Electrochemical production of sustainable hydrocarbon fuels from CO 2 co-electrolysis in eutectic molten melts

Due to the heavy reliance of people on the limited fossil fuel as energy resources, global warming has increased to severe levels due to huge CO 2 emission into the atmosphere. To mitigate this situation, a green method is presented here for the conversion of CO 2 /H 2 O into sustainable hydrocarbon fuels via electrolysis in eutectic molten salts ((KCl-LiCl; 41:59 mol%), (LiOH-NaOH; 27:73 mol%), (KOH-NaOH; 50:50 mol%), (Li 2 CO 3 -Na 2 CO 3 -K 2 CO 3 ; 43.5:31.5:25 mol%)) at the conditions of 1.5–2 V and 225–475 o C depending upon molten electrolyte used. Gas chromatography (GC) and GC-MS techniques were employed to analyse the content of gaseous products. The electrolysis results in hydrocarbon production with maximum 59.30, 87.70 and 99% faraday efficiency in case of molten chloride, molten hydroxide and molten carbonate electrolytes under the temperature of 375, 275 and 425 o C


Introduction
Over the past few decades, two major issues have captured the attention of scientists and policymakers: global warming due to the increasing levels of carbon dioxide gas (CO2) in the atmosphere, and the rapid depletion of fossil fuels as an energy resource.To tackle these complications, two solutions were proposed 1,2 .The first solution is the use of renewable energy resources such as wind, solar, nuclear or geothermal energy to minimize the greenhouse gases' emission.While the second solution is the consumption of CO2 to remove its excessive concentration from the atmosphere and to enhance energy resources by converting it into hydrocarbon fuels 3,4 .Renewable energy resources do not involve CO2 sequestration 5 .So to tackle CO2 emissions 6 , it was considered preferable to introduce some of the renewable energy resources into an existing energy infrastructure as a "drop-in" form of energy.Examples of this can include the synthesis of fuels or fertilizers from CO2 or biomass 7,8 .higher production costs, lower durability and high energy utilization, became the reason for their rejection on industrial scale implementation 22 .Molten salts exhibit the same chemistry regarding CO2 and H2O reduction as in SOEC except that CO2 can be also reduced to carbon 23 in addition to carbon monoxide depending on the operating conditions.Which can thereby affect the products.Deposited carbon on cathode can facilitate the formation of different kinds of hydrocarbons (rather than CO) in case of molten salt electrolysis.Because as soon as the fresh carbon deposit on cathode it reacts immediately with hydrogen gas produced via water reduction on the metal cathode surface itself, resulting in the formation of hydrocarbons 24 .
Molten salts are preferred over solid oxides regarding CO2-H2O co-electrolysis for a variety of reasons.Besides a wide electrochemical window, high electric conductivity, relatively low cost, reactivity with CO2 and no need of specific electrode materials (Ni-YSZ, La1-x SrxMnO3/YSZ) make them suitable candidates for this process.Moreover, the possibility of carbon or CO hydrogenation after electrolysis in molten salts is much more significant 25,26 .
Molten salts with some limitations such as slight corrosion activity particularly at high temperature and relatively high energy utilisation to maintain the heat for molten salt to avoid the solidification 27 , can still be employed to produce hydrocarbon gas or liquid fuels 28,29 .
This study systematically investigates the hydrocarbon fuel production by employing CO2-H2O co-electrolysis by using different types of molten electrolytes: molten chloride (LiCl-KCl; 58.5:41.5 mol%), molten hydroxide ((LiOH-NaOH; 27:73 mol%), (KOH-NaOH; 50:50 mol%)) and molten carbonate (Li2CO3-Na2CO3-K2CO3; 43.5:31.5:25mol%) at variable conditions of temperature and voltage depending upon the molten salt.Two feed gas insertion methods are also employed; gas flowing over the electrolyte surface (GFOE) and gas flowing inside the electrolyte (GFIE).The effect of different variables including; faradays efficiency, energy consumption and heating values at variable conditions of temperature and voltage are studied.Moreover, the product formation by electrolysis is confirmed by GC (using FID and TCD detectors) and GC-MS.

Electrochemical performance measurement
The electrochemical processes were performed by using four combinations of molten salts as electrolytes.The key point of choosing these combinations of molten salts is due to their low melting points and the ability to work at operating temperatures, as low as possible, keeping Then electrolytes were poured into a crucible present inside a corrosion resistant electrolyser's retort.Which was built in house with a flange type cover using the 316-grade stainless steel to shape the reactor to provide and control the environment needed for the molten salt electrolysis.
The dimensions of the retort were 130 mm internal diameter, 7.5 mm wall thickness and 800 mm vertical length.On the flange cover, there were some holes drilled for the insertion of ceramic tubes (for anode and cathode gas collection), observation purposes and sealing.The retort was inserted centrally in the furnace.In the retort, a stainless-steel stand was placed and a refractory brick mounted on it and above its alumina crucible containing pre-melted molten salt electrolyte (about 100 g) was placed.The two-electrode mode experimental set-up used here is a continuous system with small scale for electrolysis and hydrocarbon production.The electrolysis is conducted by using titanium metal (Purity: 99.99%, Good fellow Cambridge Ltd) as cathode and graphite (Purity: 99.99%, Advent Research Materials) as anode 37,38 .
A small rate of gas (CO2; 48.4%, + H2O; 3.2% + Ar; 48.4%) flows continuously inside the reactor where hydrocarbons and O2 gases are produced inside the molten salts on the cathode and anode surfaces respectively during the electrolysis.And the products are collected at different time intervals.This process is done by employing the Agilent E3633A 20A/10V Auto-Ranging DC Power Supply and a laptop with an EXCEL add-in to collect the instrumentation data.Two electrode tube gas outlets were present, each connected to another Dreschel bottle containing the mineral oil to observe the outlet gases produced and reduce electrolyte contamination.Gas product samples were collected using a tedler 1 L (SKC Ltd.) gas bag via a connection from the cathodic gas tube.The electrolyser setup is a modified form of previously used setups 39 .The schematic representation of experimental setup is shown in Fig. 1 To avoid their mixing, the argon gas was used which pushed the gaseous products into their respective bags.Moreover, the study is carried out with two modes of feed gas insertion inside the reactor for each electrolyte; gas flowing over the electrolyte surface (GFOE) and gas flowing inside the electrolyte (GFIE) for the comparison of hydrocarbon production in both cases.The first method (GFOE) has been used to minimize the chances of solid material's production (carbon, carbon nanotubes, graphene, carbonates solidification) 35,40 and to produce gaseous hydrocarbon products preferably.The current efficiency was calculated from the Eqs. (1-2): where Qx is the charge required for the amount of individual product produced, n is the number of electrons required, F is the charge of one electron which equals 96485 col and QT is the total charge calculated from the area under the current vs time curve.
The remaining composition of both gas standards was balanced with helium gas.The GC graphs for different calibration gas standards are shown in Fig. 2 for the comparison of electrolysis gaseous products.Furthermore, the samples were analysed by a different sophisticated GC instrument (Agilent 7890B) attached with a mass spectrometer (JEOL AccuTOF GCX) for longer chain hydrocarbons detection.Gas detecting tubes from GASTEC (ai-cbss Ltd.) were used to analyse the feed gas compositions for CO2 and H2O contents.The feed gas composition with CO2 (48.4%) + H2O (3.2%) + Ar (48.4%) was kept same for all the experiments.GASTEC 2HH is characterised to detect the higher contents of CO2 from 5 to 40% of the feed gas, with the change in colour from orange to yellow.
The GASTEC30 tube can analyse water content in the range of 0-18 mg/L, and it contains Mg(ClO4)2.However, the colour here will change from yellowish green to purple.After the analysis of cathodic gas sample, the concentration of each gas compound (Mgas) was calculated as followed by the Eq.(3): where Mgas is the concentration (%) of individual gas in the sample.Ms is the concentration (%) of the specific standard gas in the sample.Agas is the area under the peak resulting from the FID analysis for the individual gas (C2-C5) in the sample.As is the area under the peak resulting from the analysis of the specific gas standard (CH4).F i ̅ is the response factor for each gas.

Optimization of electrolytes
The selection of the molten salt is done based on the ability to generate hydrocarbon fuels from the co-reduction of CO2 and H2O (Eq.(4)).The combination of a hydrocarbon molecule starts ideally from the two known element sources: carbon (C) and hydrogen (H).Both of these elements can be effectively formed from electrochemical conversion via an appropriate molten salt.
Generally, the presence of moisture with CO2 gas in molten salt experiments is the basis for generating H2 and CH4 during electrolysis in most cases and provide feasibility to the reactions 41, 42 .

Molten chloride electrolyte
The attractive characteristic in the molten chloride case is the probability of producing CO or C directly from CO2 reduction in the presence or absence of a carbonate ion 42 .Carbonate ions (if added externally) are used as an important additive to molten chlorides to provide the oxide ions required for performing CO2 reduction 43,44 .For absorbing more CO2 gas into the molten salt leading to increase in product yield from electrolysis, the addition of oxides or carbonate salts into the molten chloride is considered preferable.But one drawback exhibited by this process was the increase in applied voltage and working temperature of resulting molten salt mixture.Which is not the favourable condition for hydrocarbon production 45 .So to tackle this problem, molten chloride electrolyte is used in this study for hydrocarbon production without any externally added oxide or carbonate salts.In the absence of H2O and carbonate ions, the reduction of CO2 to carbon or CO can be done in several steps as seen from the Eqs.(5-7) 32 . (5) In the presence of steam beside CO2 in the feed gas, the reduction of CO2 becomes more feasible as CO2 can react with hydroxide ions released from the primary reduction of H2O through to the one-electron transfer reaction.Moreover, carbonate ions can be generated even from molten chloride through the reaction of CO2 with oxide ions emitted in turn after the direct reduction of H2O to H2 gas (Eq.(9)) 46 .The carbonate ions can then be electro-reduced in turn to carbon or CO and produce hydrocarbon by reacting with H2.The overall reaction occurring at electrodes can be summarised from the Eqs.(11-13).
Overall reaction At cathode: At anode: It is preferable to perform CO2-H2O co-electrolysis at temperatures even lower than the 400 o C to form the hydrocarbons feasibly.For that purpose, electrolysis was performed at 375 o C and 2V using molten chloride (LiCl-KCl; 41: 59 mol%) with two modes of gas insertion: GFOE and GFIE.The feed gas in GFOE mode containing H2O, CO2 and Ar, was kept flowing over the LiCl-KCl (41: 59 mol%) at 1.3 bar.The feed gas pressure was applied slightly over 1 atm to increase CO2 activity and thus the opportunity of improving reduction inside the molten chloride.
The same experiment was performed at the above conditions using a feed gas containing H2O with no CO2.Both experiments in GFOE mode are performed to see the solubility of CO2 and thus its activity inside molten chlorides.Carrying out electrolysis at 2 V and 375 o C, it can be seen from Fig. 3 that there is a small difference between the two current curves resulting from electrolysis in both cases.However, the current was still relatively high in both cases and gradually decreased with time.The decline in current can be imputed generally to the drop of oxidant concentration (H2O, CO2) that is reduced on the cathode surface due to the accumulation of new products (such as H or H2 bubbles) as there is no renewal action on the cathode surface during electrolysis.Also, it can be noted that the current was slightly higher in the case where CO2 gas was absent basically due to the obstruction of CO2 gas against H2O reduction on the electrode, particularly at high pressures through the possible reduction of CO2 to CO 2 2− (Eq.( 5)).
Despite some spikes noticed in the red curve in Fig. 3, it can be seen that the drop of the curves in both cases was quite the same confirming the weak effect of CO2 inside the molten chloride in GFOE mode.So due to CO2 weak effect, the hydrocarbon could not be produced in this case (GFOE mode).The GFIE mode of gas feed introduction was chosen as an appropriate way to increase CO2 concentration and solubility (and reactivity) inside the molten chloride and collect the maximum rates of hydrocarbon products at atmospheric pressure.The rates of H2 and CH4 production, collected from the cathodic tube, changed significantly after the first 30 and 60 min of electrolysis due to the process of carbonate ions formation as can be seen by the comparison of Fig. 4(a) and (b).Where the higher production rates of CH4 (0.67 µmol/h cm 2 ) and H2 (32.00 µmol/h cm 2 ) with higher faraday efficiency (59.30%) were found after the first 30 min of electrolysis (Fig. 4(a)).While the lower production rates of CH4 (0.39 µmol/h cm 2 ) and H2 (19.10 µmol/h cm 2 ) with faraday efficiency (30.50%) were obtained after 60 min of electrolysis (Fig. 4(b)) in molten chloride (LiCl-KCl; 41: 59 mol%).
The lower faraday efficiency (30.50%) was attributed due to the higher CO 3 2− ion formation leading to the subsequent conversion to C or CO with more energy consumption.The formation of a carbonate ion can be justified due to the reaction of CO2 with OH -generated in the molten chloride after the persistent reduction of H2O as stated previously in Eq. ( 10) 46 .It is interesting to note that there is a clear increasing trend of CH4 production in both Fig. 4(a) and (b) at a lower current density of 20 mA/cm 2 , which starts dropping off beyond this limit.The increase in current density affects the products content.With the current density increase, the CH4 production reached to an optimal value.After that further rise in current density results in adverse effects on CH4 production, greatly exceeding the minimum energy requirement of H2 production that keeps CH4 production at a lower level 24,36 .Deng et al. 31 stated that LiCl-KCl electrolyte containing Li2CO3/CaCO3 showed highest current efficiency of 80-85% at the current density of 25 mA/cm 2 , which dropped off by increasing current density for the conversion of CO2 to carbon.
Comparing results for the two occasions as two gas samples were taken after 30 and 60 min, it can be noted that the production rates of both gases (CH4 and H2) were higher in the first sample after first 30 min of electrolysis as the electro-reduction of the carbonate ions (to carbon for instance) had not commenced yet.Thus, the reduction of H2O to H2 was not significantly affected.The reaction of CO2 with OH -can be confirmed in the molten chloride for the second sample as the concentration of CO2 reduced from 34.80 to 4.80% (Table 1).The hydrocarbon production confirmed through GC analysis (with FID and TCD detectors) is shown in Fig. 5 where the FID signals are showing the production of methane with the peak at 2.11 retention time while TCD signals are clearly representing the peaks of H2, O2 and CO2.No CO can be detected in the molten chloride in both cases.
Thus, the best product concentrations are obtained in case of LiCl-KCl (41: 59 mol%) electrolyte from GFIE mode at the first 30 min of electrolysis rather than prolonged electrolysis (60 min).This is because of the formation of carbonate ions in case of prolonged electrolysis, which are reduced to the C or CO gases with the consumption of more energy (Table 1).Ijije et al. 47 reported the CO2 conversion into carbon films or CO in LiCl-KCl-CaCl2-CaCO3 molten salt at 520 o C. Similarly, the absorption and conversion of CO2 was also employed in molten chloride electrolytes (CaCl2-CaO and LiCl-Li2O) at 900 and 650 o C respectively 48 .Jianbang et al. 30 has converted CO2 by electrolysis in LiCl molten salt at 650 o C.

Molten hydroxide electrolyte
The molten hydroxide salt is preferred in the case of hydrogen production leading to hydrocarbons formation.Hydrocarbon molecules can be formed basically through a H2 reaction with either C or CO as the same mechanism for molten chlorides 49 .In most experiments using molten hydroxides, the conversion of CO2 was very high but the hydrocarbon yields were still low.This can be attributed generally to the reaction of CO2 with hydroxide ions 50 .
Therefore, CO2 must be diluted to lower concentrations by mixing with argon gas before introduction to the electrolyte, as this action can help to reduce the reactivity of CO2 with the salt, driving reaction (Eq.( 14)) to the left side.The formation of carbonate ions need to be reduced to provide enough time for the prospect of electro-reduction during electrolysis.
However, CH4 gas can be formed by another way in case of molten hydroxide electrolysis (Eq. (15)) 51 .
Thus, the abundance of hydrogen gas from rapid H2O reduction in molten hydroxides can contribute towards driving reaction (Eq.indicate a distinct variation in the production rates due to the variation in gas feed modes.This outcome can be attributed to the weak reduction of CO2 in the salt in GFOE mode.The hydrocarbon production rate was significantly improved when the feed gas insertion method was changed from GFOE to GFIE.It can be seen from the Fig. 6 (a) to (b) that the CH4 rate increased largely from 1.02 to 6.12 μmol/h cm 2 by moving from GFOE to GFIE mode as CO2 was promoted to dissolve in the salt.
Therefore, the prospect of direct reduction of CO2 to CO 2 2− and CO can occur in the LiOH-NaOH salt.At the same time, the H2 rate decreased from 1142.80 µmol/h cm 2 to just 185.00 µmol/h cm 2 , confirming the possible transformation of CO2 or CO to hydrocarbons.
Nevertheless, high faraday efficiency (87.70%) in the GFOE mode rather than (15.00%) in the GFIE mode was due to the higher H2 production rate.On the other hand, low faraday efficiencies in GFIE mode were obviously because of their lower production values from the slow reduction of CO2 to CO compared with rapid H2 production.Moreover, the optimal current density range found for hydrocarbon production in case of LiOH-NaOH salt was 80-85 mA/cm 2 .
Hydrocarbon production inside the molten hydroxide can be confirmed actually by the existence of CO fuel with the cathode gas product.CO can be formed from CO2 reduction as in molten chloride experiments.But the scarcity of CO gas found in the cathodic products in both electrolytes can be interpreted due to (1) a lack of CO2 direct reduction to CO but the formation of CH4 occurs by the reaction of CO2 with excess H2 and (2) the produced amount of CO during electrolysis in all cases was too little as CO can rapidly react with excess H2 to produce CH4.The formation of gaseous product (CH4) was confirmed from GC analysis with FID detector while H2, O2 and CO2 were confirmed by TCD detectors for the GFIE mode (Fig. 7).And obtained values are presented in Table 2.The presence of very small peak of CO in Fig. 7 (b) is providing the indication of higher methane production rates than molten chloride case.
As the GFIE mode provided higher production values of methane in case of molten hydroxide so the experiment was repeated using KOH-NaOH (50:50 mol%) due to its low working temperature, under the conditions of 2V applied cell voltage and 225 o C with GFIE mode only.
Although the temperature used here was slightly lower than 275 o C as used for LiOH-NaOH (27: 73 mol%) molten salt but the production rates of H2 (164.70 µmol/h cm 2 ) and CH4 (6.12 µmol/h cm 2 ) with faradaic efficiencies (17.90%) were almost same (Fig. 6 (c)).Moreover, the composition and concentration (vol%) of other cathodic product gases were also same (Table 2).But one limiting factor was the lower resulting current in the case of KOH-NaOH (50:50 mol%) molten salt than LiOH-NaOH (27: 73 mol%) (Fig. 8).Moreover, the potentials for carbon deposition or carbon monoxide evolution are more positive than the deposition potentials of Li metal for the case of LiOH.
In contrast, in the case of KOH, the potential for the formation of C or CO is more negative than the deposition potential of potassium.The comparison suggests that carbon/CO evolution leading to the formation of methane is the more preferential product in the presence of LiOH as also observed in the previous study 45 .Therefore, the KOH-NaOH (50:50 mol%) electrolyte use was not preferred for hydrocarbon production.Consequently, the fuel production (H2, CH4) was achieved in all cases of molten hydroxide electrolytes with different product composition and concentration (vol%) as can be seen from Table 2 but the best results were provided by the LiOH-NaOH (27: 73 mol%) molten salt with GFIE mode than the other cases.

Molten carbonate electrolyte
The third kind of electrolyte used for hydrocarbon production is a ternary molten carbonate mixture (Li2CO3-Na2CO3-K2CO3; 43.5: 31.5:25.0 mol%) that is used in this research due to its relatively low melting point of 394 o C. The formation of hydrocarbons can occur directly or indirectly in a molten carbonate through the reaction of C with H2 or CO with H2 respectively which are produced primarily from the independent reductions of CO2 and H2O 28,52 .
Subsequently, experiments conducted on this salt at a range of 400-450 o C, can be perfect conditions for efficient hydrocarbon formation.In the case of electrolysis applied at conditions of 1.5 V cell voltage and 425 o C, the maximum CH4 production rate was achieved.It can be seen from Fig. 9 that a significant amount of CH4 (1.10 µmol/h cm 2 ), H2 (4.40 µmol/h cm 2 ) and CO (11.70 µmol/h cm 2 ) were obtained at the lower current density range of 4-6 mA/cm 2 .
The relevant faraday efficiency obtained was 56.20% for the production of CH4, CO and H2, which were confirmed through GC analysis using FID and TCD detectors (see Fig. 10) with production concentration values mention in Table 3.These production results are in agreement with previous studies 24,25 .Wu et al. 10 provided support to the conversion of CO2 and H2O to methane in case of molten carbonate electrolysis.It is worth mentioning that H2 and CO were the predominant gases during the experiment.Moreover, the existence of CO as clearly noted from Fig. 9 and confirmed through GC analysis with TCD detector (see Fig. 10 (b)) in a relatively significant amount (in comparison to CH4), can be imputed to the individual reduction of CO2 to CO.Previous studies stated that CO itself cannot be expected in molten carbonates at temperatures below 775 o C in cases where H2O is absent 41 .
However, some other authors have claimed that the formation of CO molecules can occur on the cathode by CO2 reduction at low temperatures (≤ 650 °C) 53 .If the reduction of CO2 to CO is preferred, then H2 gas will also be formed according to the water gas shift reaction (WGSR) which occurs due to higher temperature (< 600 o C).In contrast, CO can be generated by the reverse water gas shift reaction (RWGS) 50 .However, WGSR is more feasible at temperatures below 817 o C particularly in the event of high partial pressures of H2O (up to 16.1 mmHg) which is not the condition of present study case, so CO formation is preferred case than the H2 production leading to the hydrocarbon production.The only GFIE mode is presented here due to the same results obtained in both cases (GFOE and GFIE mode) because of the excessive CO 3 2− ions already present in Li2CO3-Na2CO3-K2CO3 (43.5: 31.5:25.0 mol%).
The existence of CO2 gas in the cathodic gas products in all the molten electrolyte cases can be due to the reasons as (1) some of the absorbed CO2 from the molten carbonates can come out with the cathodic product gas (2) CO2 can be produced accompanying the various hydrocarbon species (3) The difference between the inlet and outlet amounts of CO2 cannot be ultimately accounted as the transferred CO2 to CO and hydrocarbon products.Some other amounts of CO2 can be absorbed chemically in the molten salts (4) The 100 % CO2 gas conversion cannot be done.However, in large scale applications, the cathodic product gas with accompanied amounts of CO2 can be recycled repeatedly with feed gas to increase the ultimate CO2 conversion rate.Ji et al. 36 was able to convert CO2 and H2O into CO, H2 and CH4 products at 600 o C with the current efficiency of 51% in Li-Na-KCO3-0.3LiOHelectrolyte.

Effect of temperature and voltage
The optimum temperature used for the selected molten hydroxides was chosen on the basis of the maximum CH4 production obtained as can be seen from electrolytes respectively.The yields of hydrocarbon products (vol%) increased with the rise in temperature up to an optimum temperature value while after that further rise in temperature showed inverse effects in case of molten hydroxide and chloride salts.This was because the CO2 could not be transferred significantly to CO or hydrocarbon species because of the prospects chemisorption of CO2 in molten electrolytes at higher temperature.Ji et al. 36 provided the support to the obtained results by reporting that the reduction of co-electrolysis of CO2 and H2O decreases by increasing the temperature.
While in the case of Li2CO3-Na2CO3-K2CO3, the highest CH4 production increased up to 425 o C while after this temperature CH4 production starts decreasing which might be due to the increase in production values of other longer chain hydrocarbons (C2-C4) rather than CH4 only.
This can be due to the increase in CO2 gas solubility inside molten chloride at high temperature (475 o C) 54 .The cell voltage is a key variable that can affect energy consumption or current efficiency but it can also improve the product properties at the same time 17,55 .In case of molten chlorides and hydroxides, the average current density increased drastically (20 to 70 mA/cm 2 ) and (70 to 120 mA/cm 2 ) by increasing cell voltage from 2V to 3V as shown in Fig. 12(a) and (b).Likewise, CH4 concentration (vol%) increased but with slower production rates.
However, the alkali metal electrodeposition starts occurring at a high cell voltage, consequently affecting the current efficiencies of the products.So, at higher voltage, there is more waste of energy due to the solid metal accumulation than the desired products 56 .Therefore, the optimum voltage selected for molten chlorides and hydroxides was 2V rather than 3V.To show the effect of increasing cell voltage in molten carbonates, Fig. 12(c) illustrates the high difference between the average current (4 to 25 mA/cm 2 ) resulting from electrolysis applied at 1.5 and 2 V. The hydrocarbon formation was confirmed only at 1.5 V while carbon deposition occurred due to the rise of voltage up to 2 V as also confirmed by previous studies 54,57 .Performing both runs at 425 o C, hydrocarbon formation at 2 V was rare and not noticeable.Consequently, the optimum voltage selected for molten chloride and molten hydroxide was 2 V while 1.5 V for molten carbonates.

Formation of higher hydrocarbons
The GC analysis performed using FID detector (Fig. 10(a)) showed that along with methane production, various higher hydrocarbons were also detected in the case of molten carbonate electrolyte.Which is further confirmed by GC-MS analysis (Fig. 14).The formation of methane gas product can be justified due to the reaction of carbon or CO with H2 as follows: The Gibbs Energy values were determined at 425 o C (HSC Chemistry software, version 6.12; Outokumpu Research) as this was the temperature of the experiment.It can be seen from the first mechanism that the production of general hydrocarbons occurs basically from reaction in Eq. ( 16) with the fresh deposit of carbon and adsorbed atomic hydrogen (H), produced in turn from the individual reduction of CO2 and H2O respectively.On the other hand, the C2, C3 and C4 hydrocarbons, detected by GC analysis (with FID detector) are shown in Fig. 13 along with their production rate values (0.80, 0.50, 0.50 µmol/h.cm 2 ) and faradays efficiency (total = 55.20%).It is important to note that the accumulative faraday efficiency for all products (C1, C2, C3, C4, CO and H2) obtained in case of molten carbonates electrolysis reached to the 95% (Table 3).
The dominant peaks were of alkene products rather than alkanes in the GC analysis when detected with FID detector, such as for ethene, propene, butene and pentene at 2.73, 3.06, 7.71 and 18.11 of retention times respectively.However, GC-MS analysis are also showing the detection for some alkane products.The formation of alkene or alkanes can be justified due to the (1) reaction of C or CO with hydrogen or (2) partial oxidation of methane in molten carbonate.Furthermore, in the first mechanism the CO produced in excess can react with H2 gas to produce higher hydrocarbons (C2, C3, and C4) through two different routes.The first set of reactions (Eqs.(19-20)) results in H2O generation 58,59 whereas the second set (Eqs.(21-

22
)) produces CO2 instead 60 .The CO2 by-product method is more feasible than the method with H2O formation as shown in Table 4.
Alkane and alkene products in general are generated primarily through the CO2 route particularly in media where CO2 is highly absorbed (molten carbonates).The absorption of some amounts of generated CO2 can be sustained in the molten salt, driving the reactions (Eqs. (21-22)) to the right side and increasing hydrocarbon formation.Moreover, due to the primary production of higher CO rates and in contrast lower H2 rates, alkene hydrocarbons were found in a higher proportion than the corresponding alkanes in the final cathodic product.The ∆G data values (Table 4) confirm that the formation of higher hydrocarbon molecules (C2-C4) was possible through the production of CO2 for alkanes rather than alkenes by the process of Fischer Tropsch reaction.
Therefore, as far as adequate amounts of CO and H2 gases are produced from electrolysis, there is sufficient availability for combining on the cathode surface producing alkanes.While the justification for the formation of alkenes such as C2H4, C3H6, C4H8, rather than alkanes can be provided by the partial oxidation of CH4 gas.These conditions hold true particularly at a lower CO2 absorption level due to the feasible partial oxidation of CH4 to C2H4 rather than C2H6.
The oxidation of CH4 can be performed in two ways.Firstly, CH4 gas can react directly with O2 formed at the anode during the co-electrolysis of CO2 and H2O (Eqs.(23-24)) or also can react with O2 absorbed inside the molten salt for a short time prior to passing through the anode ceramic tube or being eluted with the cathodic gas product by the draft of feed gas.Secondly, the absorbed O2 can be transferred to a more reactive oxide anion like peroxide (diatomic O2 2- or monoatomic O -), playing a significant role in the methane oxidation mechanism particularly in the case of low CO2 concentration levels.It can also be seen from Table 4 that the formation of higher molecular weight hydrocarbons (>C2) will be more feasible (resulting in a more negative ∆G) by this mechanism with the priority on alkenes rather than alkanes.
The formation of C2H6, C3H8 and C4H10 was relatively small compared with the corresponding alkenes as also seen by GC-MS analysis (Fig. 14) as the peaks 57, 43 and 29 stands for the mass of fragments lost from C4H10 (CH3CH2CH2CH3), C3H8 (CH3CH2CH3) and C2H6 (CH3CH3) respectively.The last peaks (43 and 29) are produced from the further fragmentation of C3H8 and C4H10.Peaks 55 and 41 stands for the mass of fragments lost from 1-C4H8 (for instance) and C3H6 respectively.Peak 15 is showing the mass fragment (methyl) lost from C4H10, C3H8 and C2H6.Branco et al. 61,62 also stated the higher hydrocarbon production (C2-C4) through partial oxidation of methane in molten salt electrolytes.

Energy consumption and heating values
The energy required for the conversion of CO2 to carbon/hydrocarbons will be that needed to carry out the electrolysis and heating up of the molten salt 32 .If the heating values or energy supplied from the produced fuels are able to compensate some or all the energy consumed while performing electrolysis, the process feasibility increases 63 .This is because the yield of heat generated from the produced hydrocarbon fuel can compensate or substitute some of the normal electricity employed in large scale industrial applications.In the case of molten chloride (KCl-LiCl; 41-59 mol%) electrolyte, the heating value obtained is 162 J from the produced fuel (H2 and CH4) with the energy consumption of 278 J.While the heating values obtained are 136 and 170 J from the produced fuels (H2 and CH4) by using KOH-NaOH (50: 50 mol%) and LiOH-NaOH (27:73 mol%) respectively.And with the energy consumption of 1200 and 1000 J in KOH-NaOH (50: 50 mol%) and LiOH-NaOH (27:73 mol%) electrolysis respectively (see Table 2).
The greater the production of higher hydrocarbons (C1-C4), the greater the faraday efficiency and subsequent energy profit attained due to their ability to produce more heating energy (Table 3).It is very interesting to note that the energy obtained from the summation of heating values of cathodic products in Li2CO3-Na2CO3-K2CO3 (43.5 : 31.5 : 25 mol%) case was 94.6 J while the total consumed energy was 114.2 J with about 100% of faraday efficiency (Table 3).
The higher total efficiency results in significantly lower energy consumption of 114 J for the total fuel produced or just 0.157 kWh per mole of fuel.This value is apparently less than the energy consumed for an optimum deposit carbon operation of 0.456 kWh per mole of carbon 64 .As in all the cases, the produced hydrocarbon fuels are able to provide sufficient heating values so the CO2-H2O co-electrolysis processes are considered successful.Tang et al. 54 has optimized energy consumption for producing 1 kg of carbon from CO2 as low as 35.59 kW h with a current efficiency of 87.86% under a constant cell voltage of 3.5 V in molten carbonates.

Conclusions
This study presents a new method of CO2-H2O conversion into hydrocarbon fuel via molten salts electrolysis at relatively low temperature that is a dire need of hydrocarbon production.
While in molten hydroxide (LiOH-NaOH; 27: 73 mol %), the H2 was the predominant gas due to H2O electrolysis which contributed majorly to the production of CH4 by reacting with CO2.
The hydrocarbon production rate increased (CH4: 1.02 to 6.12 µmol h/cm 2 ) by changing the feed gas insertion mode from GFOE to GFIE by using a ceramic tube.In case of molten carbonate, the production rate of CO (11.70 µmol/h.cm 2 ) was significantly higher than H2 (4.40 µmol/h.cm 2 ) in cathodic gas product.Along with H2 and CO, other hydrocarbon species such as CH4 and olefins were also produced in molten carbonate case with 99 % of faraday efficiency while other being 59.30% and 87.70% in molten chloride and molten hydroxides respectively.
Moreover, the suitable conditions at which the fuel production was achievable are 375 o C, 275 o C and 475 o C for molten chlorides, molten hydroxides and molten carbonates under the cell voltage of 2V, 2V and 1.5 V respectively.The proposed technique holds promise as a method for converting electrical energy produced from renewable power sources into conventional fuel, this should be used in future with increased production concentrations.

TOC Graphical Abstract
The co-electrolysis of CO2 and H2O in molten chloride, molten hydroxides and molten carbonates was performed at moderate temperatures for sustainable hydrocarbons formation.eee-e- them in liquid state to enable electrolysis and promote hydrocarbon formation.The composition selection of binary mixtures (LiCl-KCl, LiOH-NaOH and KOH-NaOH) was done on the basis of thermodynamic phase diagrams as illustrated from the Fig. S1.The ternary phase diagram (Fig. S2) is clearly indicating the composition selection of ternary molten salt mixture (Li2CO3-Na2CO3-K2CO3).Therefore, the salts selected for this study along with their compositions are: LiCl-KCl (58.5: 41.5 mol%), LiOH-NaOH (27: 73 mol%), KOH-NaOH (50: 50 mol%) and Li2CO3-Na2CO3-K2CO3 (43.5: 31.5: 25 mol%), having eutectic melting temperatures of 361, 218, 170 and 397 o C respectively.The electrolyte salts were dried in an oven at 200 o C for 4 h at atmospheric pressure before their mixing to remove any sort of water impurity.

( 15 )
) to CH4 formation.The hydrocarbon production in molten hydroxide (LiOH-NaOH: 27: 73 mol%) performed in two modes: GFOE and GFIE at the conditions of 2 V applied voltage and 275 o C, is shown in Fig. 6 (a) and (b).The results

Fig. 2 .Fig. 3 .
Fig. 2. The gas chromatography analysis of calibration gas standards as reference for the comparison with other electrolysis gaseous products by (a) FID detector (b) TCD detector.

Fig. 4 .
Fig. 4. The faraday efficiency and production rates of gaseous products at 2 V and 375 o C under different current density in case of molten chloride electrolysis during GFIE mode after (a) 30 min (b) 60 min.

Fig. 5 .Fig. 6 .
Fig. 5.The gas chromatography analysis of gaseous products in case of molten chloride electrolysis at 375 o C and 2 V by (a) FID detector (b) TCD detector.

Fig. 9 .
Fig. 9.The faraday efficiency and production rate of gaseous products at 1.5 V and 425 o C under different current density in molten carbonate electrolyte.

Fig. 10 .
Fig. 10.The gas chromatography analysis of gaseous products in case of molten carbonate electrolysis by (a) FID detector (b) TCD detector.

Fig. 11 .Fig. 12 .
Fig. 11.The selection of optimum temperatures for all electrolytes on the basis of CH4 production.

Fig. 13 .Fig. 14 .
Fig. 13.The faraday efficiency and production rates of higher hydrocarbons at 1.5 V and 425 o C in case of molten carbonate electrolysis.

Table 2 .
Specification of cathodic gas products during electrolysis in molten hydroxide (LiOH-NaOH) at 275 o C and (KOH-NaOH) at 225 o C under 2V applied voltage using GC analysis.* After electrolysis with GFIE mode * After electrolysis with GFOE mode *

Table 3 .
Specification of cathodic gas products after electrolysis in molten carbonate at 1.5 V and 425 o C by using GC and mass spectrometric analysis.

Table 4 .
List of ∆G and ∆H for the generation of hydrocarbon products from the Fischer-Tropsch reaction (through CO2 and water formation) and partial oxidation of methane at 425 o C.