1 Introduction

The transition of the global energy system towards larger contributions from renewable sources requires the development of alternative concepts for energy storage. This is due to the intermittent character of both wind power and solar energy leading to large overproduction and shortage dependent on meteorological conditions. Hydrogen is considered to be a promising candidate to replace fossil fuels as an energy carrier [1], featuring an excellent gravimetric storage density of 33 kWh per kg H2 [2]. On the other hand, there is the poor volumetric energy density of hydrogen, which requires either liquefaction at cryogenic temperatures (20 K) or compression at high pressures (700 bar) [3]. Application of the “liquid organic hydrogen carrier” (LOHC) concept allows to circumvent the deficiencies arising from the exotic conditions of conventional storage technologies [2, 4–8]. The LOHC concept is based on the reversible catalytic conversion of hydrogen-lean to hydrogen-rich organic compounds through catalytic hydrogenation and dehydrogenation. The potential for enabling a decentralized energy storage system, that is, to decouple energy production and consumption, qualifies this technology to contribute immensely to the integration of renewable energies into our future energy mix.

The ideal LOHC system has to fulfill numerous requirements concerning storage density, safety, large-scale availability, and toxicity. Among the most promising LOHCs are aromatic or heteroaromatic compounds and their hydrogenated analogues [2, 4]. It was found that isomeric mixtures of benzyltoluenes/perhydro-benzyltoluenes and dibenzyltoluenes/perhydro-dibenzyltoluenes are particularly attractive candidates, offering an excellent storage capacity of 6.2 wt% H2, low toxicity, and a low melting point (−34 °C) [9]. Furthermore, they are easily available at a large scale, since the dehydrogenated compounds are well established heat transfer fluids in industry (typical tradenames are Marlotherm LH (isomeric mixture of benzyltoluenes) and Marlotherm SH (isomeric mixture of dibenzyltoluenes)) [9, 10].

Several aspects have to be considered to decide whether a LOHC couple is suitable for application, one of them being the thermal stability of the LOHC during long-term operation. Another important factor is the cost of the applied hydrogenation and dehydrogenation catalysts favoring systems with minimized noble-metal content and maximized productivity. The best performance is currently obtained by utilizing oxide-supported Ni or Ru catalysts for hydrogenation at elevated temperature (100–200 °C) and at high pressure (3–5 MPa) [10–12], while dehydrogenation is typically catalyzed by oxide-supported Pt or Pd catalysts at 150–300 °C near ambient pressure [7, 10]. In this letter, we address the differences in the mechanism and kinetics that are observed between catalytic dehydrogenation over Pd and Pt surfaces. To this end, we performed a model study on the dehydrogenation of heteroatom-free LOHC systems on Pd(111), which allows us to make direct comparison to a previous study on Pt(111) [13].

To facilitate such comparison, a detailed understanding of the underlying reaction mechanisms and microkinetics as a function of catalyst material is a key requirement. Here, model catalytic and surface science studies can provide fundamental insights into surface reactions at a molecular scale [14–17]. Studies on isomeric mixtures of benzyltoluenes using spectroscopic methods are hampered, however, by the chemical complexity of these compounds, e.g. the presence of different regioisomers in the sample. For this reason, we chose the LOHC pair diphenylmethane (DPM)/dicyclohexylmethane (DCHM) as a model system. The catalytic loading and unloading cycle involves 12 hydrogen atoms per LOHC molecule (see Fig. 1). In a recently published study, we explored the surface reactions of DCHM and DPM, as well as smaller fragments like toluene and methylcyclohexane on Pt(111) [13]. By combining different spectroscopic methods, we developed a detailed reaction mechanism for the dehydrogenation of DCHM and its side reactions. Stepwise dehydrogenation was found to occur between 200 and 360 K. The reaction is accompanied by decomposition via C-H bond scission at the methylene bridge activated at 360 K, followed by C–H and C–C bond scission at the aromatic rings above 450 K.

Fig. 1
figure 1

Schematic of the catalytic cycle of the investigated LOHC system diphenylmethane (DPM) and dicyclohexylmethane (DCHM)

Full size image

LOHC decomposition is an undesired process affecting the long-term performance of the hydrogen storage system and, therefore, needs to be minimized by selecting LOHC dehydrogenation catalysts that combine high activity with high selectivity for the aliphatic C-H bonds. Besides Pt, also Pd-based catalysts show reasonable performance in LOHC dehydrogenation [18–21]. Starting from these metals, future modification could involve the formation of alloys and/or bimetallic core/shell particles. In this work, we present the results of a mechanistic study on the surface reactions of DPM and DCHM on a Pd(111) single crystal under the same conditions as previously used on Pt(111). As in our previous study, we use synchrotron radiation-based in situ high resolution X-ray photoelectron spectroscopy (HR-XPS) and investigate both the energy-lean DPM and the energy-rich DCHM. This allows us to compare the stability of specific intermediates and to identify the differences in reactivity and selectivity between the two surfaces.

2 Results and Discussion

We start our analysis with adsorption experiments of DPM and DCHM on Pd(111) monitored by time-resolved XPS. In both experiments, the sample was kept at 170 K during adsorption from the gas phase while the C 1s region was continuously measured. We deposited approximately 5 L of DPM and 3 L of DCHM, respectively. The XP spectra recorded during adsorption are shown in Fig. 2.

Fig. 2
figure 2

(Top) C 1s (hν = 380 eV) XP spectra obtained during deposition of DPM (a) and DCHM (b) onto clean Pd(111) at 170 K. Blue and red colored peaks represent the chemisorbed monolayer of DPM and DCHM, respectively, green peaks are assigned to the physisorbed multilayer. (Bottom) quantitative analysis of the different contributions as a function of exposure

Full size image

Molecular adsorption of DPM (Fig. 2a) is confirmed by the growing C 1s peak at 284.2 eV, accompanied by two higher binding energy shoulders at 284.6 and 284.9 eV. Selected spectra and the respective fits are shown in Fig. 3a. The observed peak shape contains contributions from carbon atoms in DPM with different local environment (in the rings and in the methyl group) and with different distances to the surface; similar effects have been observed e.g. for ethylene, benzene and graphene on Ni(111) [16]. In addition, one also expects small contributions in this energy range, from vibrational excitations of C–H modes in the final state ion, with typical energies of ~400 meV [16, 22]. At higher exposures, the additional signal at 285.0 eV indicates adsorption of a physisorbed multilayer; while this signal increases, the monolayer signal is dampened, indicating layer-wise growth of DPM (see Fig. 2a, bottom). The small peak at 285.7 eV indicates minor amounts of CO, which are coadsorbed during DPM deposition.

Fig. 3
figure 3

Selected C 1s (hν = 380 eV) XP spectra obtained during adsorption and subsequent thermal treatment of DPM (a) and DCHM (b) on clean Pd(111). In addition, the fitted components are depicted (see text for details)

Full size image

Figure 2b shows the C 1s spectra recorded during adsorption of the fully hydrogenated molecule, DCHM, onto clean Pd(111). We observe one broad peak at 283.7 eV, with a shoulder at 284.2 eV, which grows simultaneously. We suggest that, similar to DPM, non-equivalent carbon atoms contribute to the observed line shape. In fact, the lack of π-interactions of DCHM with the substrate may lead to a weaker bonding and less defined adsorption geometry and, as a consequence, the carbon atoms experience an even more heterogeneous environment as compared to DPM. This effect is also indicated by larger and broader peaks for DCHM. In the multilayer regime DCHM shows a peak at 284.0 eV. Notably, in contrast to DPM, multilayer growth of DCHM starts before completion of the monolayer.

In the Temperature Programmed XPS (TPXPS) experiments, the surface reaction was followed by continuously measuring spectra while heating the crystal with a constant rate of 0.5 Ks−1, from 140 to 550 K for DPM, and from 150 to 550 K for DCHM. The spectra are depicted in Fig. 4a and b, respectively. In addition, we show the integrated areas of all components fitted as a function of temperature. Note that the end points of the adsorption experiments in Fig. 2 are the starting point of the TPXPS experiments in Fig. 4.

Fig. 4
figure 4

(Top) Thermal evolution of the C 1s (hν = 380 eV) region for DPM (a) and DCHM (b) on clean Pd(111) up to 550 K (see text for detailed peak assignments). (Bottom) quantitative analysis of the different peaks as a function of temperature

Full size image

First, we discuss the evolution of DPM. Above 240 K, we observe the first spectral changes, namely the decrease of the signal at 285.0 eV, indicating desorption of the physisorbed multilayer. During this process, the dampening of the monolayer signal decreases, as reflected by the increase of the peaks at 284.2, 284.6, and 284.9 eV. Above 200 K, the envelope of the three peaks attributed to DPM shift to higher binding energy by approximately 0.1 eV. Concomitantly, the characteristic structure of the two higher binding energy shoulders transforms into a single shoulder, indicating a surface reaction to a new species. This new species, which we refer to as “DPM II” is fitted by two peaks at 284.3 and 284.8 eV, respectively; see Fig. 3a. We will go into more detail on the nature of this species in the following discussion. The transformation of DPM to DPM II is completed at 340 K. It is superimposed by desorption of the DPM multilayer until ~280 K.

To obtain first insight into the reaction mechanism, we compared the results obtained here with our previous study of DPM on Pt(111). There, DPM undergoes dehydrogenation at the methylene bridge at 260 K, which is accompanied by a change in adsorption geometry of the molecules on the Pt(111) surface [13]. This process was reflected by a spectral feature growing at the lower binding energy side of the envelope of the peaks, which are characteristic of DPM. At significantly higher temperatures (450 K), further transformations take place, which we assign to C–H abstraction at the phenyl rings. The latter process is reflected by two growing contributions at 283.9 and 284.8 eV. For other LOHCs, such as dodecahydro-N-ethylcarbazole, we found that the dehydrogenation and decomposition mechanisms over Pt(111) [23, 24] and Pd(111) [18] are qualitatively similar, but the temperature windows of individual steps may differ substantially.

Following these lines, we transfer the mechanism proposed for Pt(111) to Pd(111), and deduce tentative conclusions on the nature of the DPM II species. The comparison of the spectrum of intact DPM at the onset of multilayer growth (Fig. 2a) with the spectrum at 350 K during heating (Figs. 3, 4a), where apart from small amounts of CO DPM II is the only species on the surface, reveals very similar signal intensities, that is, no loss of carbon on the surface. From this we can exclude that the different envelopes of DPM and DPM II are due to reorientation effects enabled by partial DPM desorption. Consequently, we propose that DPM undergoes some kind of reaction starting at 200 K. However, since the spectral changes are rather subtle (see Fig. 4a), we believe that the species formed at 200 K is still very similar to DPM. One possible and likely explanation is the dehydrogenation at the methylene group, which was found to be the first reaction step on Pt(111). This assumption is in line with the disappearance of the shoulder at 284.9 eV upon thermal treatment, assuming that the C 1s signal of the CH2 subunit of DPM may be located in this spectral region.

Above 360 K, we observe a continuous broadening of the signal. New spectral features become visible on the lower binding energy side (283.7 eV) and simultaneously at 284.2 and 284.4 eV (Figs. 3, 4a), indicating a chemical transformation of the molecule. An unambiguous separation of the decomposition products and DPM II is not possible in this temperature regime. The decomposition products can best be fitted by a combination of three peaks at 283.7, 284.2, and 284.4 eV (Fig. 3a), and we use their envelope to describe the proceeding decomposition reaction as a whole. Using this approach, all spectra between 350 and 550 K were fitted by a linear combination of the envelopes of DPM II and decomposition products. At 450 K, we observe that both the decrease of the DPM II signal and the increase of the decomposition signal slow down, which is reflected by a plateau in the integrated areas (see Fig. 3a, bottom) and then increases again above 500 K. At 550 K, DPM II is almost completely transformed. Note that due to the overlapping spectral features some uncertainty remains and the DPM II signal may also include further decomposed reaction intermediates above 450 K.

Interestingly, the overall amount of carbon does not significantly decrease until 550 K, which indicates that the Pd surface remains covered with carbonaceous species of constant concentration up to this temperature. It is known that the formation of atomic carbon is the rate-limiting step for carbon migration into the bulk in this temperature range [25]. This suggests that the ring structure of the molecule still is at least partly intact at this temperature and the observed changes are mostly due to the ongoing dehydrogenation. Note that this is also the temperature region under which the dehydrogenation is typically performed on real catalysts [7, 10]. The stability of carbonaceous species under these conditions suggests that they play an important role as reactivity modifiers. Between 600 and 800 K (not shown) the peak intensity drops to zero, indicating breakup of the ring structures and rapid dissolution of the atomic carbon into the bulk.

When comparing the observation on Pd(111) to the results obtained on Pt(111), we find drastic differences of the temperature regimes in which specific reactions steps are observed. The onset of the first decomposition step of DPM on Pd(111), which is likely to take place at the methylene bridge, is shifted to lower temperature by approximately 60 K, from 260 to 200 K. This observation is in line with previous studies, where Pd(111) showed a lower activation barrier for C-H bond scission than Pt(111) [18, 24].

Next, we address the thermal evolution of the hydrogen-rich carrier DCHM on Pd(111) (Figs. 3, 4b). Upon heating, we first observe multilayer desorption, as indicated by a decrease of the broad signal at 284.0 eV. At 260 K, multilayer desorption is completed. Between 190 and 330 K, we find a continuous shift of the DCHM monolayer peaks from 283.7 and 284.1 eV to approximately 284.3 and 284.7 eV, respectively. At the same time, the signals transform into a narrower peak shape. Our previous studies of DCHM on Pt(111) showed that the first dehydrogenation intermediate of DCHM is observed already at 200 K: [13] by abstraction of three hydrogen atoms on each cyclohexyl ring a “double π-allylic” species is formed, whose CH units are lying flat on the surface [13]. The spectral changes, that are, the shift to higher binding energies and the decreasing peak width in our present experiments on Pd(111) strikingly resemble the observations on Pt(111). This leads us to the conclusion that the mechanism of this initial dehydrogenation step is very similar on both surfaces. In Fig. 4b, the reaction products are represented by an envelope of two peaks at 284.1 and 284.5 eV, which begin to grow at 190 K at the cost of the monolayer signals (Fig. 4b, bottom).

At 280 K, the signals attributed to the double π-allyl have reached a maximum in intensity, while the DCHM monolayer signals strongly decrease, indicating that most DCHM molecules underwent partial dehydrogenation until this point. At higher temperatures, a further shift to higher binding energy and narrowing of the signals indicate formation of another reaction intermediate. At 350 K, the spectral shape agrees almost perfectly with the spectrum at 350 K of the DPM/Pd(111) experiment. This resemblance is illustrated in Fig. 3, where we compare selected spectra of DPM and DCHM at different temperatures. Both spectra recorded at 350 K are fitted with identical parameters, which lets us assume that at this temperature dehydrogenation of the cyclohexyl rings is completed and, simultaneously, the same dehydrogenation of the methyl bridge as observed for DPM/Pd(111) has occurred. Notably, we cannot rule out that the latter reaction occurs already at lower temperatures, since, on the basis of the available data, the assumed transition at the methylene bridge is not unambiguously separable from other reactions.

Interestingly, the total surface coverage (sum signal in Fig. 4) after multilayer desorption of DCHM, that is, at 280 K is very similar to that after multilayer desorption of DPM, but it decreases by ~30 % until 350 K. This behavior is attributed to the large additional amount of coadsorbed hydrogen formed through dehydrogenation of DCHM, which partly leads to desorption of e.g. rehydrogenated DPM II, unreacted DCHM or other intermediates.

We conclude that, starting at 250 K, the partially dehydrogenated intermediate undergoes further dehydrogenation to form DPM II, that is, DPM with the methyl bridge dehydrogenated. At 350 K, dehydrogenation to DPM II is completed. On Pt(111), we were able to identify an intermediate step of this reaction, namely the formation of a single π-allylic species with one ring still being partially dehydrogenated and the other being completely dehydrogenated; this intermediate is formed at 260 K and is stable up to 350 K [13]. In contrast to these findings, we could not identify such an intermediate on Pd(111). We may rationalize this observation by the fact that C-H bond scission is more rapid on Pd(111). In our previous studies on dodecahydro-N-ethylcarbazole, for instance, we have shown that the first dehydrogenation intermediate is formed at 270 K on Pt(111), while the same intermediate is formed already at 220 K on Pd(111) [18, 24]. The reason for this difference between the two metals is found in the differences in electronic structure, whereas the details of the mechanism may be intricate. For instance, the lower-lying d-bands of Pd lead to a weaker interaction with the π- and π*-orbitals of the LOHC and a weaker binding to the surface [26–28]. As some of the aliphatic parts of the allylic intermediate are not in direct contact with the surface, substantial reorientation is required to activate hydrogen at the remaining CH2 groups. Such a reorientation is facilitated on Pd due to the weaker bond to the surface. As a result, both allyl entities are rapidly dehydrogenated to phenyl rings instead of the formation of a stable single π-allylic species that is observed on Pt(111). Similar observations were reported by Hunka et al., who compared the dehydrogenation of cyclohexene to benzene on Pd and Pt single crystals [26].

Between 350 and 400 K, we do not observe any substantial changes on the Pd surface. Starting at 450 K, we identify a slight broadening of the DPM II signals on the lower binding energy side, which becomes more intense upon further thermal treatment. As discussed before, we attribute these changes to the dehydrogenation of the phenyl rings of DPM II. In fact, the spectral shape allows us to describe the formation of the decomposition products by utilizing the same envelope as in the DPM TPXPS experiment, including three components at 283.7, 284.2, and 284.4 eV (Fig. 4a). The development of the spectral shapes between 350 and 550 K is shown in Fig. 4. Qualitatively, the development is similar for DPM and DCHM, but we observe that the onset of decomposition is shifted to higher temperature by approximately 90 K for DCHM, and the integrated peak intensity remains nearly constant at a total coverage that is ~30 % lower than for DPM. These differences are noteworthy as the reactions start from very similar total coverages after multilayer desorption. The difference may be associated with the different initial DPM II coverages in in the two experiments, or with the dehydrogenation that occurs for DCHM prior to decomposition. The main desorption peak of hydrogen from clean Pd(111) is located at approximately 300–400 K, depending on the H2 coverage [29]. The presence of coadsorbates, however, may influence the desorption considerable. Coadsorbed CO, for instance, leads to a very broad hydrogen desorption pattern due to trapping of hydrogen within CO islands on the surface [30]. In our case, the high surface coverage of DPM may hinder recombination of atomic hydrogen, which is the rate-determining step for desorption [31]. Hydrogen present on the surface at higher temperatures could either block adsorption sites and thus impede further C-H bond scission or modify directly the adsorption geometry of DPM. It is also possible that a higher density of hydrogen will persist in the subsurface area of Pd [25, 32, 33], thereby modifying the electronic properties and the activation of the C–H bonds of DPM. In future experiments it will, therefore, be essential to probe the influence of preadsorbed hydrogen on the stability of the reaction intermediates systematically and up to high pressure conditions.

Above 480 K, formation of decomposition products becomes faster, which, according to the above discussion, may be associated with the progressing loss of coadsorbates, in specific hydrogen. In the temperature region above 600 K, where atomic carbon is formed, we observe a similar behavior as for DPM/Pd(111), namely the loss of carbon which is limited by diffusion of carbon atoms into the bulk.

3 Conclusions

We have studied the adsorption and reaction of the LOHC couple DPM and DCHM on Pd(111) by HR-XPS. We have performed this study analogous to our previous investigation on Pt(111), which allows us to identify similarities and differences in the dehydrogenation mechanism and kinetics for the two noble metal surfaces. The dehydrogenation and degradation mechanisms of DCHM on Pt(111) and Pd(111) are summarized in Fig. 5.

Fig. 5
figure 5

Schematic reaction mechanisms of DCHM on Pt(111) (green arrows) and Pd(111) (blue arrows). For the sake of clarity, the illustrations do not contain the information on the adsorption geometries discussed in the text

Full size image

From Pd(111), the multilayer of DPM desorbs until 280 K, leaving behind a chemisorbed DPM monolayer. Above 200 K, the DPM signals undergo slight changes, which we associate with C-H bond scission at the methylene bridge, and dehydrogenation of the phenyl rings starts at 360 K. This behavior is in contrast to the one found for Pt(111), where the methylene bridge undergoes dehydrogenation only at 260 K, and dehydrogenation of the phenyl rings does not occur before 450 K. We attribute this difference to a higher propensity of Pd(111) to cleave C–H bonds.

The reaction mechanism of adsorbed DCHM on Pd(111) is qualitatively similar as on Pt(111) up to approximately 260 K. On Pd(111) at and above 190 K, the first stable dehydrogenation intermediate is formed, which we assign to a double π-allylic structure. In this intermediate, three hydrogen atoms are abstracted from each six-membered ring. On Pt(111), this species is formed at slightly higher temperatures (200 K), in line with our previous findings that C–H bond scission is more rapid on Pd(111) than on Pt(111). Until 280 K, on both surfaces, a large ratio of chemisorbed DCHM is transformed into this partially dehydrogenated species. At 260 K, further dehydrogenation begins to take place, eventually leading to formation of the same species we observed in the DPM experiment, namely DPM which is dehydrogenated at the methylene bridge (“DPM II”). On Pt(111), we identified an intermediate at 260 K which we assigned as a single π-allyl. The latter is formed through dehydrogenation of one π-allyl entity to a phenyl ring. On Pd(111), this single π-allylic species is not observed, which, again, indicates a lower activation barrier for C–H bond scission for this surface. At 350 K, DPM II and hydrogen are the only species present on the Pd surface. On Pt(111), further dehydrogenation of the remaining π-allyl starts at 350 K and is accompanied by dehydrogenation of the methylene bridge, thereby also yielding the DPM II species. Thus, we conclude that no intact DPM is formed at any time neither on Pt(111) nor on Pd(111).

On Pd(111), DPM II formed by dehydrogenation of DCHM shows an enhanced stability as compared to its molecular counterpart formed directly by molecular adsorption of DPM from the gas phase. Specifically, decomposition sets in only above 450 K and the final carbon coverage is smaller than in case of decomposition of DPM. The effect is attributed to coadsorbed hydrogen formed through dehydrogenation.

We conclude that Pd and Pt in general show similar dehydrogenation pathways and mechanisms, but the stability of the surface intermediates is pronouncedly different for the two surfaces. In addition, our findings suggest that coadsorbates, that is, hydrogen from dehydrogenation of DCHM to DPM II, play a particularly important role on Pd, having a direct influence on the stability of the partially dehydrogenated surface intermediates. In future model studies, it will be essential to study the influence of these two coadsorbates on the stability and mechanism in situ up to realistic pressure conditions.

4 Experimental

The HR-XPS experiments were performed with a transportable UHV setup (base pressure in the low 10−10 mbar range) at the 3rd generation synchrotron BESSY II at Helmholtz-Zentrum Berlin, beamline U49/2-PGM1. The experimental setup consists of an analysis chamber and a preparation chamber, separated by a gate valve. The analysis chamber is equipped with an electron energy analyzer (Omicron EA 125 HR U7), which lies in the plane of the synchrotron ring at a horizontal angle of 50° with respect to the incoming synchrotron light. The preparation chamber is equipped with a low energy electron diffraction (LEED) optics (ErLEED, Specs), an ion sputter gun (Specs IQE 11/35) for sample cleaning, and gas dosing facilities. For more details on this transportable HR-XPS setup see ref [34].

The Pd(111) single crystal (MaTecK GmbH) was cleaned by several cycles of Ar+ sputtering (E = 1 keV, 5 × 10−6 mbar Ar, 300 K), annealing to 1200 K in vacuum and exposure to oxygen (5 × 10−7 mbar O2 at 1200 K followed by cooling down in O2 to 600 K) [35–37]. The quality of the crystal was controlled by low-energy electron diffraction (LEED) and XPS. DPM and DCHM were deposited by physical vapor deposition (PVD) using a differentially pumped doser, which is separable from the main chamber by a gate valve and may be baked out separately. During evaporation, the background pressure in the analysis chamber rises to approximately 4.0 × 10−9 mbar. In order to avoid radiation induced damage of the surface, the sample was shifted vertically by 0.05 mm prior to each measurement. The C 1s spectra were acquired at a photon energy of 380 eV with an overall resolution of approximately 180 meV at normal emission (acquisition time: typically ~15 s/spectrum). All spectra were referenced to the Fermi edge and treated by subtracting linear backgrounds. The peak components were fitted with asymmetric Doniach–Sunjic line profiles [38] convoluted with a Gaussian function [39, 40].

The numbers and positions of the peaks within the envelopes describing the surface species were evaluated on the basis of our previous publication of DPM, DCHM, and their molecular building blocks on Pt(111) [13] and on literature addressing photoelectron spectroscopic studies of hydrocarbons on Pd(111) (see Refs. [18, 19, 41] and references therein).

Diphenylmethane was purchased from Sigma Aldrich (≥99 %). The hydrogenation of diphenylmethane to dicyclohexylmethane was carried out using a 300 mL stainless steel Parr batch-autoclave equipped with a four blade gas-entrainment stirrer (1200 rpm). The volume of 150 mL unloaded LOHC was placed in the pressure vessel and a constant molar ratio of 400/1 (LOHC/catalyst metal) of the catalyst (Ru/C, 5 wt%) was added. To ensure an inert atmosphere, the gas volume of the reactor was replaced three times with Argon (4.6, Rießner Gase) by flushing the reactor. After heating up the LOHC to the desired reaction temperature (120 °C) using an external electrical heating jacket, the autoclave was pressurized with 10 bar of hydrogen 5.0 (Linde) and this pressure was kept constant during the experiment by continuous dosing of hydrogen.