrolysis. The shown values are averages of the replicates (n) and the error bars represent the 95 % confidence interval.

Additionally, the selectivity and the yield for the production of n‐decane were studied (see Table 2). The selectivity for the dimerization is at least 50 %, even if the CE_decane is only 10.3±18.5 % as for the Pt/Ir mixture on Ti. This clearly underlines that the reaction mechanism towards Kolbe or non‐Kolbe products does not only depend on the electrode material but is also influenced by the reaction conditions such as pH, temperature, and current density.^[20] Both platinized Ti electrodes have a similar yield of 52.3±6.9 % and 60.4±43.0 %, respectively, in terms of n‐decane production as pure Pt (Y _decane=63.7±12.4 %) even if the CE_{hexanoic acid} and CE_decane are not comparable. Wadhawan et al. reported only a Y _decane of 24±3 % to 45±5 % in a one‐chamber electrochemical cell with a comparable current density converting n‐hexanoic acid into n‐decane.^[17] The selectivity of the reaction for the dimerization to n‐decane for both platinized materials and pure Pt is above 60 %, meaning on a molar basis nearly two‐thirds of the reaction products are n‐decane. Selectivities (S _decane) for the dimerization between 31.3 and nearly 60 % are reported for Pt foil depending on the reaction conditions.[¹⁶, ¹⁹] Interestingly, the observed differences of the materials on the suitability for Kolbe reaction were already indicated by analysis using cyclic voltammetry (see Supporting Information, section 4).

Table 2.

Carbon balance, yield of n‐decane per n‐hexanoic acid, and selectivity for n‐decane production of the different anodes.^[a]

Electrode material	Carbon balance	Ydecane [%]	Sdecane [%]
pure Pt (n=6)	74.6±13.6	63.7±12.4	69.9±1.8
platinized Ti Type A (n=3)	77.2±51.4	60.4±43.0	63.9±2.3
platinized Ti Type B (n=6)	65.0±7.5	52.3±6.9	66.9±0.9
Pt/Ir on Ti (n=3)	31.2±67.6	21.2±52.6	49.9±17.5
Ru MMO on Ti (n=3)	4.6±1.1	–	–
Ir MMO on Ti (n=3)	7.8±4.2	–	–

[a] “n” provides number of replicates, “±” represents the 95 % confidence interval, and “−“ indicates that the values could not be calculated, as no n‐decane was gained.

Of special interest, however, is the difference of platinized Ti Type A and B that were subjected to surface analysis together with the benchmark material pure Pt.

Surface and corrosion analysis

Surface properties were studied at the mm‐ and μm‐scale. Since only pure Pt as well as platinized Ti Type A and B showed a CE_decane of at least 25 %, the other materials were excluded from the physical‐chemical examination. The materials were studied before (“as‐received”) and after use for Kolbe electrolysis. Further, the loss of Pt to the electrolysis solution due to corrosion was examined.

At the mm‐scale optical profilometry reveals a remarkable difference between the surfaces of the pure Pt electrodes and the two types of platinized Ti (see Figure 2; Figure S4). Whilst the surface of pure Pt is relatively smooth with regular longitudinal grooves, the platinized Ti electrodes possess a more regular surface structure consisting of spherical elevations (Figure 2 B, C). A closer look at the latter reveals further differences amongst the platinized Ti electrodes: The Type A electrode is very finely structured with small spherical elevations appearing like fine granules, which are attached to the surface, leading to an increased overall roughness. These, however, are much coarser on the Type B electrode. Surprisingly, on the mm‐scale there are no significant differences between new and used electrodes (Figure 2). To quantify the physical surface properties, the roughness R _α, which provides the arithmetical mean deviation of the surface profile, was calculated (Table 3). As can be seen, R _α of all three electrode materials is in the same order of magnitude, which does apparently not reflect the different visual perception of the surface structure discussed above. More important, however, it cannot conclusively explain their different electrochemical performances.

Figure 2.

Exemplary microscopic pictures of the electrode surface taken with a optical profilometer. Per new (A–C) and per used electrode material (D–F), 2 technical replicates were done in total. (A, D) Pure Pt; (B, E) platinized Ti Type A; (C, F) platinized Ti Type B.

Table 3.

Roughness R _α±SD calculated from the profilometric height profiles of the different electrode materials. For each material two height profiles were taken.

Electrode material	Rα [μm]
pure Pt new	1.531±0.15
pure Pt used	5.174±0.69
platinized Ti Type A new	1.182±0.16
platinized Ti Type A used	1.099±0.01
platinized Ti Type B new	3.177±0.81
platinized Ti Type B used	2.820±0.47

Subsequently, the surface was examined on the μm‐scale using scanning electron microscopy (SEM) coupled with energy‐dispersive X‐ray spectroscopy (EDX) (Figure 3 with all pictures are shown in Figures S5 and S6).

Figure 3.

Exemplary SEM (left column) and EDX (right column) pictures of the electrode surface showing the same section of the electrode. Per electrode material 4 technical replicates were done each for the new (C, D, G, H) and the used electrodes (A, B, E, F, I, J) in total. (A, B) Pure Pt; (C–F) platinized Ti Type A; (G–J) platinized Ti Type B. Scale bar: 50 μm.

SEM micrographs acquired on smaller fields of view are partially in contrast to the profilometry data. For instance, the used pure Pt electrode, considered rough according to profilometry, appears smooth in the SEM image (Figure 3A). Also, the fine structure measured by profilometry on new, as received platinized Ti Type A electrodes appears cracked and shows many small indentations in the SEM image (Figure 3C). In contrast to this, the new platinized Ti Type B electrode (Figure 3G) is nearly as smooth as the pure Pt and there are only a few spots where small granules can be seen. Here, the surface looks like it consists of valleys and mountain ranges. Comparing the two platinized Ti electrodes after use, the surface of the Type B electrode has hardly changed (Figure 3I). Only some small cracks and brittle‐looking areas appear after six times of use for 4 h each at 150 mA cm⁻². The surface of the platinized Ti Type A electrode, however, has deteriorated significantly. Larger areas exposing the underlying Ti serving as current collector formed due to detachment of the Pt coating during the electrolysis (Figure 3E). After the electrolysis, the detached Pt can be found in the aqueous phase of the reaction solution (see Table 5). For Type B, the SEM micrographs look more stable, as here the surface is very similar to the surface of pure Pt before as well as after use.

Table 5.

Average concentration±SD of Pt in the aqueous phase after each electrolysis run.^[a]

Electrode material	cPt [ng mL−1]
pure Pt (n=3)	24.63±9.47
platinized Ti Type A (n=3)	99.00±10.03
platinized Ti Type B (n=6)	25.85±3.83
control (n=3)	0.13±0.06

[a] For pure Pt, the platinized Ti Type A electrode and the control 3 samples each from independent experiments and 6 samples from independent experiments for the Type B electrode were measured.

To complement the assessment of the electrode surface structure using SEM, EDX analysis was carried out to examine the distribution of Pt and Ti on the surface (Figure 3, right column; Figures S5 and S6). In addition to the Pt signal only the background noise was recorded for the pure Pt electrode, likewise for 3 out of 4 recordings of the new platinized Ti Type B electrode and 2 out of 4 recordings of the used platinized Ti Type B electrode. This shows clearly that the entire surface of both electrodes was covered by Pt. The remaining recordings reveal clear differences as shown in Figure 3D,F,J with a high EDX‐signal for Ti (red) and a high signal for Pt (green). Already for the new platinized Ti Type A electrode, many small, sharply defined spots of Ti are observed. This indicates that the Pt coating (on the Ti base material) possesses holes of 3–40 μm, and thus both Pt and Ti face the electrolyte solution. We reason that during electrolysis this leads to an increased occurrence of the competitive reaction to the Kolbe electrolysis, the evolution of oxygen from water electrolysis. This lowers the CE_decane to 27.4±15.2 % for the platinized Ti Type A compared to CE_decane=55.5±3.5 % for pure Pt. On the other site the CE${}_{\mathrm{O}{}_2}$ increases from 0.6±0.6 % for pure Pt to 16.7±2.5 % for the platinized Ti Type A. In contrast, EDX analysis of the new platinized Ti Type B electrode does not reveal any Ti on the electrode surface, which leads to the conclusion that the water electrolysis is suppressed, as it is for monolithic Pt electrodes. This is perfectly in line with the low CE${}_{\mathrm{O}{}_2}$ of only 1.4±0.9 %. The EDX maps of the electrodes after electrolysis reveal that for the Type A electrode the exposed surface of Ti has increased significantly. There are many areas of Ti exposure of diameters ranging from 3–120 μm. Only for 2 out of the 4 recordings of the used Type B electrode Ti was detected, but only isolated and with a small diameter of 5–30 μm. In order to quantify the optical assessment of the surface coverage, the EDX recordings were analyzed, and the degree of coverage was calculated (Table 4). This shows that even for the platinized Ti Type A electrodes the Ti exposure is minor. Yet, together with the electrochemical performance (see “Electrochemical characterization”) it provides clear evidence that even a small proportion of 1–3 % Ti exposed to the aqueous electrolyte solution has a strong influence on the CE.

Table 4.

Area%±SD of Pt and Ti on the electrode surface. Per electrode material 4 technical replicates were done diagonally across the electrode.

Electrode material	Ti area [%]	Transition area [%]	Pt area [%]
pure Pt	0.00±0.00	0.00±0.00	100.00±0.00
platinized Ti Type A new	0.98±0.47	1.06±1.08	98.02±1.49
platinized Ti Type A used	2.63±1.42	3.96±2.34	93.49±3.72
platinized Ti Type B new	0.15±0.26	0.12±0.21	99.74±0.45
platinized Ti Type B used	0.54±0.16	2.18±2.15	97.41±1.87

To further investigate this corrosion, the loss of Pt to the aqueous electrolyte solution after electrolysis was analyzed by inductively coupled plasma mass spectrometry (ICP‐MS). The results in Table 5 show that all electrodes lose Pt during the electrolysis. The Pt concentration in the aqueous phase after electrolysis with the platinized Ti Type B electrode as anode is with 25.85±3.83 ng mL⁻¹, comparable to the concentration when pure Pt is used as anode (c _Pt=24.63±9.47 ng mL⁻¹). However, the Type B electrode is only platinized, and here the loss of Pt leads to a loss of electrocatalytic surface area for the Kolbe electrolysis as discussed above. This problem does not occur with a pure monolithic Pt electrode that can be considered as infinite reservoir of electrocatalyst. The average Pt concentration for the platinized Ti Type A electrode is 99 ng mL⁻¹, which is almost 4 times higher than the concentration when using the Type B electrode or pure Pt, but still is highly suitable for Kolbe electrolysis. This increased corrosion of Pt has to be taken into account when considering the use of different electrode materials at technical scale with the platinized Ti Type A electrode wearing out significantly fast. Most likely the corrosion process is due to detached Pt particles caused by the gas evolution at the electrode surface.

Consequences for implementation

The transfer of Kolbe electrolysis from laboratory to industrial scale requires anodes with a low capital expenditure (capex). Although there is a great potential for process improvement in electrochemical reactor engineering,^[21] for simplicity reasons we are now considering an volume‐to‐electrode ratio of 5 cm² L⁻¹ as used here. For a 100 L reactor a 500 cm² electrode would be required. A suitable monolithic pure Pt electrode of this size would cost at least 24 € cm⁻², being 12.000 €. The platinized Ti Type B that is just as suitable for Kolbe electrolysis as pure Pt costs only 330 € for 500 cm² (0.66 € cm⁻²), about 36 times less than monolithic Pt (see Table S2). The platinized Ti Type B material thus offers the possibility of combining high performance with an acceptable capex when the Kolbe electrolysis is scaled up to a technical scale. This is further substantiated by the expected lifetime of the electrode of almost 10000 h, more than one year (see Table S4). As this assumes constant operation, which is highly unlikely for Power‐to‐X aiming for the utilization of surplus electric energy from fluctuations from renewables,^[22] an even longer lifetime can be expected. In this line a further decrease of capex can be expected when replacing Ti by equally corrosion‐resistant materials that are also suitable for platinization using high‐temperature electrolysis, with alloyed stainless steels being most promising.

The derived operational expenditures (opex) of 1 L of organic product mixture that may serve as model for drop‐in fuel additive without the need of down‐streaming^[13] can be simplified as follows: Per 1 mol of converted n‐hexanoic acid, 65.7 mL of liquid fuel mixture is produced with pure Pt, and 56.7 mL with platinized Ti Type B as anode (see the Supporting Information, section 8). This means in order to produce 1 L of fuel mixture it is necessary to convert 15.2 mol of n‐hexanoic acid using 6.22 kWh (pure Pt) or 17.6 mol of n‐hexanoic acid using 6.66 kWh (platinized Ti Type B) (see Table S3). When considering the electric energy price this equals an opex of 1.14 € (1.38 US$, pure Pt) or 1.22 € (1.47 US$, platinized Ti Type B) per L fuel mixture. Although this opex does not consider, for example, the costs for agitation of the reaction solution, this calculation has two key messages: First, there is only a minor price increase in the opex of <10 % when using platinized Ti as anode material, but a significant decrease in capex as shown above. Second, the costs of fuel by Kolbe electrolysis are within a reasonable range (albeit tax, etc. is not included), but at the same time a green premium might be worth considering.^[23]

Conclusion

Based on a highly reproducible experimental setup this study shows that commercially available platinized Ti can be used as anode for the efficient galvanostatic conversion of n‐hexanoic acid to n‐decane in aqueous solution with a coulombic efficiency (CE_{hexanoic acid}) of 93.1±6.7 %. Thereby, the Kolbe product n‐decane is produced with a selectivity of 66.9±0.9 % and a CE_decane of 48.3±3.2 %. This performance together with the economic consideration demonstrates that platinized Ti can fully replace pure Pt as anode material for the Kolbe electrolysis.

However, we have demonstrated that the surface properties of the platinized Ti strongly influence the CE and selectivity. Thus, apparently similar materials at similar capital expenditure show strongly deviating electrochemical suitability. Consequently, further studies need to address the structure–function relationships in more detail as it was done, for instance, for the fabrication of electrode materials for Li‐ion batteries.[²⁴, ²⁵] This will shed light on the impact of the different production procedures and steps on the electrode properties and hence their suitability to be used for the Kolbe electrolysis or other electro‐organic syntheses, such as hydrogenations.^[26]

With this study the production of fuel additives by Kolbe electrolysis of medium‐chain carboxylic acids (MCCAs) comes closer to application. Consequently, as follow‐up the scaling not only of electrodes but of entire electrochemical cells and the operation of these devices needs to be addressed as well as the conversions of mixtures of MCCA. Aspects to be considered include electrode topographies and geometries, like the use of foams or other 3D electrodes to increase the active electrode surface area, or the insertion of turbulence promoters inside flow channels to increase mass transfer.^[21] Improvement of electrochemical synthesis can also be achieved by reactor engineering, such as the use of parallel plate flow cells,^[21] microchannel reactors,^[27] or pulsing techniques combined with artificial intelligence.^[28] Finally, to show applicability, the operation at scale with different loads at different current densities, as it can be expected from fluctuations from renewable electric power, needs to be addressed.

Experimental Section

General remarks

All chemicals were of at least analytical grade or as stated. All solutions were prepared with bidistilled water (Milli‐Q IQ 7000, Merck KgaA, Darmstadt, Germany). If not stated otherwise all experimental potentials refer to Ag/AgCl [saturated KCl, 0.197 V vs. standard hydrogen electrode (SHE)]. A list of all used symbols and abbreviations can be found in the Supporting Information.

Chemicals

Na₂SO₄ (anhydrous, ≥99 %) was obtained from Carl Roth GmbH+Co. KG (Karlsruhe, Germany). H₂SO₄ (98 %), NaOH (pellets), acetone, n‐hexane for dilution series (for GC), dichloromethane (dried), acetic acid (p. A.), iso‐butyric acid (for synthesis), ethanol (p. A.), 2‐propanol (for liquid chromatography), 2‐butanol (p. A.), 2‐pentanol (for synthesis), 3‐pentanol (for synthesis), 1‐hexanol (for synthesis), 1‐heptanol (for synthesis), 2‐heptanol (for synthesis), 1‐octanol (pure), n‐pentane (for spectroscopy), and n‐hexanoic acid‐3‐methylbutyl ester (for synthesis) were supplied by Merck KGaA (Darmstadt, Germany). 1‐Propanol (analytical standard), 1‐butanol (analytical standard), 1‐pentanol (99 %), n‐propanoic acid (99 %), n‐butyric acid (99 %), n‐valeric acid (99 %), n‐hexanoic acid (99 %), n‐heptanoic acid (99 %), n‐octanoic acid (99 %), n‐nonanoic acid (99 %), n‐decanoic acid (99 %), n‐hexane as analytical standard (95 %), n‐heptane (99 %), n‐octane (99 %), n‐nonane (99 %), n‐decane (99 %), n‐undecane (99 %), n‐dodecane (99 %), n‐tridecane (99 %), n‐tetradecane (99 %), n‐pentadecane (99 %), n‐hexadecane (99 %), n‐heptadecane (99 %), n‐octadecane (99 %), isobutyl hexanoate (98 %), cyclohexanone (99.8 %), and undecanoic acid methyl ester (analytical standard) were purchased from Sigma‐Aldrich (St. Louis, USA). Iso‐valeric acid (98 %) was supplied by Fluka Chemie GmbH (Buchs, Switzerland) and 2‐hexanol (98 %) as well as 3‐heptanol (98 %) by abcr GmbH (Karlsruhe, Germany). The carrier gases helium and argon (99.999 vol % purity) and the calibration gases for level 1, 2, 5, and 6 (composition see Table S1) for GC as well as N₂ (>99 vol %) were obtained from Air Products GmbH (Hattingen, Germany). The calibration gases for level 3 and 4 (composition see Table S1) for GC were derived from Air Liquid S.A. (Paris, France) and Praxair N.V. (Oevel, Belgium).

Experimental setup

Electrochemical decarboxylation of n‐hexanoic acid by Kolbe electrolysis, as well as acidification after the electrolysis were performed in a three‐neck 250 mL flask (Schott AG, Mainz, Germany, Figure 4). The central neck contained a butyl‐rubber stopper equipped with the working electrode (WE), the counter electrode (CE), the reference electrode (RE), a tailor‐made electrode holder, as well as two needle ports, one for gassing the reaction solution and acidification and the other one as gas outlet. The electrode holder was printed from poly(lactic acid) (BASF/Innofil3D, Emmen, Netherlands) using a 3D printer (Ultimaker 2+, Ultimaker, Utrecht, Netherlands) keeping the WE and the CE chamber in a defined position with a 18 mm distance between WE and CE. The two external necks contained a pH electrode and a conductivity electrode as shown in Figure 4. The whole setup was fabricated gas tight, except the gas inlet and gas outlet.

Figure 4.

Scheme of electrochemical cell.

All experiments were carried out under galvanostatic conditions by a DC power source (2230‐30‐1 triple Channel DC Power Supply, Keithley/Tektronix GmbH, Köln, Deutschland) using a three‐electrode setup consisting of a WE with a geometric surface area of 2 cm², an Ag/AgCl sat. KCl reference electrode (SE 11, Xylem Analytics Germany Sales GmbH & Co. KG/Sensortechnik Meinsberg, Waldheim, Germany), and a CE (Platinum foil, 1×1.2 cm², Goodfellow, Huntingdon, UK). The WE materials used in the experiments were: platinum foil (1×1 cm², Goodfellow, Huntingdon, UK), platinized titanium (1.4×1.4 cm², single sided, Magneto/Evoqua Water Technologies, Pittsburgh, USA; from now on called platinized Ti Type A), platinized titanium (1×1 cm², Umicore, Schwaebisch Gmuend, Germany; from now on called platinized Ti Type B), platinum–iridium mixture deposited on titanium (Pt/Ir on Ti, 1.4×1.4 cm², single sided, Magneto/Evoqua Water Technologies, Pittsburgh, USA), ruthenium multi‐metal‐oxide deposited on titanium (Ru MMO, 1×1 cm², Magneto/Evoqua Water Technologies, Pittsburgh, USA), and iridium multi‐metal‐oxide deposited on titanium (Ir MMO, 1.4×1.4 cm², single sided, Magneto/Evoqua Water Technologies, Pittsburgh, USA). If the WE material was only single sided the uncoated side was insulated with two‐component epoxy‐glue (EA3430 LOCTITE, Henkel AG & Co. KG, Düsseldorf, Germany or kwikweld, JB Weld, Sulphur Springs, USA). The DC power supply was connected to the WE serving as anode and the CE serving as cathode, respectively. The cell potential between WE and CE (E _cell) was measured using the DC power supply. The anode potential was measured relative to the RE with an additional multimeter (Autoranging Mini MultiMeter MN16, Extech Instruments, Nashua, USA). The system was operated in two‐chamber configuration with the cathode chamber separated via a custom‐made glass tube interfaced via an ion exchange membrane (fumasep FKS‐PET‐130, FUMATECH BWT GmbH, Bietigheim‐Bissingen, Germany). The working volume of the anode chamber was 200 mL and that of the cathode chamber was 9 mL.

The n‐hexanoic acid solution in the anode chamber was purged with nitrogen for at least 15 min before each electrolysis. For all connections gas tight Tygon tubes (Saint‐Gobain, Charny, France) were used. All experiments were performed at room temperature and the n‐hexanoic acid solution in the anode chamber was continuously mixed via a magnetic stir bar at 1000 rpm. The pH and the conductivity were measured continuously during the experiment using a SevenExcellence S470 (Mettler‐Toledo, Greifensee, Switzerland) with an InLab Micro Pro pH electrode and an InLab 710 conductivity electrode (both Mettler‐Toledo, Greifensee, Switzerland). Both electrodes were calibrated with the commercial buffer solutions (Mettler‐Toledo, Greifensee, Switzerland) directly before each experiment. After each electrolysis, the anode was cleaned with acetone and afterwards rinsed with water. All other inlet parts of the setup were also cleaned with water.

Kolbe electrolysis

To prepare the reaction solution 450 mL of water were introduced into a 500 mL flask and 33 mL of n‐hexanoic acid was added. Thereafter, the flask was weighed in order to determine the exact amount of n‐hexanoic acid used for each experiment. Subsequently, the pH of the solution was adjusted to pH=7 using NaOH pellets and H₂SO₄. The conductivity was adjusted by adding Na₂SO₄ to a concentration of 0.25 mol L⁻¹. Afterwards, water was added to exactly 500 mL. The solution for the cathode chamber was prepared to a similar pH and conductivity by adding NaOH pellets, H₂SO₄, and Na₂SO₄.

The 500 mL n‐hexanoic acid solution was divided into 200 mL for the electrolysis and 200 mL as blank solution (100 mL remained unused). The 200 mL for the electrolysis were filled into the electrochemical cell and the reactor was weighed. Thereafter, the cathode chamber was filled with 9 mL of the cathode chamber solution, and the complete experimental setup was assembled. The gas outlet was connected to a N₂‐mass flow meter/controller (MFM; LOW‐ΔP‐FLOW F‐101D, 60 mLn min⁻¹, Bronkhorst High‐Tech B.V., Ruurlo, Netherlands) controlled via a Flow‐Bus (Bronkhorst High‐Tech B.V., Ruurlo, Netherlands) with a microGC (3000 Micro GC, INFICON, Cologne, Germany) in by‐pass in order to determine the composition and the volume of the produced gas. The electrolysis was carried out for 4 h with a constant current of 300 mA at room temperature and 1000 rpm stirring.

After the electrolysis, the system was kept gas tight. The N₂‐mass flow meter/controller was changed to a CO₂‐mass flow meter/controller (EL‐FLOW Select F‐201 CV, 500 mLn min⁻¹, Bronkhorst High‐Tech B.V., Ruurlo, Netherlands) also controlled via a Flow‐Bus (Bronkhorst High‐Tech B.V., Ruurlo, Netherlands). Through the needle port 50 % H₂SO₄ was added to adjust to pH 2. Thereby, the dissolved CO₂ was gassed out and the volume of CO₂ was measured in order to allow calculation of the carbon balance. Also, the pH of the blank solution was adjusted to pH 2 with H₂SO₄. The acidification of the reaction and the blank solution led to an increased phase separation between organic phase and aqueous phase.

Gas‐phase analysis

During electrolysis, the gas composition was analyzed with a four‐channel microGC equipped with a thermal conductivity detector, which was calibrated for the components to be analyzed (see Table 6). Measurements were carried out at the beginning of the electrolysis (t ₀) and every 10 min during the first 90 min of the experiment. During the remaining 150 min of the experiment the gas composition was determined every 30 min.

Table 6.

Specifications of the used microGC‐TCD and method for gas composition analysis.^[a]

Column	Carrier gas	T column [°C]	t injection [ms]	p column [psi]	Analyzed components
14 m molsieve with 2 m Plot U pre‐column, 1 μL backflush injector	argon	100	0	25	H2; O2; N2
8 m Plot Q, variable volume injector	helium	80	25	20	CO2; propylene; propane; 1‐butene; butane
8 m OV‐1, 1.2 μm thick, variable volume injector	helium	60	250	20	propylene/propane; butene/butane; iso‐pentane; 1‐pentene; pentane; 2‐methyl‐2‐butene; 2‐cis‐pentene; 2‐trans‐pentene; hexane
10 m Stabilwax, variable volume injector	helium	60	250	15	–

[a] For all columns: sample inlet temperature=100 °C; injector temperature=100 °C; running time=420 s.

From the mass flow controller (see section above), the measured volume $v_{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{d}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ [mL] is gained and the mole fraction y_i of each individual gas component i [%] is obtained from the microGC‐TCD measurement. Determined gaseous reaction products are shown in Figure S1 (green boxes).

Liquid‐phase analysis

GC−MS for quantification of liquid electrolysis products: To quantify the amount of the reaction products (see Figure S1, orange boxes) in the liquid phase, first a phase separation was carried out in a separating funnel for both the reaction solution and the blank solution for at least 2 h. After phase separation both phases (aqueous and organic) were separately filled into flasks. The electrochemical reactor flask and the separation funnel were rinsed with acetone, which was added to the aqueous phase. The weight of both phases was determined. The density of the phases was determined by weighing 1 mL. The aqueous phase was diluted in acidic water (water adjusted with H₂SO₄ to pH=2) and the organic phase in n‐hexane and parallel in dichloromethane. Dilutions of 1 : 10, 1 : 100, and 1 : 1000 were analyzed.

The dilution series of the aqueous phase as well as both dilutions series of the organic phase were analyzed via GC–MS (GC 7890 A and MSD 5975 C InertXL, Agilent, Santa Clara, USA), using a DB‐FATWAX capillary column (30 m×250 μm×0.25 μm, Agilent, Santa Clara, USA) with helium as carrier gas, undecanoic acid methyl ester as internal standard for the organic phase, and cyclohexanone as internal standard for the aqueous phase. The initial temperature was 50 °C (held for 2 min) and with a temperature ramp of 15 K min⁻¹ it was increased to 250 °C. Carboxylic acids (CA), n‐alkanes, alcohols, and esters were identified using retention times and mass spectra of pure compounds. CA (aqueous: C₂−C₆ and C₇−C₈, each 3 levels; organic: C₄−C₁₀, 4 levels), n‐alkanes (organic: C₅−C₇ and C₈−C₁₈, each 4 levels), alcohols (aqueous: C₂−C₆, 3 levels; organic: C₃−C₈, 3 levels), and some esters (organic: C₁₀−C₁₁, 3 levels) were quantified using external standards. The concentration of esters that were not calibrated (a very minor share) was estimated using an average response factor of all calibrated esters.

Determination of Pt concentration in aqueous phase: The loss of Pt from the surface of the electrode during electrolysis was determined using ICP‐MS measurements (ELEMENT™ XR ICP‐MS, Thermo Scientific, USA) of the aqueous phase after phase separation. Calibration was conducted with external Pt standards (0.05–1 ng mL⁻¹, 5‐point‐calibration) at a mass of 195. Re was added as internal standard to the sample solution at a concentration of 3.5 ng mL⁻¹ and measured at a mass of 185. All samples were diluted 1 : 100 for the measurement. The limit of quantification was 0.065 ng mL⁻¹. For control the aqueous phase of an n‐hexanoic acid solution (identically treated) was used.

Electrode surface analysis

Profilometry: To determine the physical electrode surface properties an optical profilometer (Sensofar SNeox 3D, Sensofar Metrology, Terrassa, Spain) with green LED illumination and 10× objective was used to acquire topography maps of the electrode surfaces with a lateral resolution of approximately 1 μm. In order to determine the surface roughness two topography maps per electrode were acquired as 8‐bit grey‐scale images with the topography encoded in the pixel brightness. These were then further analyzed using the software Fiji (Fiji is just ImageJ, Version: ImageJ 1.52p, https://imagej.net/Fiji,^[29]) in order to obtain a map showing the deviation of the topography from the average height level: First, the average grey value of all pixels was calculated and subtracted from the image. Second, the absolute value was calculated for each pixel. Subsequently, the average grey value of the obtained image was determined as before. This average grey value was then divided by 255, which is the maximum range of grey values of an 8 bit image, and multiplied by the highest altitude value on the scale bar in order to calculate the roughness R _α. Thus, the arithmetic mean of the absolute deviation from the average height is defined as roughness R _α in this study.

SEM/EDX: To determine the physical and chemical properties of the surface at a microscopic level, SEM imaging in combination with EDX was used. For that a Zeiss Merlin VP Compact field‐emitting scanning electron microscope (Carl Zeiss Microscopy, Oberkochen, Germany) with the software package SmartSEM was used to acquire four SEM images per electrode. Additionally, to visualize the elemental distribution on the electrode surface via EDX a Bruker Quantax FlatQuad spectrometer (Bruker Nanoanalytics, Berlin, Germany) with the Software Esprit was used. In order to excite the L‐alpha line of Pt an electron acceleration voltage of 18 kV was chosen. The beam current amounted to approximate 250 pA throughout the experiments. The elemental distribution was mapped as a grey‐scale image (white: high concentration of Ti on the surface; black: absence of Ti on the surface), which was then used for further processing with Fiji (Fiji is just ImageJ, Version: ImageJ 1.52p, https://imagej.net/Fiji,^[29]). For convenience, the grey‐scale image was subsequently inverted, so black refers to a high Ti concentration on the surface and white to the absence of Ti, which then shows the presence of Pt on the surface in this study. Afterwards, different threshold limits were set to determine the percentage of Pt, Ti, and the transition from one to the other element on the electrode surface. The percentage is calculated from the amount of pixels within a given threshold range divided by the amount of all pixels of the image. Within a threshold of 0–50 only Ti was considered being present on the surface, between a threshold of 50–100 it was considered as transition area, and with threshold from 100–255 only Pt was considered being present on the surface. For illustration purposes, the element map was also color coded, with green reflecting the presence of Pt on the electrode surface and red the presence of Ti (see Figure 3).

Data processing and calculations

Quantification of gas components: For quantification of individual components in gas mixtures data from the MFM and the microGC‐TCD measurements were combined. This allowed to calculate the absolute amount of each gaseous component that was produced during electrolysis. First, the total molar norm volume of the gas mixture, $V_{{\mathrm{m}}_{\mathrm{m}\mathrm{i}\mathrm{x}}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ , was calculated as the sum of the incremental molar norm volumes, ${\Delta V}_{{\mathrm{m}}_i}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ , as follows [Eq. 1]:

$${\displaystyle V_{{\mathrm{m}}_{\mathrm{m}\mathrm{i}\mathrm{x}}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}=\sum \Delta V_{{\mathrm{m}}_i}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}$$ (1)

with Equation 2:

$${\displaystyle {\Delta V}_{{\mathrm{m}}_i}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}=\frac{y_i}{100\%}·V_{{\mathrm{m}}_i}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}\mathrm{a}\mathrm{n}\mathrm{d}{V}_{{\mathrm{m}}_i}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}=\frac{M_i}{{\varrho }_i^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}}$$ (2)

where $M_i$ is the molecular weight of the gaseous component $i$ , ${\varrho }_i^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ is the gas density of the component $i$ at norm conditions, and $y_i$ the mole fraction [%] of each individual gas component $i$ measured (see “Gas‐phase analysis”). This and all following calculations were performed for each sampling point individually. Norm conditions are defined as T=273.15 K and p=1.01325 bar in this paper.

The volume fraction, ${\varphi }_i^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ , is gained by Equation 3:

$${\displaystyle {\varphi }_i^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}=\frac{{\Delta V}_{{\mathrm{m}}_i}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}{V_{{\mathrm{m}}_{\mathrm{m}\mathrm{i}\mathrm{x}}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}}$$ (3)

As each gaseous component shows a different response in the N₂‐calibrated MFM, the measured volume of the gas mixture is corrected by a conversion factor. The reciprocal conversion factor C _mix ⁻¹ of the gas mixture is calculated as the sum of the reciprocal incremental conversion factors ΔC_i ⁻¹ [Eq. 4]:

$${\displaystyle \frac{1}{C_{\mathrm{m}\mathrm{i}\mathrm{x}}}=\sum \frac{1}{{\Delta C}_i}}$$ (4)

with Equations (5) and 6:

$${\displaystyle \frac{1}{{\Delta C}_i}={\varphi }_i^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}\times \frac{1}{C_i}}$$ (5)

$${\displaystyle C_i=\frac{{\varrho }_{\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{i}\mathrm{b}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}\times c_{\mathrm{p};\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{i}\mathrm{b}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}{{\varrho }_i^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}\times c_{\mathrm{p};i}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}}$$ (6)

where ${\varrho }^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ is either the gas density of component $i$ or that of the calibration gas of the MFM at norm conditions, $c_{\mathrm{p}}$ is either the specific heat capacity of component $i$ or that of the calibration gas, and ${\varphi }_i^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ is the calculated volume fraction of each gas component [see Eq. (3)].

The actual incremental gas volume $\Delta v_{\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{l}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ of the gas mixture is determined via the average of the conversion factors $C_{\mathrm{m}\mathrm{i}\mathrm{x}}$ of two successive gas measurements denominated as $t_n$ and $t_{n+1}$ , calculated with Equation (4), and the measured incremental gas volume $\Delta v_{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{d}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ of the gas mixture between these two points in time [Eq. 7]:

$${\displaystyle {\Delta v}_{\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{l}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}=\frac{C_{\mathrm{m}\mathrm{i}\mathrm{x};t_{\mathrm{n}}}+C_{\mathrm{m}\mathrm{i}\mathrm{x};{\mathrm{t}}_{n+1}}}{2}\times \Delta v_{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{d}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}$$ (7)

Up to this point the calculation is based on norm conditions since these are the calibration conditions for the MFM. Therefore, a conversion to standard conditions (STP) is necessary [Eq. (8)]. Standard conditions are defined as T=273.15 K and p=1.00 bar according to the IUPAC definition.

$${\displaystyle {\Delta v}_{\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{l}}^{\mathrm{S}\mathrm{T}\mathrm{P}}=\frac{{\Delta v}_{\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{l}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}·p^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}{p^{\mathrm{S}\mathrm{T}\mathrm{P}}}}$$ (8)

To calculate the absolute amount of substance of each component $n_i$ , the incremental amounts of substance $\Delta n_i$ are summed up [Eq. 9]:

$${\displaystyle n_i=\sum \Delta n_i}$$ (9)

with Equations (10) and 11:

$${\displaystyle \Delta n_i=\frac{{\Delta v}_i^{\mathrm{S}\mathrm{T}\mathrm{P}}}{V_{{\mathrm{m}}_i}^{\mathrm{S}\mathrm{T}\mathrm{P}}}}$$ (10)

$${\displaystyle \Delta v_i^{\mathrm{S}\mathrm{T}\mathrm{P}}=\left(\frac{{\varphi }_{i;{\mathrm{t}}_n}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}+{\varphi }_{i;{\mathrm{t}}_{n+1}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}{2}\right)\times {\Delta v}_{\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{l}}^{\mathrm{S}\mathrm{T}\mathrm{P}}}$$ (11)

where ${\varphi }_i^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ are the calculated volume fractions of the gaseous component at two successive measurements [see Eq. (3)], $\Delta v_{\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{l}}^{\mathrm{S}\mathrm{T}\mathrm{P}}$ is the actual incremental volume of the gas mixture calculated with Equation (7), and $V_{{\mathrm{m}}_i}^{\mathrm{S}\mathrm{T}\mathrm{P}}$ is the calculated molar norm volume of the component at STP conditions.$V_{{\mathrm{m}}_i}^{\mathrm{S}\mathrm{T}\mathrm{P}}$ is calculated in the same way as $V_{{\mathrm{m}}_i}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ [see Eq. (1)], but ${\varrho }_i^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ is replaced by ${\varrho }_i^{\mathrm{S}\mathrm{T}\mathrm{P}}$ .

Additionally, the dead volume of the experimental setup was taken into account. The gas composition in the dead volume corresponds to the composition determined via the last microGC measurement. The conversion of the determined dead volume $v_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}^x$ into the dead volume $v_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}^{\mathrm{S}\mathrm{T}\mathrm{P}}$ under STP conditions is done as follows [Eq. 12]:

$${\displaystyle v_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}^{\mathrm{S}\mathrm{T}\mathrm{P}}=\frac{v_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}^x\times T^{\mathrm{S}\mathrm{T}\mathrm{P}}}{T^x}}$$ (12)

where $T^x$ is the ambient temperature and T ^STP=273.15 K.

The amount of substance in the dead volume $n_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d};i}$ can hence be derived from Equation 13:

$${\displaystyle n_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d};i}=\frac{{\Delta n}_i\times v_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}^{\mathrm{S}\mathrm{T}\mathrm{P}}}{\Delta v_{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{d}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}}}$$ (13)

where $v_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}^{\mathrm{S}\mathrm{T}\mathrm{P}}$ is the calculated dead volume [see Eq. (12)] under STP conditions, ${\Delta n}_i$ is the incremental amount of substance between the last two measurements of the experiment, and $\Delta v_{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{d}}^{\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{m}}$ is the measured incremental gas volume of the gas mixture between these last two measurements.

Coulombic efficiencies: The CE for each component $i$ , CE_i, for electrochemical n‐hexanoic acid decarboxylation and the formation of different products (Table 7) was derived from the charge $Q_i$ calculated for each reaction in relation to the total charge ($Q_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ ) measured during electrolysis according to Equation (14). $Q_i$ was calculated from the consumed or produced amount of substance $n_i$ within the reaction electrolysis solution (the volume was assumed to be constant):

$${\displaystyle {\mathrm{C}\mathrm{E}}_i=\frac{Q_i}{Q_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}}\times 100\%}$$ (14)

Table 7.

Assumed number of transferred electrons z_i for n‐hexanoic acid and different electrolysis products.

Component $\mathbf{i}$	zi
Substrate
n‐hexanoic acid (decarboxylation)	1
Products
n‐decane	2
n‐pentane	1
n‐octane	6
n‐nonane	4
n‐undecane	4
n‐dodecane	2
3‐pentanol	2
2‐pentanol	2
1‐pentanol	2
n‐hexanoic acid‐2‐pentyl ester	2
n‐hexanoic acid pentyl ester	2
CO2	1
H2	2
O2	4
Σ prop[a]	5
Σ but[b]	3
Σ pentene[c]	1[d] 2[d]
n‐pentane	1
n‐hexane	0

[a] Equals the combined measurement of propylene/propane via microGC. [b] Equals the combined measurement of butene/butane via microGC. [c] Sum of 1‐pentene, 2‐methyl‐2‐butene, 2‐cis‐pentene, and 2‐trans‐pentene measured via microGC. [d] z=1 was used for the same molar amount as n‐pentane since they are produced in equimolar amount via disproportionation (see Scheme 1); z=2 was used for the additional produced amount of Σ pentene.

with Equation 15:

$${\displaystyle Q_i=n_i\times z_i\times F}$$ (15)

where $n_i$ is the amount of each substance in mol, $z_i$ is the number of transferred electrons per molecule (see Table 7), and F=96.485 C mol⁻¹ is the Faraday constant. $Q_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ can be derived from the integrated current ($i$ ) that was set over time [Eq. 16]:

$${\displaystyle Q_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}=\int i\left(t\right)\mathrm{d}t}$$ (16)

Since all experiments were carried out galvanostatically, $Q_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ was always 4320 C, meaning in all experiments 0.44 Faraday equivalents were transferred.

Additionally, the CE_overall of the electrolysis was calculated according to Equation (17) from the sum of $Q_i$ of all products (excluding CO₂, H₂ and O₂):

$${\displaystyle {\mathrm{C}\mathrm{E}}_{\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{l}\mathrm{l}}=\frac{\sum Q_{i,\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{s}}}{Q_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}}\times 100\%}$$ (17)

Carbon balance, yield, selectivity, and rates of the electrolysis: The carbon balance for the electrolysis was calculated as the ratio of the molar amount of carbon that was found in all formed products ($\sum n_{\mathrm{C},\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{s}}$ ) and the molar amount of carbon that was consumed from the substrate n‐hexanoic acid ($n_{\mathrm{C},\mathrm{h}\mathrm{e}\mathrm{x}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{c}\mathrm{i}\mathrm{d}}$ ) [Eq. (18)]. The consumed amount of carbon from n‐hexanoic acid is the difference between the gravimetrically determined molar amount before and the molar amount measured by GC–MS after the electrolysis.

$${\displaystyle \mathrm{c}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}\mathrm{b}\mathrm{a}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}=\frac{\sum n_{\mathrm{C},\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{s}}}{\Delta n_{\mathrm{C},\mathrm{h}\mathrm{e}\mathrm{x}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{c}\mathrm{i}\mathrm{d}}}\times 100\%}$$ (18)

The n‐decane yield, $Y_{\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{e}}$ , is given by Equation 19:

$${\displaystyle Y_{\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{e}}=\frac{2\times n_{\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{e}}}{\Delta n_{\mathrm{h}\mathrm{e}\mathrm{x}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{c}\mathrm{i}\mathrm{d}}}\times 100\%}$$ (19)

In order to assess the selectivity, $S_{\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{e}},$ of the electrolysis of n‐hexanoic acid to n‐decane, the molar amount of n‐decane was related to the amount of all products formed (excluding CO₂, H₂ and O₂) [Eq. 20]:

$${\displaystyle S_{\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{e}}=\frac{n_{\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{e}}}{\sum n_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{s}}}\times 100\%}$$ (20)

For determining the consumption rate ($r_i$ ) of n‐hexanoic acid and the production rate of n‐decane, the degraded or produced amount of respective substance ($\Delta n_i)$ was divided by the electrode surface area (A _{electrode surface}=2 cm²), the reaction volume (V=0.2 L) and the electrolysis time (t=4 h) [Eq. 21]:

$${\displaystyle r_i=\frac{\Delta n_i}{A_{\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{f}\mathrm{a}\mathrm{c}\mathrm{e}}\times V\times t}}$$ (21)

Statistical analysis: All experiments were done in at least three independent replicates (n≥3). In this regard independent replicates means that the n‐hexanoic acid solution, the following electrolysis of the solution, and finally the sample preparation were performed fully independent for each single replicate. All values are given as the mean±confidence interval (CI, α=0.05) if not stated otherwise.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

K. Neubert, M. Schmidt, F. Harnisch, ChemSusChem 2021, 14, 3097.