Rehydrogenation of Sodium Borates to Close the NaBH 4 ‐ H 2 Cycle: A Review

: In 2007, the US Department of Energy recommended a no ‐ go on NaBH 4 hydrolysis for onboard applications; however, the concept of a NaBH 4 ‐ H 2 ‐ PEMFC system has the potential to become a primary source for on ‐ demand power supply. Despite the many efforts to study this technology, most of the published papers focus on catalytic performance. Nevertheless, the development of a practical reaction system to close the NaBH 4 ‐ H 2 cycle remains a critical issue. Therefore, this work provides an overview of the research progress on the solutions for the by ‐ product rehydrogenation leading to the regeneration of NaBH 4 with economic potential. It is the first to compare and analyze the main types of processes to regenerate NaBH 4 : thermo ‐ , mechano ‐ , and electrochemical. Moreover, it considers the report by Demirci et al. on the main by ‐ product of sodium borohydride hydrolysis. The published literature already reported efficient NaBH 4 regeneration; however, the processes still need more improvements. Moreover, it is noteworthy that a transition to clean methods, through the years, was observed.


Introduction
The negative impact of the excessive use of fossil fuels has been a major concern since the beginning of the century, mainly in the current decade [1]. Thus, the need to reduce greenhouse gas concentrations to sustainable levels is mandatory to allow the future generations to live on an Earth at least as habitable as today, which has resulted in a demand for clean and abundant energy technologies [2]. In this context, hydrogen (H2) arises as one of the most promising fuels of the future. Hydrogen can be efficiently oxidized in Polymer Electrolyte Membrane Fuel Cells (PEMFC) to provide electricity to a wide range of applications (e.g., portable and maritime applications) [3,4], involving an electrode (anode and cathode)/membrane electrolyte system. The energy stored in the H2 fuel is converted into electricity and heat, producing only water as a by-product [5,6]. Thus, hydrogen is supplied in the anode and splits into electrons and protons, while oxygen is reduced at the cathode. The protons are carried through a polymer electrolyte membrane (PEM) to recombine with oxygen, whereas the electrons proceed to the cathode through an external circuit, providing the electrical output [6].
Hydrogen is an energy vector with a strong potential for energy storage due to its high gravimetric density (142 MJ.kg −1 ) and lightness. Nevertheless, it is lighter than air, flammable, and is usually combined with other elements [3,7]. Due to its abundance, hydrogen can be produced by several methods using a wide range of sources, through fossil fuels to renewable energy resources. Currently, 96% of worldwide hydrogen is generated from fossil fuels; nevertheless, to be a low-carbon energy carrier, new generation methods must be developed [2,[7][8][9].
Chemical hydrides, such as metal-boron hydrides and ammonia borane, have received strong attention, since they exhibit an impressive volumetric hydrogen storage density on a material-only basis and simultaneously allow the generation and storage of hydrogen [10,11]. Among chemical hydrides, sodium borohydride (NaBH4) is a promising hydrogen carrier due to its high hydrogen storage capacity (10.8 wt %), safe handling, and fast kinetics of hydrogen release [12][13][14]. NaBH4 releases molecular hydrogen via hydrolysis, at room pressure and temperature, as described by the chemical reaction presented in Equation (1) [3,15]. NaBH 4 2 H 2 O → NaBO 2 4 H 2 (1) The reaction is spontaneous and exothermic (ΔH = −217 kJ.mol −1 ), and the release of hydrogen can be accelerated by a catalyst [16,17]. The only by-product that forms upon hydrolysis usually retains two or four molecules of water, which decreases the hydrogen yield and generation rate. Therefore, an excess of water is needed, and the effective hydrolysis reaction of NaBH4 is observed according to Equation (2), where × represents the hydration factor [18][19][20].
Although the hydrolysis of NaBH4 combines the best properties for hydrogen generation and storage, including solubility in water, rapid controllable hydrolysis, overall stability, and moderate exothermicity, in 2007, the US Department of Energy (DOE) recommended a no-go on NaBH4 hydrolysis for onboard applications [34]. This decision was based on three key issues: storage capacity targets, regeneration of NaBH4, and costs [35]. In order to address these issues, the research dedicated to the synthesis and regeneration of NaBH4 improved [36].
Bearing in mind the research articles concerning the NaBH4 hydrolysis by-product recycling that have been published in peer-reviewed international journals since the early 2000s, only three studies were dedicated to this topic [37][38][39] prior to the US DOE report in 2007. About 96% of the papers on this topic [20, were published after the DOE no-go recommendation ( Figure 1). Despite the efforts, since 2009, to develop an efficient and reliable method to regenerate NaBH4, the reported research is low when compared with the total number of published studies on NaBH4 hydrolysis. In the past 20 years, 12 review-type articles were published regarding the hydrogen generation through boron-based hydrides [21,22,35,36,[69][70][71][72][73][74][75][76]. Although 10 of these articles focused on NaBH4 hydrolysis, only five overviews highlight the importance to recycling the by-product in order to close the NaBH4-H2 cycle [21,22,35,36,71].
In 2015, Demirci et al. [21] analyzed 260 research articles published between 2000 and 2013 that were dedicated to "the hydrogen cycle with hydrolysis of sodium borohydride" and reported that only 8.9% of the papers were related with NaBH4 regeneration. According to the authors, more than 60% of the research articles focused on catalysis. There is a noteworthy lack of research in recycling, which is a key issue to close the NaBH4-H2 cycle. Therefore, 217 research papers [11,17,20,23,24,[28][29][30][31]50,[52][53][54][55][56][57][58][59][60][61], published from 2014 up to the publication of this paper, were selected to understand the evolution of the research on "hydrogen from NaBH4 hydrolysis" after the report by Demirci et al. [21]. The selected articles were published in peer-reviewed international journals, and all the review-type articles and conference proceedings were discarded. According to their main focus, the research articles were divided in four topics: "catalysis", "fundamentals", "others", and "regeneration" (Figure 2). The "catalysis" topic highlights the performance and characterization of catalysts; "fundamentals" focuses on reaction mechanisms, kinetics, and thermodynamics; the topic "other" emphasizes the experimental and computational study of reactors, prototypes, and developed systems; and the "regeneration" topic is dedicated to the synthesis of NaBH4 from the reaction by-products. It is noteworthy that the high majority of the studies, about 80%, continue to be dedicated to the topic "catalysis". Despite the need to close the hydrogen cycle from NaBH4 hydrolysis, which is strongly dependent on the development and implementation of an efficient, low-cost, and reliable recycling method [21,36,61], only 5.5% of the published articles highlight the topic "regeneration". Although important advances have been made through the years, there is still a low impact of this topic. The usual path to synthesize NaBH4 is the Brown-Schlesinger process, which was developed in the early 1950s [15,22]. This process is expensive, not clean, and inefficient. It consists of seven steps, which require high pressures and temperatures and produce various by-products [36,58]. The main reaction is presented in Equation (4), where sodium hydride (NaH) and trimethylborate (B(OCH3)3) react at 225 to 275 °C [15,42]. 3 NaH → NaBH 4 NaOCH 3 (4) An alternative method to produce NaBH4 at industrial scale is the Bayer process, which is represented in Equation (5) [33,37]. Na2B4O7•7SiO2 reacts with Na at 400-500 °C under a hydrogen atmosphere. It is noteworthy that this process exhibits high potential risks due to the high temperature and hydrogen pressure operations. Moreover, it produces Na2SiO3 as a by-product, which is a residue with low commercial value that is difficult to discard [33,58].
Although the synthesis of NaBH4 from hydrolysis by-products remains a great challenge, the studies published to date present promising advances [25,33]. Nevertheless, a low-cost fully reliable alternative method to synthesize NaBH4 has not yet been developed. Closing the NaBH4-H2 cycle is essential to implement a NaBH4-H2-PEMFC system as a primary energy source to provide on-demand power for portable and maritime applications [4,11]. Moreover, NaBH4-H2-PEMFC systems emerge as a promising off-grid technology capable of providing a fully reliable electricity supply in emerging countries and locations without a reliable grid.
Among the many possible hydrogen carriers, liquid ammonia and LOHC (liquid organic hydrogen carriers) have been gaining increased importance [22,70,274]. Ammonia has a very high hydrogen content (17.8% in weight) and simultaneously an impressive volumetric hydrogen density (108 kg H2/m 3 in liquid ammonia at 20 °C and 8.6 bar). However, ultra-pure hydrogen recovery from the ammonia cracker in the reconversion to hydrogen process is still a great challenge [28,275]. LOHC exhibit moderate H2 content, reversibility, moderate dehydrogenation temperature, commercial availability, and compatibility with the existing gasoline infrastructure, but the reactor systems, process heat integration, and concerns such as catalyst recovery and regeneration, separation, and capture of CO2 from the gaseous products formed need further developments [274]. In comparison, chemical hydrides present, as referred above, a high hydrogen content and low to moderate dehydrogenation temperatures (the hydrolysis reaction is an easy and controllable process and allows simultaneously hydrogen storage), but they suffer from irreversibility and energy-consuming regeneration. Therefore, a major challenge for the effective use of these hydrogen carriers is the development of alternative or improved routes to rehydrogenate the by-products formed. The specific advantages of the sodium borohydride carrier were already enlightened. Thus, this overview focuses on the published alternative methods to rehydrogenating the borates leading to the NaBH4 regeneration, since this is a critical issue that needs to be overcome in order to close the NaBH4-H2 cycle and allow the implementation of this technology [36,59,61].

Regeneration of NaBH4
The regeneration of sodium borohydride can be organized into three major groups of processes based on the energy source: thermochemical, mechano-chemical, and electrochemical. Thermochemical processes are based on reactions that involve high pressure and/or temperature, and they comprise most of the reactions to effectively regenerate NaBH4. The mechano-chemical processes are similar to the thermochemical ones, but the source of energy used in this type of process relies on mechanical forces. Lastly, the electrochemical processes use electric energy to produce sodium borohydride by reducing or oxidizing other borates. Until 2015, the regeneration of sodium borohydride was mainly achieved through thermochemical processes [41,42] ( Figure 3). However, since 2016, there has been a transition to cleaner processes to regenerate this compound, with mechano-chemical processes being the most used ones [20,43,45,50,[52][53][54][55][56][57][58][59][60][61]68]. The lack of efficient NaBH4 electrosynthesis also contributed to this "mechanochemical processes boom" to regenerate NaBH4. An overview of the NaBH4 regeneration yields through thermochemical and mechano-chemical processes is presented in Figures 4 and 5, respectively. Regeneration yields based on thermochemical processes are presented according to the temperature (closed circles) and pressure (open circles) of the reaction ( Figure 4). Almost all experiments were performed using pressure values between 2 and 4 MPa and temperature values between 800 and 900 K. For mechano-chemical processes, the regeneration yields are presented along with the time of milling (closed circles) and the year of publication of the articles (open circles) ( Figure 5). As observed, higher yields are achieved when using higher milling times; nevertheless, regenerating NaBH4 for more than 35 h (2000 min) is a considerable energy-consuming process. These higher yield values have been obtained since 2017, which may reflect a gradual improvement of these processes. It is noteworthy that the same color points in both figures correspond to the same first author, indicating that a wider variety of groups has progressively been focusing on mechano-chemical sodium borohydride regeneration.  Regarding the electrochemical synthesis of NaBH4, studies conducted at Penn State University (PSU), in combination with DOW Jones Industrial Average and Los Alamos National Lab (LANL), indicated that no NaBH4 was electrochemically obtained, and it would not be possible to effectively regenerate it from NaBO2. This statement contributed to a lack of papers on the subsequent years using electrochemical methods to regenerate NaBH4. Moreover, although several patents have claimed that the electrochemical regeneration was achieved, there are no published studies reproducing the results. It is noteworthy to highlight some papers that synthesize NaBH4 through electrochemical processes as an intermediate step for desulfurization experiments [276][277][278][279].
The following sub-sections describe the results reported so far regarding the regeneration of NaBH4. It is noteworthy that most of these results were obtained using commercially available boron compounds. Thus, there is an emergent need for developing synthesized boron compounds to regenerate NaBH4, as the ultimate goal is to use the byproducts of its hydrolysis for the regeneration process, which can react differently compared to commercial boron compounds.

Thermochemical Processes
Thermochemical processes are chemical reactions that occur at high pressure and/or temperature. As previously referred, NaBH4 is usually synthesized through two thermochemical processes: (i) the Brown and Schlesinger process and (ii) the Bayer process.
The main disadvantage of the Brown and Schlesinger process is the formation of several by-products. To overcome this issue, Liu et al. [280] developed a modified closed system where NaBO2 and ammonia borate ((NH4)3(BO3) are simultaneously produced and recycled from NaBH4. According to the authors, the metallic Na is obtained by the electrolysis of sodium chloride (NaCl) in a seawater treatment. Although this study presents a major improvement to the Brown and Schlesinger process, it continues to be a multi-step process that needs to use metallic Na. In contrast, the Bayer process allows a direct production of NaBH4. Nevertheless, it has some associated issues: (a) the need to use metallic Na, (b) the disposal of the by-product (Na2SiO3), and (c) the risk of explosion due to high temperature (873 K) and hydrogen pressure [33,58].
Therefore, several authors attempted to synthetize NaBH4 through different methods. Since the early 2000s, only 12 papers were published reporting the NaBH4 hydrolysis by-product recycling through thermochemical processes (Table 1) [38-40,44,46-49,62,64 -66]. All reported studies were carried out using either magnesium hydride (MgH2) or magnesium under a hydrogen atmosphere. The thermochemical studies using MgH2 reported higher conversion yields; however, this method also presents high costs due to the commercial value of MgH2 [47]. On the other hand, although the NaBH4 regeneration using Mg under a hydrogen atmosphere presented lower costs, this method is not suitable due to the need to use hydrogen as a reactant. Since the basis of these studies is to generate hydrogen as an energy carrier, this method does not seem to be profitable to close the NaBH4-H2 cycle. The thermochemical reactions are highly affected by the parameters considered, as observed in Figure 4. The published articles analyzed in this overview focus on two main parameters: the reaction temperature and pressure. Moreover, some articles also studied the influence of time and the benefit of specific additives.
[38] studied the mixture of NaBO2 with MgH2 (Equation (6)), at 7 MPa H2 and 823 K, after 2 h of reaction, and they obtained a conversion yield of 97.0%. This is the higher yield reported for the NaBH4 regeneration using a thermochemical process without additives.
Ten years later, Ou et al.
[48] studied the influence of temperature variation in Equation (6) at 3.1 MPa H2. The authors increased the reaction temperature from 293 to 884 K followed by a consequent decrease to 293 K and reported a NaBH4 regeneration of 90.4% yield. The calculus of the conversion yield was based on H2 pressure variation, and 85.2% of the obtained value had been associated to the formation of NaBH4 during the isothermal period between T = 873 K and T = 884 K. Therefore, Ou et al. [49] added two more parameters to the study: pressure and time of reaction (tr). In order to promote the hydrogen adsorption for NaBH4 formation, the authors set the H2 pressure between 2.1 and 3.1 MPa and observed that after 1.5 h, at 907 K, they achieved the maximum borate conversion. Nevertheless, at 857 K, the authors observed that it took 8 h to obtain the same NaBH4 conversion.
Eom et al.
[46] also studied Equation (6), in turn, at 6 MPa. After 1 h of reaction, the authors observed that the conversion yield increased with temperature ( Figure 6a). The results obtained stand out, since the yield is considerably lower than the ones previously reported [38,48,49]. However, the authors justified the results with the tough in MgH2 decomposition promoted by the high H2 pressure.
According to these results, it is possible to conclude that the temperature presents a main role in thermochemical processes.
Eom et al. [46] also studied, at 873 K, the influence of H2 pressure (Figure 6b) as well as other reducing agents. Thermodynamically, Ca is more suitable than Mg as a reducing agent. Nevertheless, the authors concluded that Ca does not act as an effective catalyst in the H2-Hconversion, since only a small amount of NaBH4 was formed.

Effect of Additives and the Hydration of the Boron Compound
In addition to the effect of temperature, Kojima et al.
[38] also studied the influence of additives in the reaction. The authors performed experiments with silicon (Equations (7) and (8)) at 7 MPa H2 and 823 K, and they obtained NaBH4 with a 98% yield after 2 h of reaction.
[39] studied the influence of Al as a reducing agent and observed that the formation of NaBH4 did not occur. However, the authors reported a maximum NaBH4 conversion of 65.8% when the compounds tetrasodium diborate (Na4B2O5) and sodium oxide (Na2O) were combined with Al (Equation (9)).
The studies were carried out at 873 K with a NaBO2/Na2O molar ratio between 1.5 and 4 ( Figure 7). The authors concluded that although the presence of Na2O has a positive impact on the regeneration of NaBH4, the NaBO2/Na2O molar ratio must be optimized [64]. Furthermore, Liu et al. [40] studied the influence of several additives in the reaction of NaBO2 with Mg and H2, such as Fe, Co, Ni, and Cu. The authors reported that Fe and Ni affect the reaction similarly. The addition of 5 wt % of these metals not only increased the NaBH4 yield (Fe: 88%; Ni: 82%) but also decreased the reaction temperature (Fe: ≈283 K; Ni: ≈313 K). Moreover, increasing the Fe concentration up to 20 wt % promotes the NaBH4 generation rate with similar conversion yields. The authors also observed that the addition of 5 wt % of Co converted NaBH4 with 85% yield, and the addition of the same amount of Cu did not show any significant effect in the reaction. Additionally, the authors concluded that none of the additives decreased the activation energy of the reaction (156.3 kJ.mol−1) A year later, Liu et al. [65] reported a eutectic alloy of Mg with 23.5% Ni in the recycling of NaBO2. The studies were performed at 673 K and 3.0 MPa H2, and a slight increase of 3 wt % H2 absorbed on the Mg surface was observed, which promoted an increase of 8% in NaBH4 conversion yield. The authors concluded that the Mg alloy induced the hydrogen dissociation and diffusion during the reaction.
Furthermore, Liu et al. [66] also studied the NaBH4 by-product recycling without performing its initial dehydration. Bearing in mind Equation (2), the authors studied the influence of the by-product hydration between x = 0 and x = 2 and observed that while the hydration factor increases, the conversion yield considerably decreases. Moreover, the authors reported that for x = 4, no NaBH4 was formed. Therefore, the authors realized that although it is a high-energy consumption step, the dehydration of the by-product is necessary.

Mechano-Chemical Processes
The mechano-chemical processes use mechanical movement instead of heat to generate the required energy to form the products of the reaction. These reactions are typically performed in a high-energy ball mill, at room temperature, with the reactants in the solid state (powder). The impact of the milling balls on the reactants reduces the size of the particles, which increases the contact between them and, consequently, promotes the reaction. It is noteworthy that although these reactions occur at room temperature, there is an increase of temperature inside the mill during the reaction that promotes the formation of NaBH4. Similar to thermochemical processes, the mechano-chemical processes are carried out using MgH2 or Mg under hydrogen atmosphere as reactants. Although mechano-chemical processes are more environmentally friendly, since they are low-energy consumption processes, the high times of milling are a considerable limitation. Moreover, the implementation of the ball milling process to synthesize NaBH4 in a larger scale is not feasible mainly due to cost issues [58].
Since 2003, several authors reported the regeneration of NaBH4 through ball milling ( Table 2). In addition to MgH2, some authors studied the reaction with Si compounds in order to find a low-cost alternative. Moreover, the influence of several additives was studied, such as sodium peroxide (Na2O2), sodium hydroxide (NaOH), sodium carbonate (Na2CO3), sodium hydride (NaH), and methanol (CH3OH). The influence of the hydration of sodium borates was also studied, and it is noteworthy that NaBH4 was regenerated with high yield without an initial dehydration step.
Nevertheless, the NaBH4 regeneration through ball milling is a multi-step process, since after the ball mill reaction, it is necessary to proceed to NaBH4 extraction. In this additional step, it is necessary to use a solvent (e.g., isopropyl amine) to dissolve NaBH4 from the final reaction mixture and then filter and dry to obtain the final NaBH4. The information regarding the conditions inside the mill, mainly the inside temperature achieved and the inside pressure atmosphere, are rarely provided. However, it is possible to conclude from the reported articles that the inside temperature is strongly influenced by the time of milling, the speed of milling, and the ball mill material composition. Moreover, despite the poor information, researchers clearly observed a preference for H2 or Ar atmospheres inside the mill.

Effect of Time, Ball to Powder Ratio, and Excess of Reactant
Kong et al. [63] were one of the first authors to report the ball mill reaction of NaBO2 with MgH2 (Equation (10)). The influence of the following parameters was studied and optimized: MgH2/NaBO2 molar ratio, ball to powder ratio (BPR), and time of milling (t).
According to the authors, it took 1 h of ball milling to start to synthesize NaBH4 and an extra hour for the conversion to be complete, with a yield of 71%. The authors also concluded that BPR considerably influences the conversion yield, which is higher with the increase of BPR [63].
In 2010, Varin et al. [281] studied the milling of MgH2 up to 100 h in H2 and Ar atmospheres. Although it was concluded that both atmospheres did not influence MgH2 molecules, the formation of Mg(OH)2 was observed. However, in the H2 atmosphere studies, the formed Mg(OH)2 was naturally reduced to MgH2 and water during the reaction, which allowed the authors to conclude that a H2 atmosphere is more suitable to perform this process.
Çakanyldirim et al. [68] analyzed the effect of the milling time and the excess of MgH2 in the synthesis of NaBH4 through ball milling (Figure 8)  In 2017, Ouyang et al.
[54] also studied the influence of the same parameters (milling time, excess of Mg and BPR) in the ball milling reaction of NaBO2 with Mg, expecting to achieve higher yields by increasing those parameters. Nevertheless, the obtained conversion was below the expectations, and a maximum yield of 67.7% was reported after 15 h of milling, with 37.5% excess of Mg and a BPR of 30:1.
Furthermore, bearing in mind the borates structure presented in the NaBH4 regeneration process, Ar et al. [56] reported ball mill studies where B2O3 reacts with NaNH2 and MgH2 (Equation (11) Several experiments were carried out with time of milling between 400 and 800 min. According to the authors, the optimal value for the time parameter is 500 min of milling, since above this value, researchers observed the formation of MgO instead of the desirable NaBH4. Moreover, Ar et al. reported that the excess of MgH2 is necessary since, at the end of the reaction, it was observed that a portion of this compound was fixed to the reactor walls and did not react. On the other hand, if the amount of MgH2 is too high, it may decrease the contact between the reactants and, consequently, the NaBH4 yields will be lower. Thus, the authors concluded that the optimal value for the amount of MgH2 is 35% excess, which is in agreement with the results published by Çakanyldirim et al. [68].
In 2019, Le et al.
[58] reacted NaBO2 with a Mg-Al based alloy at 7 MPa H2 with a BPR of 20:1. The reported results showed a conversion close to 100% after 36 h of milling. Moreover, the authors stated that the milling time could be reduced to 12 h without any considerable loss since, at that point, the NaBH4 conversion was about 97%.

Effect of Additives and the Hydration of the Boron Compound
In addition to the influence of the parameters mentioned before, Çakanyldirim et al. [68] also studied the influence of several additives (Al, Na, and Na2CO3) and observed that the all the studied additives promoted a negative effect in the conversion yield ( Figure 9).
Lang et al.
[53] studied the influence of methanol (CH3OH) as an additive to the NaBO2-MgH2 mixture and reported a slight increase of NaBH4 conversion yield when the time of milling was superior to 10 h ( Figure 10). Furthermore, several authors researched the possibility to regenerate NaBH4 through ball milling using the borates in its hydrated form. Since the by-product of NaBH4 hydrolysis is usually obtained in its hydrated form, and its dehydration is a high-energy consumption step, the possibility of ball milling the hydrated borates with high conversion yields makes this method more feasible. Thus, Huang et al.
The reaction was performed in a vibrational ball mill under Ar atmosphere, and the time of milling as well as the reactants molar ratio were analyzed. The results presented a conversion yield of 86%, after 20 h, with a Mg:Mg2Si:NaBO2 molar ratio of 2.5:1:1.
Zhong et al.
[57] also studied the recycling of the dehydrated sodium metaborate at room temperature in an Ar atmosphere; however, it was ball milled with Mg2Si (Equation (13)). The process regenerates NaBH4 with 78% yield, after 20 h of milling, with a Mg2Si/NaBO2•2H2O molar ratio of 3:1.  (14) and (15)) in a high-energy ball mill at room temperature and pressure during 15 h. Both borates were mixed in a 1:12.5 molar ratio and 68.6% and 64.1% of NaBH4 were obtained for the NaBO2•2H2O and NaBO2•4H2O, respectively.  (16) According to the authors, borax presents more similarities with NaBH4 and has a reactivity higher than NaBO2, which makes the regeneration of NaBH4 easier. A maximum yield of 78.9% was obtained after 30 h of milling and 23.75% excess of Mg.
The NaBH4 regeneration from Na2B4O7 with MgH2 was previously studied by Li et al. [37] in 2003. The experiments were carried out during 1 h, and the influence of three additives was studied: Na2O2, NaOH, and Na2CO3. The higher NaBH4 conversion (78%) was achieved with the use of Na2CO3 as an additive.

Effect of the Mill
As referred above, there are different types of ball mills, which can also influence the reaction. The work published by Hsueh et al.
[51] compares the results of NaBH4 regeneration, which are presented in Equation (10) and were performed in a planetary ball mill and a shaker mill. Initially, the reactant mixture was ground for 6 h. Then, the experiments were performed by maintaining the amount of borate constant and increasing the excess of MgH2 in both mills ( Figure 11). Hsueh concluded that the shaker mill was more suitable than the planetary mill due to the difficulty of decomposing the reaction and promoting the NaBH4 formation.

NaBH4 Formation and Energy Efficiency
To better optimize the NaBH4 synthesis, it is essential to understand NaBH4 formation.
Ouyang et al.
[54] studied the NaBO2•2H2O conversion into NaBH4 and reported the structural rearrangement of the borate to NaB(OH)4. The authors analyzed and described the ball mill reaction between NaBO2•2H2O and Mg and observed the formation of magnesium hydroxide (Mg(OH)2), after 10 min of milling, as an intermediate in the reaction. This compound was also consumed after 1 h, as shown in Equation (17). The formation of Mg(OH)2 plays an important role in this reaction, since its consumption during the process is responsible for its appearance of molecular hydrogen in the reaction medium. Those observations were also reported by Chen (18)). The authors reported that this step is important to promote the NaBH4 regeneration yield.
[50] calculated the energy efficiency of the consumption/regeneration of NaBH4 based on the recycling of NaBO2 with MgH2 in a high-speed vibrating mill. The results showed an energy efficiency of about 50%, which led to the conclusion that the system is feasible. Nevertheless, it is important to find novel techniques to improve the energy efficiency of this process, for example by reusing the heat released from the reactions.

Electrochemical Processes
As previously referred, the regeneration of NaBH4 can occur through the input of electric energy. When it is applied in the right conditions, it can reduce or oxidize compounds. This process takes place inside an electrochemical cell, which consists of an anode (where oxidation reaction occurs) and a cathode (where reduction reaction occurs). The electrolyte is essential to warrant the correct means to promote the reduction and/or oxidation of any intended compound. Moreover, the electrochemical cell can also be used to convert the chemical energy of a compound in electric energy. These processes are desired due to its low energy consumption, cleanliness, and reduced cost compared to the thermochemical processes.
The regeneration of sodium borohydride from sodium metaborate can theoretically be obtained mainly through the direct electroreduction, as shown in Equations (19) and (20). Cathode: Anode: Although there have been claims of sodium borohydride synthesis, those results have not been replicated. This fact is associated with the formation of other compounds at the potentials applied at the cathode: for example, the water decomposition into H 2 and OH . Other electrochemical processes have been referred in the literature; however, the results obtained were similar to the direct electroreduction process. From what the authors could find, the first paper reported in the literature that confirms the NaBH 4 electrosynthesis was that of Zhu et al. [280,282] The authors reported a 15.1% current efficiency for the NaBH 4 electrosynthesis at atmospheric pressure and room temperature. The electrochemical cell was composed of a copper working electrode and a lead counterelectrode. Then, 1 mol/L H 2 SO 4 solution was added to the anode and 1 mol/L NaOH and 0.2 mol/L NaBO 2 solution was added to the cathode. A cathodic peak was observed at −1.17 V, which is indicative of the BO 2 reduction. Moreover, the increase of NaOH favors the electrosynthesis; however, excessive NaOH hardens the products' adsorption and conduction. Similar observation was made for the NaBO 2 concentration. Nevertheless, the process' efficiency is too low.
Efficient electrosynthesis of NaBH 4 was achieved by Shen et al. [276]. Given the use of NaBH 4 as an effective reducing agent and its high cost, the authors considered its electrosynthesis as an intermediate step in a desulfurization experiment of a coal water slurry at ambient pressure and room temperature. The electrochemical cell consisted of cathode and anode reservoirs separated by a KCl bridge. On the cathode side, a solution containing 0.1 to 0.45 mol/L NaBO 2 and 0.025 mol/L NaOH was initially added. Then, the coal was added, and the mix was maintained under stirring. The anode side was composed of an iron sulfide (FeS 2 ) and NaOH solution. The cathode was Pb and the anode was graphite (C). After the synthesis of BH 4 in the cathode, its hydrolysis promoted the formation of hydrogen that reduced the FeS 2 to hydrogen sulfide (H 2 S). Then, in the anode, the reaction with NaOH allowed the separation and removal of the sulfur. Equations (19) and (21) The BO 2 concentration, electrolytic time, and voltage were studied between 6 and 30 g/L, 1 and 5 h, and 2.5 and 3.3 V, respectively. An optimal total sulfur removal (TSR) of 36.6% was achieved with 12 g/L, 4 h, and 3.0 V, respectively. Considering the BO 2 as an intermediate essential step of the process, these parameters also benefited the BH 4 formation.
The group continued the experiments in the years forward, and the electrochemical cell for the process was also optimized, promoting the TSR and the BH 4 formation.
Considering similar electrochemical cells-however, with a combined anode and cathode reservoir [277]-the group studied again the BO 2 concentration, electrolytic time, and voltage. However, the optimal values did not suffer a considerable change (0.2 mol/L, 4 h and 3.5 V). Moreover, the NaOH concentration and electrodes were analyzed. The NaOH concentration was studied up to 0.375 mol/L. It was observed that 0.05 mol/L was the optimal value to promote the TSR, since further increase favored the pH and reduced the hydrogen generation from the BH 4 hydrolysis. Regarding the electrodes, five anode/cathode combinations were considered: Pb/Pb, C/Pb, Ti/Pb, C/Cu, and C/Zn. The results were in favor of the last. In this way, graphite not only better promotes the TSR but also presents significant advantages to consider, such as easy to process, reduced waste and price, and good thermostability and conductivity. Following that, the group studied the desulfurization of gasoline [278]. The electrochemical cell was composed of a BDD electrode in the cathode (working electrode) and a C electrode in the anode (counter electrode) and the cathode and anode were separated by a cation exchange membrane. After BH 4 formation, a gasoline model was added to the cathode, and after the reaction, NiCl 2 •6H 2 O was added to the last. Moreover, pulse voltage was studied during the BH 4 electrosynthesis, and the researchers observed a considerable increase in the TSR (almost twice higher). Cyclic voltammetry tests were performed using 0.15 mol/L NaBO 2 , and a reduction peak was observed between −1.2 and −1.8 V, while the hydrogen and oxygen evolution reactions were initiated at voltage below −1.8 V and above 0.6 V, respectively. Forward and reverse pulse voltages were optimized considering these limits. Overall, the best TSR achieved was with a forward and reverse pulse of −1.5 V (applied after 1.5 s) and 0.3 V (applied after 0.5 s), respectively, which were explained considering that it avoided at maximum the hydrogen evolution reaction (happening simultaneously with the BH 4 electrosynthesis), and the reverse pulse voltage was enough to attract the BO 2 ions to the BDD electrode while avoiding the oxygen evolution reaction. Regarding the BH 4 electrosynthesis step, other cell parameters were also studied: NaBO 2 and NaOH concentrations, electrolytic time, and stirring rate. First, 0.15 mol/L of NaBO 2 was maintained as an optimized value, since higher concentrations did not contribute to further increasing the TSR. The optimal NaOH concentration was 0.1 mol/L, since although it is important to maintain the cell as alkaline to avoid too fast BH 4 hydrolysis, its excess reduces the oxygen evolution voltage, promoting it and making the BO 2 transfer to the BDD electrode difficult. Excess stirring rate was also observed to affect the BO 2 transfer, so 400 rpm was the best value defined. The higher efficiency of the desulfurization process was 97%.
More recently, the group studied the addition of ionic liquids (IL) to the desulfurization process of gasoline. The use of ILs as electrolyte has been increasingly reported in the literature due to their characteristics, which make them more suitable than water and other electrolytes. In the BH 4 electrosynthesis, they avoid the need to control the hydrogen and oxygen evolution reactions and present a wider electrochemical window. Moreover, they can be recovered at the end of the experiment. Similar electrochemical cells to previous study were considered by the group. In this experiment, the working electrode was a glassy carbon electrode, and the counter-electrode was a Pt wire. The IL was added to both the electrodes reservoirs in the cell, while water was only present due to the NaBO 2 solution and the need of hydrogen for the BH 4 electrosynthesis.
Cyclic voltammetry tests were performed, which indicated a reduction peak between −2.2 and −3.0 V, which was associated to the BH 4 -. Moreover, the hydrogen and oxygen evolution reactions were observed at −3.0 and 2.0 V, respectively. As expected, the electrochemical window using IL is wider than that of water. Forward and reverse pulse voltage was equally applied, and their parameters were optimized. The optimal values to promote the BH 4 electrosynthesis and the TSR (97%) were a forward pulse voltage of −2.6 V (0.6 s forward pulse time), reverse pulse voltage of 0.5 V (0.5 s reverse pulse time), and pull-off time of 0.8 s. The times that were associated to the ones referred and applied were the forward and reverse pulse voltage and turn-off. If they are too long, the reactions do not benefit; if they are too short, the reactions do not occur. Based on these results of the group and assuming that an efficient electrosynthesis of BH 4 is essential for the desulfurization of the coal water slurry, it is possible to affirm that NaBH 4 can be obtained through an electrochemical process and that it is an alternative to the thermo-and mechano-chemical processes. Nevertheless, it is essential to promote the NaBH 4 regeneration as the main aim instead of an intermediate step in another process.
In addition, as noted in the thermo and mechano-chemical regeneration processes section, the study of hydrated NaBO 2 is crucial, since it is the main by-product of the NaBH 4 hydrolysis and it is demonstrated that its recycling is not as easy. Considering the practicability and possibly low cost of these processes, the electrochemical regeneration of NaBH 4 must be considered. In that order, both the use of ILs and the pulse voltage are improvements to the process and must be studied.

Conclusions
Sodium borohydride stands as a promising hydrogen carrier for on-demand power supply. Despite the DOE no-go recommendation in 2007 for NaBH4 hydrolysis for onboard applications, over 13 years later, several findings and improvements have been made in this area. Therefore, NaBH4-H2-PEMFC systems have emerged as a feasible candidate to supply energy for portable applications, maritime applications, as well as an off-grid technology capable of providing an electricity supply in emerging countries and locations without a reliable grid. Nevertheless, closing the NaBH4-H2 cycle is strongly dependent on the regeneration of NaBH4. In this overview, a survey of alternative methods to recycle NaBH4 hydrolysis by-products was made, and three types of methods stands out: thermo-, mechano-, and electrochemical. Although the first published studies were mostly dedicated to thermochemical processes, since they present high conversion yields, over the years, they have been replaced by mechano-chemical processes. This transition shows a concern for the development of environmentally friendly methods, since mechano-chemical processes use a mechanical source of energy instead of heat and, consequently, do not require high temperatures and pressures. Nevertheless, although it is possible to obtain NaBH4 conversion yields as high as in thermochemical processes, they require high milling times to achieve those same yields. Moreover, both methods typically use expensive hydrides, such as MgH2, to promote the NaBH4 synthesis. Thus, electrochemical processes arise as a promising low-cost and clean alternative. Although the regeneration of NaBH4 from NaBO2 has been confirmed, as an intermediate step in a desulfurization experiment, the study of this process is very limited, and published articles dedicated to electrochemical methods to regenerate NaBH4 are rare. There are a few issues in this process that need to be overcome: mainly, the water decomposition and the formation of other by-products. Thus, further studies to find alternative electrochemical reactions are needed in order to take advantage of the potential of this process. Therefore, the rehydrogenation of the borates formed in the NaBH4 regeneration is a technology that needs urgent improvement, and it is essential to close the NaBH4-H2 cycle and, consequently, implement NaBH4-H2-PEMFC on-demand systems. After all, these systems may be the future off-grid technology that would address the increasing renewable energy supply. Funding: This work was supported by the "Programa Operacional Espaço Atlântico; Fundo Europeu de Desenvolvimento Regional (FEDER)" through the Transnational Cooperation Project Hylantic-Atlantic Network for Renewable Generation and Supply of Hydrogen to promote High Efficiency (EAPA_204/2016). FEDER also supported this work via CEFT.