Device for Controlled Production of Hydrogen

: In this work, the production of hydrogen from the sodium borohydride (NaBH 4 ) reaction was studied using an experimental bench test in a passive device operating with or without minimal external energy input. The system consists of a reactor in which a mixture based on sodium boro-hydride powders and an organic acid is conﬁned. A ﬂow of water feeds the area in which the solid mixture is conﬁned, which undergoes a hydrolysis reaction and this generates gaseous hydrogen. The hydrogen thus produced, already saturated with water vapor, is particularly suitable for feeding polymer electrolyte fuel cells for the production of electricity because it does not require further humidiﬁcation. The borohydride–organic acid coupling studied for this device, and its chemical process, provides high reaction and conversion kinetics, presenting remarkable chemical stability over time.


Introduction
Hydrogen is commonly thought of as a source of energy; in fact, it must be extracted from water (by electrolysis) or from fossil fuels (by reforming), which in both cases involves an expenditure of power.For this reason, hydrogen is not a source but a "carrier" or an "accumulator" of energy.On the other hand, it is becoming commonplace to downgrade hydrogen to a simple "form" of energy by not considering all the potential and advantages that would derive from using the best technologies to produce and use hydrogen [1].The best sources of energy are renewable ones (solar, photovoltaic, hydroelectric, etc.), which, however, have the drawback of depending on random variables such as the sun, wind, water flow, etc.Since they supply energy in an absolutely discontinuous way, in order to exploit them it is necessary to store the excess energy produced when the environmental conditions allow it (strong wind, high sun, etc.) to then reuse it when production decreases (absence of wind, rivers with non-optimal flow, low sun) [2].The transformation of the excess energy produced into hydrogen, and its storage, could expand the use of these sources.In addition, the hydrogen produced in surplus could also be used to obtain other chemical and/or industrial products such as ammonia (today it is obtained from petroleum hydrogen mainly to produce fertilizers), methanol (today it is obtained from petroleum), etc., thus obtaining savings in the use of fossil fuels, exhaustible and highly polluting sources.The main obstacles to the widespread use of hydrogen as an energy vector are related to unsolved problems in the methods of accumulation, transport and distribution, and in the choice of the optimal energy production system.One of the most promising systems for the use of hydrogen is represented by fuel cells that, by being fed with this gas, are able to directly produce electricity with high conversion efficiencies [3][4][5][6][7][8].Hydrogen can be stored and accumulated in several ways, as compressed or liquefied gas [9] contained in metal hydrides and in chemical hydrides (NaBH 4 , KBH 4 , LiH, NaH) [10][11][12][13][14][15][16][17][18].The latter technology [19][20][21], through the use of alkali borohydrides in particular, could solve the major problems of the accumulation, transport and distribution of hydrogen, at least in mobile and/or portable applications [22].In this case, the hydrogen is trapped in the chemical bonds of the boron and the alkali metal, forming a salt.An aqueous solution Hydrogen 2023, 4 435 composed, for example, of 50% sodium borohydride and 50% water, gives, by reacting on a suitable catalyst, hydrogen with an energy ratio similar, by volume, to petrol.In addition to hydrogen, the other product of the aforementioned reaction is sodium borate, a natural chemical compound commonly used in detergents which can be transformed back into sodium borohydride.
When borohydride is dissolved in water at sufficiently low acidity values (pH), hydrogen is produced which can be sent directly to the anode of a fuel cell thus obtaining electricity.Borohydride powder, if stored in the absence of humidity, is stable, non-flammable, non-explosive, does not produce polluting emissions or unwanted by-products and, finally, it is an excellent hydrogen storage system at room temperature for use in portable applications [23,24].
Sodium borohydride is a thermally stable and hygroscopic crystalline white salt which decomposes by hydrolysis according to the exothermic (∆H = 217 kJ mol −1 ) reaction: The following equation [25] conveniently expresses the rate of decomposition of borohydride aqueous solutions in its half-life (time required for the hydrolysis of 50% of the initial borohydride) as a function of pH and temperature: where t 1/2 is in minutes and T is in degrees Kelvin.Thus, the rate of decomposition can be controlled by varying the acidity and/or temperature.The complete hydrolysis of 1 g of borohydride, releases 2.374 L of hydrogen under normal conditions with a kinetic that slows down in a short amount of time due to the increase in pH caused by the formation of the basic metaborate salt.By maintaining a constant adequate acidity, it is therefore possible to completely transforming the salt into hydrogen by direct hydrolysis according to reaction (1).There are various works which describe the production of hydrogen, starting from borohydride according to reaction (1), using metal-based catalysts such as, for example, ruthenium, nickel and its alloys, and cobalt and its oxides [26][27][28].In all articles, an aqueous solution of borohydride with suitable quantities of added sodium or potassium hydroxide was used for the purpose of stabilizing the borohydride itself and preventing its spontaneous decomposition.Indeed, James et al. [29] claimed that appropriate salts of metals such as nickel, cobalt and iron, and of noble metals such as ruthenium, platinum, rhodium, iridium, etc., are effective catalytic accelerators of the reaction (1) in basic solutions.However, all these works highlight the problem of the chemical instability of the catalysts themselves, which tend to deactivate over time.The only exception seems to be the one described in the article by Pozio et al. [30] in which the instability problem is solved by using magnetic catalysts (cobalt, nickel or alloys) blocked inside a magnetic field of suitable intensity.There are also works in which the metal catalysts reported above are mixed with borohydride in a solid state and only subsequently combined with water as, for example, in the system described in Prosini et al.'s patent [31] in which nickel acetate is used.This expedient of using solid solutions essentially serves to avoid the use of basic solutions and their relative instability and encumbrance.In this case, however, the catalyst must be thoroughly mixed with the borohydride and this leads to a reduced yield.As an alternative to the use of catalysts that favor reaction 1, there are works that exploit the direct reaction between a basic aqueous solution of borohydride and an acidic aqueous solution [32][33][34][35][36].The management of an acid solution flow is, however, quite problematic due to problems related to corrosion.
In this article, an innovative device for the controlled production of hydrogen from borohydride is described in order to overcome the highlighted technical obstacles.The device operates by a method that has technical characteristics intended to avoid the disadvantages of the prior studies while, at the same time, operating without any or with minimal external energy supply, with dimensions that may be small in weight and size.A mixture of a solid organic acid and borohydride was used and added directly to pure water.The organic acid under consideration in the present article has a minimum length of its hydrocarbon chain of C2, must be solid under normal conditions so as to be mixable with the solid borohydride, and must be very soluble in water.In this work, we tested oxalic and citric acid, but tartaric, ascorbic or other organic acids that have a high number of carboxyl (COOH) functional groups could also be used.The organic acid most suitable depends on technical reasons or on environmental safety.As an example, the use of citric acid has the advantage that the waste product of the reaction (metaborate and sodium citrate) hydrolysis can be released to the environment without pollution.On the other hand, the use of oxalic acid helps reduce the weight of the reactants, but its sodium oxalate product is toxic and cannot be dispersed in the environment.The device configuration using solid acid/borohydride mixture highly simplifies the production system with respect to others options.In fact, it offers considerable advantages in terms of weight and a geometry that make it easily portable and integrated, able to operate with water coming from the normal water supply system or even with water of low purity and involving the use of particularly low-cost organic acids.

Static and Kinetic Test
The measurements were carried out both statically [37] using the configuration shown in Figure 1 and dynamically using the purpose-designed reactor.In the first case, the static tests were carried out by injecting water directly into a mixture of solid sodium borohydride (Aldrich, St. Louis, MO, USA) and oxalic acid 1).The components were thoroughly mixed together and placed in a closed 50 mL flask equipped with a syringe, a small glass tube that conveys the generated gas into a graduated container full of water, upside down and with the mouth dipped into the water held in a third container (Figure 1).The flask was kept at 25 • C and then an amount of water was added so as to measure the hydrogen contained in the hydride.The pH of the produced solution was measured with a pH meter (AMEL, Milan, Italy mod.338).
minimal external energy supply, with dimensions that may be small in weight and size.A mixture of a solid organic acid and borohydride was used and added directly to pure water.The organic acid under consideration in the present article has a minimum length of its hydrocarbon chain of C2, must be solid under normal conditions so as to be mixable with the solid borohydride, and must be very soluble in water.In this work, we tested oxalic and citric acid, but tartaric, ascorbic or other organic acids that have a high number of carboxyl (COOH) functional groups could also be used.The organic acid most suitable depends on technical reasons or on environmental safety.As an example, the use of citric acid has the advantage that the waste product of the reaction (metaborate and sodium citrate) hydrolysis can be released to the environment without pollution.On the other hand, the use of oxalic acid helps reduce the weight of the reactants, but its sodium oxalate product is toxic and cannot be dispersed in the environment.The device configuration using solid acid/borohydride mixture highly simplifies the production system with respect to others options.In fact, it offers considerable advantages in terms of weight and a geometry that make it easily portable and integrated, able to operate with water coming from the normal water supply system or even with water of low purity and involving the use of particularly low-cost organic acids.

Static and Kinetic Test
The measurements were carried out both statically [37] using the configuration shown in Figure 1 and dynamically using the purpose-designed reactor.In the first case, the static tests were carried out by injecting water directly into a mixture of solid sodium borohydride (Aldrich) and oxalic acid H2C2O4 × 2H2O (Aldrich) or citric acid C6H8O7 × H2O (Aldrich) (Table 1).The components were thoroughly mixed together and placed in a closed 50 mL flask equipped with a syringe, a small glass tube that conveys the generated gas into a graduated container full of water, upside down and with the mouth dipped into the water held in a third container (Figure 1).The flask was kept at 25 °C and then an amount of water was added so as to measure the hydrogen contained in the hydride.The pH of the produced solution was measured with a pH meter (AMEL, Milan, Italy mod.338).The kinetics tests (K-1, K-2 and K-3) were performed with a designed device.This was a reactor constituted by a cylindrical shape polyamide vessel with a height of 22 cm, a diameter of 8 cm and an empty weight of 400 g (Figure 2).It consisted, essentially, of two distinct sides, an upper tank (A) capable of housing water and an inferior reactor chamber (B) in which a mixture of solid borohydride and solid organic acid (C) was present.Water was poured in the upper tank through a filling nozzle (D) arranged on the top wall and equipped with a vent hole which kept the upper tank at a constant atmospheric pressure.The water poured in the upper tank did not necessarily need to be of a particular degree of purity.The upper tank communicated with the lower side via an opening (E), through which water flowed via a manual needle valve (F).Finally, the reactor chamber comprised a conduit (G) through which the hydrogen produced by the hydrolysis reaction escaped from the device.The kinetics tests (K-1, K-2 and K-3) were performed with a designed device.This was a reactor constituted by a cylindrical shape polyamide vessel with a height of 22 cm, a diameter of 8 cm and an empty weight of 400 g (Figure 2).It consisted, essentially, of two distinct sides, an upper tank (A) capable of housing water and an inferior reactor chamber (B) in which a mixture of solid borohydride and solid organic acid (C) was present.Water was poured in the upper tank through a filling nozzle (D) arranged on the top wall and equipped with a vent hole which kept the upper tank at a constant atmospheric pressure.The water poured in the upper tank did not necessarily need to be of a particular degree of purity.The upper tank communicated with the lower side via an opening (E), through which water flowed via a manual needle valve (F).Finally, the reactor chamber comprised a conduit (G) through which the hydrogen produced by the hydrolysis reaction escaped from the device.As soon as the water encountered the solid borohydride/organic acid mixture, reaction 1 was triggered and the hydrogen gas produced flowed outwards through the duct.The organic acid acted as a buffer system in such a way that avoided the increase in pH caused by basic hydrolysis.By mixing the borohydride powder with a defined quantity As soon as the water encountered the solid borohydride/organic acid mixture, reaction 1 was triggered and the hydrogen gas produced flowed outwards through the duct.The organic acid acted as a buffer system in such a way that avoided the increase in pH caused by basic hydrolysis.By mixing the borohydride powder with a defined quantity of a suitable studied solid organic acid, it was thus possible to obtain a constant production of hydrogen.In fact, the reaction of water with borohydride (1) was associated with the hydrolysis of the organic acid itself, which kept the pH at a constant value and thus allowed rapid kinetics.The residue obtained was obviously a concentrated solution of sodium metaborate mixed with non-polluting organic acid and its final pH was neutral In the first kinetic test, 0.8 g of solid sodium borohydride (Aldrich) and 1.44 g of citric acid C 6 H 8 O 7 × H 2 O (Aldrich) were thoroughly mixed together and placed in the reactor chamber (B).An amount of 3.2 mL of distilled water was placed in the upper tank (A) and the needle valve was regulated to ensure a flow of water equal to 1.6 mL min −1 .Two additional hydrogen production tests were conducted, using the same amount of solid mixture but a different water flow rate.In particular, in each sample, the reactor was loaded with 28 g of solid reagents and 100 mL of water (Table 2).

Static Tests
In the first static test (S-1), 1.00 g of water was injected in the closed flask (Figure 1) containing the mixture of borohydride/oxalic acid and, in less than 60 s, a complete development of the hydrogen contained in the hydride (474 mL under normal conditions) was observed.The final product resulted in a pasty semi-solid phase of sodium borate and sodium oxalate.
In the second experiment (S-2), 0.80 g of water was injected in the closed flask containing the mixture, and, in this case, a complete development of the hydrogen contained in the hydride was also observed, while the final product resulted in a homogeneous but very viscous solution, with a final pH of about 7.5.
In the third test (S-3), 1.00 g of water was injected in the closed flask containing borohydride/citric acid.In this case, a 100% yield of hydrogen was consistently obtained but, in addition, the final product was a fluid and homogeneous solution, with a pH between 6.5 and 7.0.
From these data, a minimum weight ratio of water/borohydride of four is needed in order to produce H 2 with a yield of 100% and a ratio of five is needed to obtain a final homogeneous solution of borate and sodium salt with a neutral pH between 6.5 and 7.0, which is disposable in the environment.

Kinetic Test
The graph in Figure 3 shows the volume of hydrogen produced as a function of time under the conditions described above in the K-1 experiment (0.8 g NaBH 4 ).In 120 s, approximately two liters of hydrogen were produced, equal to around 15 mL of hydrogen per second.The pH of the borate/organic acid solution after the test was 7.5.

Figure 3.
Hydrogen volume produced vs. time () using the experimental apparatus with 0.8 g NaBH4, 1.6 citric acid and 3.2 mL of water at 25 °C.Regression line (---) is also showed.Two additional tests were carried out with large amounts of solid salts (10 g) and citric acid (18 g), but with two different water flows (Table 2).Figure 4 shows the two curves relating to the hydrogen produced as a function of time in the respective tests K-2 and K-3.In the K-2 example, the weight ratio water/borohydride used was around 5.5 while, in example K-3, the ratio was exactly equal to 4.0.In both examples, the solid mixture consisting of NaBH4 and citric acid had a height of 1.5 cm, or a volume of about 75.4 mL.The temperature during example K-2 increased to 62-70 °C while, in example K-3, it increased to 74 °C.Two additional tests were carried out with large amounts of solid salts (10 g) and citric acid (18 g), but with two different water flows (Table 2).Figure 4 shows the two curves relating to the hydrogen produced as a function of time in the respective tests K-2 and K-3.In the K-2 example, the weight ratio water/borohydride used was around 5.5 while, in example K-3, the ratio was exactly equal to 4.0.In both examples, the solid mixture consisting of NaBH 4 and citric acid had a height of 1.5 cm, or a volume of about 75.4 mL.The temperature during example K-2 increased to 62-70 • C while, in example K-3, it increased to 74 • C. Figure 3. Hydrogen volume produced vs. time () using the experimental apparatus with 0.8 g NaBH4, 1.6 citric acid and 3.2 mL of water at 25 °C.Regression line (---) is also showed.Two additional tests were carried out with large amounts of solid salts (10 g) and citric acid (18 g), but with two different water flows (Table 2).Figure 4 shows the two curves relating to the hydrogen produced as a function of time in the respective tests K-2 and K-3.In the K-2 example, the weight ratio water/borohydride used was around 5.5 while, in example K-3, the ratio was exactly equal to 4.0.In both examples, the solid mixture consisting of NaBH4 and citric acid had a height of 1.5 cm, or a volume of about 75.4 mL.The temperature during example K-2 increased to 62-70 °C while, in example K-3, it increased to 74 °C.In both examples, the residue had a volume of 35 to 38 mL and was in the form of a liquid-solid mixture of a similar density to honey.The theoretical amount of hydrogen which could be produced was 23.74 L (c.n.) and, at the end, a yield of approximately 100% was obtained experimentally.
When the water drained through the reactor, the citric acid and the borohydride dissolved simultaneously only at the point of contact.In this area, there was a high concentration of the two components with a decrease in pH.The acidity was such that the development of hydrogen was immediate and continuous.Controlling the kinetics of the hydrolysis reaction was essential to account for this type of hydrogen production.Citric acid is a weak triprotic carboxylic acid, with pK a1 3.13, pK a2 4.76 and pK a3 6.34.In addition, the system citric acid/sodium citrate acted as a buffer that adjusted the pH from 3 to 6.2.This means that the kinetics of the hydrolysis reaction depended on the pH (eq.2) in a controlled way.By adding water, the local acidity decreased rapidly but not excessively, as in the case of adding a solution of hydrochloric acid.At the same time, borohydride hydrolysis produced sodium citrate that buffered the system, preventing excessive pH rise, thus stopping the hydrolysis reaction itself.In addition, the control of the hydrolysis reaction avoided the production of excessive heat that would have led to an excessive increase in the temperature by the evaporation of the water itself.
In the same area of water fall, however, a metaborate product was formed which can trap two molecules per boron atom (NaBO 2 × 2H 2 O) that precipitate being in conditions of supersaturation.
In that state, the system was in a state of generating feedback.The water generated hydrogen and, at the same time, limited the production.This effect did not occur in the case where an acid solution (e.g., HCl) was added to the solid borohydride.The trend here showed production with discontinuous peaks followed by slowdowns.Ondemand hydrogen systems should preferably have a start-up and shut-down dynamic to provide on/off control for hydrogen generation.This feature was illustrated in the borohydride/citric acid solid system where the control was strictly linked to the water flow.A fine manual valve can be used for water flow control, or better, a small pump directly controlled by the fuel cell control system.
The results permitted the evaluation, in 87.5 L, of the production of hydrogen for a device comprising an upper tank with a capacity of around 140 cc of water.For such a production of hydrogen, it would be necessary to fill the reactor with about 34 g of NabH 4 and 60 g of organic acid.Considering also the total weight of the device (400 g) and the estimated weight of the reagents (234 g), an energy density of the device as a whole of 379 Wh kg −1 can be obtained (126 L under normal conditions per Kg).The hydrogen output rate is determined exclusively by the needle valve and it is therefore necessary to use a valve with the finest possible regulation in order to obtain the gas flows necessary for the selected application in order to interrupt the water delivery.
Two important considerations concern the fact that (1) tap water can easily be used, (2) although the organic acid must be introduced in the same way as a reagent (its recovery after the reaction is, in fact, unthinkable); however, its cost is completely negligible (less than 10 euros per kg).
Finally, in terms of application, its ideal location is expected to be in the production of hydrogen for the small fuel cell cells currently on the market, with a power of 10-100 Watts, that are useful for powering portable electronic devices such as computers, PDAs, cell phones, etc.For such an application, 100-1000 cc/min is required.By properly calibrating the outgoing hydrogen stream, a variable range between 3 and 12 h could be easily guaranteed.The described hydrogen production system can be actively managed with a small energy consumption.In fact, in the case of the device used in the experiment, the water supply was about 1.4-1.9mL min −1 .For example, there are, on the market, piezoelectric pumps that can control liquid flows of 1-5 mL min −1 with a consumption of 0.25 Watt and a weight of just 2 grams.There are also micropumps that allow dosing up to 50 nL min −1 of liquid; this means, in practice, the possibility of descending to extremely low hydrogen production flows for miniature applications.The control of the micropump could, moreover, be slaved to a system for controlling the entire energy generator, hydrogen generator and fuel cell, so as to optimize the performance and hydrogen output.
The device tested can be further improved from a conceptual perspective by considering the specifications required to power a fuel cell.By way of example, let us imagine feeding a small stack in "dead end" mode with an improved device, i.e., the anode duct is closed and the gas is consumed only if the current is supplied.Since our system supplies a cell in "dead end" mode, it will be necessary that, in a completely passive way, the supply of hydrogen stops as soon as the cell stops supplying a current.For this purpose, the diagram of the device described above (Figure 2) can be modified by inserting a check valve or non-return (A) located immediately downstream of the adjustment seat of the flow regulator valve (Figure 5).The production of hydrogen in the lower compartment, where the borohydride is present, produces an increase in pressure increase and sends hydrogen gas to the cell, whose anodic side is intercepted by a valve located downstream of the "deadend" point (B).In the anode compartment, the gas is consumed by the electrochemical reaction in the cell or stagnates, raising the pressure in the circuit.
could, moreover, be slaved to a system for controlling the entire energy generator, hydrogen generator and fuel cell, so as to optimize the performance and hydrogen output.
The device tested can be further improved from a conceptual perspective by considering the specifications required to power a fuel cell.By way of example, let us imagine feeding a small stack in "dead end" mode with an improved device, i.e., the anode duct is closed and the gas is consumed only if the current is supplied.Since our system supplies a cell in "dead end" mode, it will be necessary that, in a completely passive way, the supply of hydrogen stops as soon as the cell stops supplying a current.For this purpose, the diagram of the device described above (Figure 2) can be modified by inserting a check valve or non-return (A) located immediately downstream of the adjustment seat of the flow regulator valve (Figure 5).The production of hydrogen in the lower compartment, where the borohydride is present, produces an increase in pressure increase and sends hydrogen gas to the cell, whose anodic side is intercepted by a valve located downstream of the "dead-end" point (B).In the anode compartment, the gas is consumed by the electrochemical reaction in the cell or stagnates, raising the pressure in the circuit.The increase in pressure in the cell and in the generator has the consequence of first slowing down and then stopping the flow of water from the tank above.The increase in the pressure in the lower part of the generator up to the predefined value of the differential pressure of the check valve (crack pressure), leads to the closure of the pipe that supplies water to the chemical hydrides and the arrest of the reaction.A vent valve located on the tank (not shown) can discharge any limited increase in pressure due to there being no reacted water in the hydride compartment.As the cell works again, the pressure in the anode compartment, and therefore in the generator, drops again, the non-return valve will be able to reopen, the water able to go down and the reaction able to proceed.A filter can also be placed before H2 goes into the anodic side in order to stop any possible salt particulate.

Conclusions
A hydrogen production system for portable energy applications, based on the hydrolysis of a NaBH4 solution, has been researched and developed.The system consists of increase in pressure in the cell and in the generator has the consequence of first slowing down and then stopping the flow of water from the tank above.The increase in the pressure in the lower part of the generator up to the predefined value of the differential pressure of the check valve (crack pressure), leads to the closure of the pipe that supplies water to the chemical hydrides and the arrest of the reaction.A vent valve located on the tank (not shown) can discharge any limited increase in pressure due to there being no reacted water in the hydride compartment.As the cell works again, the pressure in the anode compartment, and therefore in the generator, drops again, the non-return valve will be able to reopen, the water able to go down and the reaction able to proceed.A filter can also be placed before H 2 goes into the anodic side in order to stop any possible salt particulate.

Conclusions
A hydrogen production system for portable energy applications, based on the hydrolysis of a NaBH 4 solution, has been researched and developed.The system consists of a reactor where a mixture of sodium borohydride powders and an organic acid is contained and fed with water.The prototype tested exhibited some very interesting features: • a high chemical stability of the solid mixture used, resulting in the safe storage and transport for mobile production systems; • reaction products that have a neutral pH and can be disposed of without any particular problems or chemical and biological hazards; • reagents that have minimal or reduced toxicity; • extremely low water usage and a minimal footprint of the production; • a reaction conversion close to 100%; • practical absence of catalysts; • use of water of low quality and purity, avoiding the need to carry water refills (possible use of wastewater); • a passive production system or one with a very limited energy consumption, selfregulating in terms of pressure and capable to produce hydrogen on demand.
The hydrogen thus produced, already saturated with water vapor, is particularly suitable for feeding small polymer electrolyte fuel cells for the production of electricity because it does not require further humidification.

Figure 5 .
Figure 5. Modified experimental H2 generator connected to a fuel cell: check valve (A), dead end valve (B).

Figure 5 .
Figure 5. Modified experimental H 2 generator connected to a fuel cell: check valve (A), dead end valve (B).