H2O Hydrolysis to Generate

Jul 20, 2015 - Experiments using the CoCl2 catalyzed NaBH4/H2O hydrolysis were conducted at ambient and cold temperatures as well as under pressure. T...
11 downloads 7 Views 5MB Size
Download PDF
Article pubs.acs.org/IECR

Feasibility Studies of the NaBH4/H2O Hydrolysis to Generate Hydrogen Gas to Inflate Lighter than Air (LTA) Vehicles P. A. Mosier-Boss,* C. A. Becker, G. W. Anderson, and B. J. Wiedemeier SPAWAR Systems Center Pacific San Diego, California 92152, United States ABSTRACT: Experiments using the CoCl2 catalyzed NaBH4/H2O hydrolysis were conducted at ambient and cold temperatures as well as under pressure. The purpose of these experiments was to (1) provide guidance in designing a portable reactor capable of producing 330 standard cubic feet (SCF) of hydrogen gas and (2) determine how to control the reaction so that it could be used to generate hydrogen gas to inflate lighter than air (LTA) vehicles.

1. INTRODUCTION Lighter than air (LTA) vehicles support a variety of payloads at altitudes ranging from a few hundred feet to several thousand feet. Depending upon the payload, they have a number of potential uses in surveillance, reconnaissance, force protection, border/port security, over the horizon (OTH) communication relays, incident awareness and assessment, and monitoring air quality and atmospheric conditions. Because of their great utility, the use of LTA vehicles for military and commercial applications is going to increase. The two primary lifting gases used in LTAs have been hydrogen and helium.1,2 One advantage of hydrogen over helium is cost. Helium is used in a number of manufacturing processes.1,3,4 Since these processes have increased worldwide, the demand for helium has increased dramatically. Unfortunately, helium is a nonrenewable source. As demand has increased, supplies worldwide have dwindled, driving prices up. Besides cost, other advantages of hydrogen over helium are (1) hydrogen has more lift than helium so larger payloads are possible and (2) hydrogen can be stored in a chemically dense form as a metal hydride. The later advantage is useful for those missions that require long-term storage of the hydrogen as storing helium/hydrogen in tanks is not practical due to leakage, safety, and size concerns. For the above reasons, we began to explore the use of metal hydride generated hydrogen gas to inflate LTAs. Our objective is to build a reactor capable of generating 330 SCF of hydrogen gas. Of the metal hydrides, we chose to use the sodium borohydride chemistry for this purpose. This chemistry was chosen because of the stability of sodium borohydride in alkaline solution and its byproduct (sodium metaborate) is relatively environmentally benign. Also the sodium borohydride chemistry has been used in hydrogen generators that were integrated with fuel cells.5−9 The chemical reaction between NaBH4 and H2O is as follows:10 NaBH4 + (2 + x)H 2O → NaBO2 ·x H 2O + 4H 2↑

RuCl3, RhCl3 > H 2PtCl 6 > CoCl 2 > NiCl 2, OsO4 > IrCl4 > FeCl 2 > > PdCl 2

While RuCl3, RhCl3, and H2PtCl6 out-perform CoCl2, these salts are comparatively very expensive. Others have also used cobalt-based catalysts to accelerate sodium borohydride hydrolysis.12−15 Given the volume of hydrogen gas needed to inflate an aerostat, CoCl2 was chosen to be the catalyst in these experiments. Additional constraints on the reactor are to minimize the amount of water so as to reduce weight, be able to operate between 5 and 25 °C, generate hydrogen gas at a reasonable and constant rate, to keep the temperature of the reagents below 60 °C to minimize the amount of water vapor entering the LTA, and to keep the temperature of the generated gas below 60 °C so as to not damage the LTA. Smaller scale experiments were conducted in the laboratory to determine how to meet these constraints. This communication discusses the results of those experiments.

2. EXPERIMENTAL SECTION 2.1. Chemicals. Sodium borohydride (GFS Chemicals, CAS 16940-66-2), anhydrous cobalt(II) chloride (SigmaAldrich, CAS 7646-79-9), and distilled water (Arrowhead, CAS 7732-18-5) were used as received. The percent CoCl2 used in these experiments is calculated relative to the mass of NaBH4. In the presence of NaBH4 and H2O, CoCl2 reacts to form the catalyst, Co2B:16 BH4 − + 2Co2 + → Co2B +

(2)

17

While the catalyst, Co2B, can be reused, no attempts to do so were done in these experiments as the intended application of the hydrolysis reaction was single use. 2.2. Experimental Apparati to Measure Hydrogen Gas Generation under Ambient Conditions. For water volumes of 50 mL, the reaction was carried out in a 250 mL tapered wall

(1)

where x represents the excess hydration factor. In the absence of a catalyst, this reaction goes very slowly. Brown and Brown11 examined a number of metal salts to catalyze this reaction. They showed that the reaction rate as a function of catalyst type was as follows: This article not subject to U.S. Copyright. Published XXXX by the American Chemical Society

1 H 2 + 3H+ 2

Received: May 4, 2015 Revised: July 10, 2015 Accepted: July 20, 2015

A

DOI: 10.1021/acs.iecr.5b01647 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Industrial & Engineering Chemistry Research

Article

Figure 1. (a) Photograph of the glass reaction vessel used to generate 50−100 L of hydrogen gas at ambient pressure where SS is stainless steel and PVC is polyvinyl chloride. (b) Photograph of the reaction vessel as well as the sensors used to monitor the hydrogen generation reaction.

Figure 2. (a) Photograph of the glass reaction vessel used in experiments conducted under pressure. The dashed line indicates the level of the reagents and the asterisks indicate the placement of the thermocouplesT1, T2, T3, and T4, from lowest to highest. (b) Schematic of the pressure can. Not shown is the manual release valve and the 200 psig burst disc.

flask that had three necks with 24/40 ground glass joints and a fourth neck with a #7 Ace thread (Ace Glass, p.n. 6963−34). Temperature measurements were made using a K-type thermocouple. This thermocouple was placed in a thin glass sleeve that was immersed in the reagents through the #7 Ace thread. A known amount of sodium borohydride was then

placed inside the flask followed by 49 mL of water. To initiate the reaction, a 1 mL sample containing a known amount of CoCl2 was pipetted into the flask. Data collection was done under computer control. To generate 50−100 L of hydrogen gas, the reaction vessel shown in Figure 1a was used. As the catalyzed reaction between B

DOI: 10.1021/acs.iecr.5b01647 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Industrial & Engineering Chemistry Research

Article

water and NaBH4 is exothermic, water flowing through the copper cooling coils is used to remove the heat generated during the reaction. Either a peristaltic pump or a bilge pump, with a check valve to keep the water flowing in one direction, was used to pump cooling water through the copper coils. Use of the peristaltic pump made it possible to adjust the rate of water flow through the coils and study the effect of water flow rate on both the cooling and the reaction rate. The bilge pump enabled automatic control, using a computer, of the cooling. For automatic control of the cooling, the temperature of the reaction was monitored using a thermocouple inside the reaction vessel. The cooling would then be turned on and off when the reaction reached a given temperature set point. Figure 1b shows that additional K type thermocouples monitored the inlet and outlet water temperatures of the copper cooling coils. An Omega FMA4000 Digital Mass Flow Meter was placed on the gas outlet to measure the flow rate of the hydrogen gas that was generated. A 50 μm filter was placed between the reaction vessel and the flow meter to prevent particulates and water vapor from entering the flow meter. Data collection was done under computer control. To conduct an experiment, a known amount of CoCl2 was placed on the bottom of the graduated cylinder. Enough NaBH4 to generate 50−100 L of hydrogen gas was then placed in the stainless steel basket. A 1 in. diameter PVC rod in the center of the stainless steel basket was used to concentrate the reaction near the coils. This stainless steel basket was then placed within the Cu coils as shown in Figure 1a. A K type thermocouple (Omega), Figure 1b, was placed inside the reaction vessel. This thermocouple was in contact with the reagents during the reaction and was used to monitor the temperature of the reagents. The ratio of the water to NaBH4 was 4.6 by mass. In some experiments, two K type thermocouples were placed inside the reaction vessel at different heights. To initiate the reaction, water was added to the glass reaction vessel through one of the ground glass joints. 2.3. Experimental Apparatus to Measure Hydrogen Gas Generation under Pressure. Figure 2a shows a photograph of the glass reaction vessel used in the experiments. Cooling water flows through the copper coils. A dashed line in Figure 2a indicates the initial level of the reagents. A plastic test tube is taped on the water outlet. This plastic test tube houses a thermocouple that is used to monitor the temperature of the water as it exits the cooling coils. A tape measure is epoxied outside the glass vessel. This tape measure is used to measure the level of the reagents during the course of the experiment. A glass sleeve housing four thermocouples at different heights is placed inside the glass reaction vessel as shown in Figure 2a. Asterisks indicate the position of the thermocouples. The lowest thermocouple is in contact with the reagents. This is also the thermocouple that is used to monitor/control the reaction. The thermocouple above the lowest one is initially at the reagent/air interface. The remaining two thermocouples are placed above the reagents in the air. Once assembled, the glass reaction vessel was placed on a stand inside the reaction can, shown schematically in Figure 2b. Lead weights, shown in Figure 2a, hold the glass reaction vessel in place. The reaction can is a 316 L stainless steel, three gallon capacity, portable, standard mouth ASME pressure tank (McMaster-Carr, p.n. 6778K21). The pressure tank has a 185 psig maximum pressure rating at 37.8 °C and a maximum operating temperature of 149 °C. An adjustable 10-228 psig proportional pressure release valve (Swagelok, RL3) and a

pressure gauge with a pressure transducer (Omega model PX835−300GI) are connected to the pressure tank. The pressure release valve is connected to the gas outlet and to a hose that goes to the Omega FMA4000 Digital Mass Flow Meter to measure the flow rate of the generated hydrogen gas. A 50 μm filter was placed between the reaction vessel and the flow meter to prevent particulates and water vapor from entering the flow meter. The pressure tank has pass-throughs for the thermocouples that are placed inside the glass reaction vessel and the plastic tube taped to the water outlet, Figure 2b. This experimental setup can be used to conduct testing under ambient conditions as well as under pressure. When pressure tests were done, the reaction can was pressurized using nitrogen gas. For safety, the pressure can also had a 200 psig burst disc and a manual emergency release valve on it, both of which are not shown in the schematic. As shown in Figure 2b, a tube inside the reaction can connects to the inlet of the copper coils of the glass reaction vessel. A hose on the water inlet connects this tube to a water can, which is not shown. The water can is another 316L stainless steel, three gallon capacity, portable, standard mouth ASME pressure tank (McMaster-Carr, p.n. 6778K21). A pressure gauge and a water pump are attached to the water can. The water can houses the cooling water that is pumped through the copper coils. The water can is connected to a nitrogen tank throughout the course of the experiment. It is maintained at a pressure that is ∼20 psig greater than the reaction can pressure. The water pump has an on/off switch. When the temperature inside the glass reaction vessel reaches a given set point, the pump switches on and pumps water from the water can through the copper coils in the reaction vessel inside the reaction can. Because the water running through the coils dumps into the reaction can, the glass reaction vessel was placed on top of a stand. Besides being able to conduct experiments at different pressures, this setup also allowed us to conduct experiments using either chilled or ambient temperature water for the cooling. All experiments and data collection were under computer control. In addition to temperature, pressure, and gas flow measurements, videos of the reactions were obtained using a GoPro camera and a LED light source inside the pressure can (not shown in the schematic). To start the reaction, a known amount of CoCl2 was placed inside a shallow polyethylene cup, which was then lowered into the stainless steel basket shown in Figure 2a. A known amount of NaBH4 was then added to the stainless steel basket. Monofilament fishing line was used to tie the basket to a metal screw that is attracted to magnets. A magnet on the outside of the pressure can, Figure 2b, holds the metal screw and the basket containing the reagents above 100 mL of water in the glass reaction vessel. Reaction starts when the magnet is pulled away dropping the stainless steel basket with reagents into the water.

3. RESULTS AND DISCUSSION 3.1. Effect Water Type. Depending upon mission scenarios, it may be desirable to use seawater instead of distilled or deionized water. The advantage for maritime applications is that, if seawater could be used, we would not have to carry water with us, which would result in weight reduction and would simplify the design of a hydrogen gas generator. Experiments were conducted using seawater and deionized water using the 250 mL tapered flask described vide supra. The results are summarized in Figure 3. For deionized C

DOI: 10.1021/acs.iecr.5b01647 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Industrial & Engineering Chemistry Research

Article

weight and affect the lift capacity of the aerostat/LTA vehicle. Also the foam is alkaline which could potentially cause chemical damage to the aerostat/LTA vehicle. Experiments were conducted at 25 °C and colder temperatures. 3.2.1. Continuous Flow Experiments Using Ambient Temperature Water for Cooling. In one series of experiments, we allowed water to flow through the cooling coils continuously prior to, and after, the addition of the water to the glass reaction vessel. It was found that, regardless of the water flow rate and the amount of catalyst, a continuous flow of water though the Cu coils resulted in so much cooling that the reaction was essentially quenched. We then tried a protocol where we allowed the reaction to reach a set temperature before turning on the flow of the cooling water and leaving it on. With the addition of water, the temperature of the reaction increased rapidly. When the temperature reached 45 °C, the cooling was turned on and was left on throughout the rest of the reaction. No foaming was observed in this reaction. With continuous cooling, the temperature of the reaction decreased. At the same time, a decrease in the flow rate occurred, indicating that continuous cooling was quenching the reaction. Upon termination of the data acquisition, it was observed that the reaction had not gone to completion as indicated by the total hydrogen (it went to ∼75% completion) and the fact that bubbles of hydrogen gas continued to be released long after data acquisition was stopped. Additional experiments were conducted with higher percentages of CoCl2 and with lower cooling water rates. Under these conditions, the reactions achieved ∼90% completion, however, foaming was considerable. Consequently, this method of cooling, while simple, did not meet all the criteria needed to control the reaction. The peristaltic pump has a dial which can be used to vary the water flow rate through the coils. An experiment was conducted where we allowed the reaction to reach a set temperature before turning on the flow of the cooling water through the Cu coils and leaving it on. However, instead of keeping the flow rate constant, we varied it manually. The results are summarized in Figure 4. After the addition of water to the reagents, the sharp spike, due to the formation of the Co2B catalyst, was observed in the flow rate, Figure 4b. The reaction temperature was observed to increase rapidly, Figure 4a. After 4.7 min from when the water was added to the reagents, the reaction temperature reached 50 °C, at which point the cooling water was turned on at a rate of 1 mL s−1. The temperature and gas flow rate remained constant, Figure 4a and b, respectively. Eighteen minutes after the water was added to the reagents, the temperature of the reaction began dropping below 50 °C. At this time, the water flow rate through the coils was reduced to 0.8 mL s−1, which caused the reaction temperature and gas flow rate to increase. When the temperature of the reaction reached 52 °C, the water flow rate through the coils was increased to 2.3 mL s−1. Despite this increased flow of cooling water through the Cu coils, the temperature continued to increase and reached a maximum of 60 °C, Figure 4a. Likewise, the hydrogen gas flow rate increased and maxed at 6.54 L min−1. The total generation of hydrogen gas was fairly linear, Figure 4c. The results in Figure 4 show that varying the flow rate of water through the cooling coils adequately controlled the reaction. 3.2.2. Cooling Using Ambient Temperature Water and Manually Toggling the Pump On/Off. We then began to conduct experiments where the flow of cooling water through the Cu coils was toggled on/off over a set temperature range. In

Figure 3. Plots of reaction temperature as a function of time when either deionized (DI) water, tap water, or seawater is used. Experimental conditions: 50 mL water, 7.5% CoCl2, and H2O/ NaBH4 ratio of 33.3:1.

water, the temperature of the reaction ramped up to 52 °C and the total reaction time was 5.6 min. For seawater, the temperature maxed at 30 °C and the reaction time was >30 min. An experiment was done using tap water. As can be seen in Figure 3, the temperature got up to 42 °C, and the reaction time was 11.8 min. The results indicate that something in tap water and seawater is poisoning the catalyst, Co2B. Addition of sodium chloride to the reagents had no effect on the reaction. Tap water contains 4 ppm chloramine18 and seawater contains dissolved organics that have carboxyl, amine, and sulfhydryl groups present.19 In polarography experiments done using tap water instead of distilled water, the half wave potentials of the transition metal ions were more negative than those observed in distilled water. This was attributed to complexation between the chloramine and the transition metal ions. In seawater, transition metals appear as both inorganic and organic species.20 Inorganic forms of the transition metals include hydrated metal ions and complexes with inorganic ligands. Organic forms of the transition metals include complexes with organic ligands such as proteins or humic substances. Since both chloramine and the dissolved organics can form complexes with transition metals, it is believed that these complexes impede the ability of Co2B to catalyze the reaction between NaBH4 and H2O. The results summarized in Figure 3 indicate that seawater cannot be used with transition metal catalysts. Since we will have to carry our own water when using the NaBH4/CoCl2 chemistry, we want to minimize the amount of water needed. Experiments were then conducted with deionized water to determine the minimum H2O/NaBH4 ratio that could be tolerated. The temperature of the reaction is dependent upon both the H2O/NaBH4 ratio and the % CoCl2 relative to NaBH4. Experiments showed that the minimum ratio of water to sodium borohydride, by mass, that can be tolerated is 4.6. Anything lower results in a very viscous sludge due to the precipitation of NaBO2 crystals. Increasing viscosity resulted in foaming and reaction temperatures approaching 90 °C. 3.2. 50−100 L Hydrogen Generation Experiments. Once the minimum H2O/NaBH4 ratio had been determined, experiments were conducted using the apparatus shown in Figure 1. The purpose of these experiments was to optimize the amount of catalyst and the cooling as well as to minimize foaming. Foam needs to be avoided as it would introduce water and chemicals into the aerostat which would increase the D

DOI: 10.1021/acs.iecr.5b01647 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Industrial & Engineering Chemistry Research

Article

reagents were fully in contact with the cooling coils at the beginning of the reaction, we began to do tests that would generate 50−60 L of hydrogen gas. Figure 5 summarizes the results of an active cooling experiment using automatic control at a set temperature of

Figure 4. Results obtained when water was flowed continuously through the cooling coils at a varying flow rate where (a) is a plot of reaction temperature, Trxn, and ΔT for the water outlet and inlet, (b) is a plot of the flow rate and (c) a plot of total hydrogen gas all as a function of reaction time. This was a 100 L experiment using a H2O/ NaBH4 ratio of 4.6:1 and 3% CoCl2. The cooling water was at ambient temperature and the peristaltic pump was used. Water flow rates, in mL s−1, are indicated. The set temperature was 50 °C.

Figure 5. Results obtained for automatic control of cooling using water at ambient temperatures and a water flow rate of 2.5 mL s−1 where (a) is a plot of reaction temperature, Trxn, and ΔT for the water outlet and inlet, (b) is a plot of the flow rate and (c) plot of total hydrogen gas all as a function of reaction time. This was a 60 L experiment using a H2O/NaBH4 ratio of 4.6:1 and 3.25% CoCl2. The set temperature was 50 °C.

a 100 L run using 4% CoCl2 relative to the NaBH4, a peristaltic pump was toggled on and off, manually, to keep the reaction temperature between 50 and 55 °C. It took 4.3 min for the reaction to get to 55 °C, at which time toggling the cooling on/ off was begun. The gas flow rate followed the reaction temperature and spiked at a flow rate of 14.2 L min−1 (at the same time the temperature maxed at 64.9 °C). This reaction foamed badly. The plot of total gas as a function of time was not linear indicating that the gas was not generated at a constant rate. This experiment was repeated using less CoCl2 (2.5%) and manually toggling the cooling to keep the temperature between 55 and 60 °C. Because less catalyst was used, it took 9.6 min for the reaction to reach 55 °C. Foaming was minimized, and the plot of total gas as a function of time was more linear. 3.2.3. Cooling Using Ambient Temperature Water and Automatically Shutting the Pump On/Off. As was discussed vide supra, two methods to control the cooling water going through the coils were explored−varying the water flow rate through the coils with continuous cooling and toggling on/off. To automate varying the water flow rate would require a flow meter and a valve to control flow rate whereas toggling on/off requires a temperature sensor and an on/off switch. Given that temperature sensors, such as thermocouples, are compact and inexpensive and on/off switches are simpler than valves, it was decided that toggling on/off would be easier to automate. We then began conducting experiments where the toggling on/off was automated. To do this, a bilge pump was used in place of the peristaltic pump and toggling on/off, for a given set temperature, was under computer control. As shown in Figure 1a, the Cu cooling coils are below the 400 mL mark on the glass cylinder. While conducting the 100 L tests, it was noticed that, when the water was poured into the reaction vessel, the reagents were initially at the 500 mL mark, which is above the cooling coils. Therefore, cooling would not be homogeneous. It was thought that this could be contributing to the foaming issues we were observing. To ensure that the

50 °C and 3.25% CoCl2. Figure 5a indicated that it took ∼5 min for the reaction to reach the set temperature, at which time the computer turned on the bilge pump to flow water through the coils. The number of saw-teeth indicated the number of on/ off cycles of the pump (22 in this case). The temperature of the water outlet became 14 °C hotter than the inlet indicating that the coils were removing heat from the reaction. The foam level was observed to remain at, or below, the levels of the Cu coils. The rate of hydrogen gas generation was nearly constant, Figure 5b, and the total gas as a function of time was linear, Figure 5c. This was a well controlled reaction. 3.2.4. Cooling Using Chilled Water and the Effect of Controlling Thermocouple Placement. As some deployments of LTAs would be done in colder climates, we began to conduct experiments using chilled cooling water. In these experiments, the bilge pump and hoses were immersed in an ice water bath, and the glass reaction vessel was also chilled by immersing it in an ice water bath for ∼1 h. Prior to initiating the reaction, the glass reaction vessel was taken out of the ice bath. The reaction was initiated by pouring chilled water into the reaction vessel. Foaming was a greater problem when using chilled water for cooling. It was observed that as the reaction proceeded, crystals of NaBO2 would come out of solution. The solution would become more viscous and layers would form. A reaction was imaged using an FLIR thermal camera. The images showed that the reaction temperature was not homogeneous and that the temperature was hotter at the air/liquid interface. Because of this inhomogeneity, we then did 60 L experiments placing two thermocouples inside the reaction vessel at two different levels to see what effect control thermocouple placement would have on the reaction profile. The lower thermocouple was placed at the 50 mL mark and the upper at the 200 mL mark in the E

DOI: 10.1021/acs.iecr.5b01647 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Industrial & Engineering Chemistry Research

Article

reagents (the liquid level started at the 250 mL level). In one experiment, the lower thermocouple was used to monitor/ control the reaction. The results are summarized in Figure 6. It

Figure 7. Results obtained for automatic control of cooling using chilled cooling water and a water flow rate of 3.2 mL s−1 where (a) is a plot of reaction temperatures, Trxn,upper and Trxn,lower, as well as water inlet, Tin, and outlet, Tout, temperatures, (b) is a plot of the flow rate, and (c) plot of total hydrogen gas all as a function of reaction time. This was a 60 L experiment using a H2O/NaBH4 ratio of 4.6:1 and 4.0% CoCl2. The set temperature was 45 °C, and the Trxn,upper thermocouple was used to monitor/control the reaction.

Figure 6. Results obtained for automatic control of cooling using chilled cooling water and a water flow rate of 3.0 mL s−1 where (a) is a plot of reaction temperatures, Trxn,upper and Trxn,lower, as well as water inlet, Tin, and outlet, Tout, temperatures, (b) is a plot of the flow rate and (c) plot of total hydrogen gas all as a function of reaction time. This was a 60 L experiment using a H2O/NaBH4 ratio of 4.6:1 and 4.0% CoCl2. The set temperature was 45 °C, and the Trxn,lower thermocouple was used to monitor/control the reaction.

not prevent foaming. Ways to alleviate it are to reduce the amount of CoCl2, decrease the reaction set point, or increase the amount of water for the reaction. These actions would also result in longer reaction times. Should weight be a constraining factor, increasing the amount of water may not be a viable option. There are antifoaming agents, but chemical methods of foam control may cause contamination problems as well as a reduction in mass transfer.21 We then began to explore mechanical methods of foam control. It is known that many oils foam when trapped gas is suddenly released under conditions of an abrupt drop in pressure.22 We then began to conduct tests under pressure to see if this would minimize the foaming. An added advantage of pressure is that it would reduce the amount of water vapor that will get into the gas stream. The apparatus shown in Figure 2 was used for these tests. The measurements that were made are described vide supra. 3.3.1. Experiment Conducted at Ambient Pressure Using Chilled Cooling Water. To establish a baseline for comparison, an experiment was done at ambient pressure. In this experiment, the glass reaction vessel was placed on top of a stand so that it was not in contact with the cooling water exiting the Cu coils. The results are summarized in Figure 8. Figure 8a shows that it took 2.2 min for the reaction to get to temperature before the cooling was activated. The temperature of thermocouple T1 in the reagents stayed steady at 45 °C. For 7 min, thermocouples T2, T3, and T4 tracked thermocouple T1. At the 7 min mark, the temperatures of thermocouples T2, T3, and T4 diverged from that measured by thermocouple T1. The maximum temperatures measured by thermocouples T2, T3 and T4 were 54, 62, and 65 °C, respectively. Images from the GoPro camera showed that 5 min into the reaction, the entire field of view was engulfed in foam. From this time until 14 min into the reaction, all four thermocouples were immersed in foam. Figure 8b shows the hydrogen gas flow rate as a function of time. The flow rate varied between 2 and 5 L min−1. The total flow rate as a function of time, Figure 8c, was nearly linear. Except for the foaming, this was a well controlled reation.

took 9 min for the reaction to reach the set temperature of 45 °C, at which point the computer turned on the bilge pump to flow chilled water through the coils. As shown in Figure 6a, the temperature of the lower thermocouple was maintained at 45 °C. It can also be seen that the two thermocouples tracked one another, until ∼20 min into the reaction. At this point, the temperature of the upper thermocouple began to increase. The temperatures registered by the upper thermocouple were 5−30 °C hotter than the lower thermocouple. At the maximum temperatures of the upper thermocouple, the foam reached the 500 mL mark. The flow rate, shown in Figure 6b, was fairly linear at ∼1.4 L min−1. However, ∼ 20 min into the reaction, the flow rate increased and maxed out at 5.7 L min−1. The increase in flow rate coincided with the increase in temperature measured by the upper thermocouple. The total flow as a function of time, Figure 6c, was nonlinear. The temperatures recorded by the upper thermocouple as well as the nonlinearity in the flow rate and total flow as a function of time indicated that this was not a well controlled reaction. In the second experiment, the upper thermocouple was used to monitor/control the reaction. Other experimental conditions were similar to those used in the experiment using the lower thermocouple for control. The results are summarized in Figure 7. It took almost 20 min to get to the set temperature. The temperatures stayed at, or below, 45 °C, Figure 7a. The reaction took longer to reach completion. The flow rate was fairly constant, Figure 7b, and the total flow as a function of time, Figure 7c, was nearly linear. Foaming did reach the 400 mL mark. Despite the foaming, this was a well controlled reaction. 3.3. Hydrogen Generation Experiments Conducted under Pressure. Foaming has proven to be a big issue in these experiments. Even experiments using the tapered wall flask did F

DOI: 10.1021/acs.iecr.5b01647 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Industrial & Engineering Chemistry Research

Article

for 8 min at which time the temperature recorded by T2 dropped below that of T1. This indicated that, 22.5 min into the reaction, T2 was above the solution/air interface. Most of the hydrogen gas generation occurred during the first 8 min of the reaction. The total flow rate as a function of time was nonlinear. Using the video obtained from the Go-Pro camera, the foam level was measured as a function of time. A plot of foam level as a function of time is shown in Figure 9a. There was an initial spike to the foam at 1.7 in due to the formation of the Co2B catalyst. Compared to the video for the experiment summarized in Figure 8, the foam did not fully engulf the field of view indicating that a pressure of 13 psig does suppress foaming. For the experiment conducted under a pressure of 50 psig, the set temperature was 45 °C. Thermocouple T1 was always in contact with the solution. It took less than 1 min for T1 to reach the set temperature at which time the cooling was activated. Even with active cooling, the temperature recorded by T1 shot past the set point and maxed at 61 °C. The temperature of T1 then came down and active cooling kept it at 45 °C. This temperature spike was not observed in the experiment done at an applied pressure of 0 psig, Figure 8a. Most of the hydrogen gas was produced during the first 3 min of the reaction, and the total flow as a function of time was nonlinear. Figure 9a shows a plot of foam level as a function of time. This plot showed that a pressure of 50 psig suppresses the foaming better than a pressure of 13 psig. The suppression of foam by pressure is aptly demonstrated by the photographs shown in Figure 9b and c. Figure 9b shows an image taken of a reaction at 50 psig that had not gone to completion. No foaming was observed. When the pressure was released, Figure 9c, the foam filled the camera’s field of view. These experiments showed that pressure did suppress the foaming. However, thermal control was not as good as in the absence of pressure. 3.3.3. Experiments Conducted under Pressure Using Chilled Cooling Water. In these experiments, the glass reaction vessel was placed on a stand inside the reaction can, Figure 2b. The experiment summarized in Figure 8 was done under no applied pressure and using chilled water for cooling. This experiment was repeated using a pressure of 50 psig. Videos taken using the Go-Pro camera inside the pressure can showed that foaming was suppressed. The results of this experiment are summarized in Figure 10. Figure 10a shows the temperature data. Like the experiment done under 0 psig pressure, it took ∼2.3 min for the T1 to reach the set temperature of 45 °C. However, unlike the 0 psig experiment, T1 continued to increase and maxed at 64 °C. The temperature at the air/ reagent interface maxed at 106 °C, and the temperatures in the

Figure 8. Results obtained for automatic control of cooling using chilled cooling water at 5.7 °C and a water flow rate of 4.6 mL s−1, where (a) is a plot of reaction temperatures, T1 through T4, as well as water outlet, Tout, temperature, (b) is a plot of the flow rate, and (c) plot of total hydrogen gas all as a function of reaction time. This experiment was done in the pressure canister at an applied pressure of 0 psig, using a H2O/NaBH4 ratio of 4.6:1 and 3.0% CoCl2. The set temperature was 45 °C and thermocouple T1 was used to monitor/ control the reaction.

3.3.2. Experiments Conducted under Pressure Using Ambient Temperature Water. In these experiments, the glass reaction vessel was not on a stand inside the reaction can, Figure 2b. Consequently, the glass reaction vessel would be in contact with the cooling water that exited the Cu coils. Experiments were conducted at applied pressures of 13 and 50 psig. For the experiment conducted under a pressure of 13 psig, the set temperature was initially 45 °C. Thermocouple T1 was always in contact with the solution. It took 1.1 min for T1 to reach the set temperature at which time the cooling turned on. Even with active cooling, the temperature recorded by T1 shot past the set point and maxed at 50 °C. The temperature of T1 then came down and active cooling kept it at 45 °C. Fourteen minutes into the reaction, the set temperature was changed to 50 °C. It took 3.5 min for T1 to reach 50 °C. The temperature of T1 stayed steady at 50 °C and, after 24 min, began to decrease due to completion of the reaction. The thermocouple placed near the reagent/air interface, T2, showed higher temperatures than T1 for the first 14 min maxing at 55 °C. When the set temperature was changed to 50 °C, T2 tracked T1

Figure 9. (a) Plot of foam level as a function of time for applied pressures of 13 (black) and 50 (gray) psig. Images of the same reaction at applied pressures of (b) 50 psig (dashed white line is at the 2.4 in mark) and (c) 0 psig. In (c), the foam completely fills the field of view. G

DOI: 10.1021/acs.iecr.5b01647 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Industrial & Engineering Chemistry Research

Article

Figure 10. Results obtained for automatic control of cooling using chilled cooling water at 5.4 °C and a water flow rate of 4.6 mL s−1 where (a) is a plot of reaction temperatures, T1 through T4; set temperatures, Tset, as well as water outlet temperature, Tout, (b) is a plot of the flow rate and (c) is a plot of the pressure, P, and total hydrogen gas all as a function of reaction time. This experiment was done in the pressure canister at an applied pressure of 50 psig, using a H2O/NaBH4 ratio of 4.6:1 and 3.0% CoCl2. Thermocouple T1 was used to monitor/control the reaction.

Figure 11. Results obtained for automatic control of cooling using chilled cooling water at 5.1 °C and a water flow rate of 4.6 mL s−1, where (a) is a plot of reaction temperatures, T1 through T4; set temperatures, Tset, as well as water outlet temperature, Tout, (b) is a plot of the flow rate and (c) is a plot of the pressure, P, and total hydrogen gas all as a function of reaction time. This experiment was done in the pressure canister at an applied pressure of 50 psig, using a H2O/NaBH4 ratio of 4.6:1 and 2.0% CoCl2. Thermocouple T1 was used to monitor/control the reaction.

air above the reagents maxed at 106 and 97 °C. The temperature of the water exiting the Cu cooling coils maxed at 73 °C, indicating that the coils were carrying away a great deal of heat. Figure 10b is a plot of flow rate as a function of time showing that most of the hydrogen gas was produced during the first 4 min of the reaction. Figure 10c shows plots of gas pressure and total hydrogen gas produced as a function of time. An increase in pressure was observed that coincided with the maximum flow rate and temperature. The total hydrogen gas generation was linear. At the end of the experiment, crystals were observed in the spent reagents. This indicated that, under pressure and using chilled cooling water, the NaBO2 comes out of solution. When this happens, it takes out water. With less water, the reaction gets hotter and a thermal runaway occurred as shown by the temperature measurements, Figure 10a. The experiment was then repeated using the same amount of CoCl2 and a lower set temperature to see if this would prevent the thermal runaway. When it was shown that this did not prevent the thermal runaway, the experiment was repeated using 2% CoCl2, instead of 3%, as well as using a lower set temperature. The results are summarized in Figure 11. Temperature measurements, Figure 11a, indicated that the reaction was kept under control. The highest temperature reached was 67 °C, and this was at the air/reagent interface. The temperatures above the interface were lower. Figure 11b shows the flow rate as a function of time. No flow was recorded during the first 11 min due to the fact that the pressure release valve was not functioning properly. However, Figure 11c shows a plot of gas pressure as a function of time. The gas pressure is increasing linearly during the first 11 min, indicating that the gas generation was fairly constant during this period. After 11 min, the pressure reached 54 psi and the release valve opened. When this happened, it can be seen that the flow rate, Figure 11b, was fairly constant and total hydrogen gas, Figure 11c, as a function of time, was linear. These results show that the reaction can be controlled and the hydrogen gas generated at a reasonable rate.

4. CONCLUSIONS Experiments were conducted using the CoCl2 catalyzed NaBH4/H2O hydrolysis reaction to generate hydrogen gas. In order for the reaction to complete in a reasonable time, either distilled or deionized water should be used. The minimum ratio of H2O/ NaBH4 that could be tolerated was 4.6:1. It was found that the reaction rate could be controlled by either varying the flow rate of water through the Cu cooling coils or, by keeping the flow rate constant, and automatically turning the cooling on/off upon reaching a set temperature. Measurements showed that the temperature of the reacting solution was inhomogeneous and was hotter at the air/reagent interface. Consequently, placement of the monitoring/controlling thermocouple influences the control of the reaction. Experiments done under pressure showed that pressure suppresses foaming. It was shown that the reaction could be controlled at operational temperatures of 25 °C and lower.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: (619) 553-1603. E-mail: [email protected]; pboss@ san.rr.com (P.A.M.-B.) Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Cady, H. P. Beginning of the Helium Industry. Ind. Eng. Chem. 1938, 30, 845. (2) Stephenson, C. Zeppelins: German Airships 1900−40; Osprey Publishing: Oxford, 2012. (3) Banks, A. Helium. J. Chem. Educ. 1989, 66, 945. (4) Snyder, W. E.; Bottoms, R. R. Properties and Uses of Helium. Ind. Eng. Chem. 1930, 22, 1189. (5) Kim, T. NaBH4 (Sodium Borohydride) Hydrogen Generator with a Volume-Exchanged Fuel Tank for Small Unmanned Aerial Vehicles Powered by a PEM (Proton Exchange Membrane) Fuel Cell. Energy 2014, 69, 721. H

DOI: 10.1021/acs.iecr.5b01647 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Industrial & Engineering Chemistry Research

Article

(6) Lee, J.; Kim, T. Micro PEM Fuel Cell System with NaBH4 Hydrogen Generator. Sens. Actuators, A 2012, 177, 54. (7) Oh, T. H.; Kwon, S. Performance Evaluation of Hydrogen Generation System with Electroless-Deposited Co-P/Ni Foam Catalyst for NaBH4 Hydrolysis. Int. J. Hydrogen Energy 2013, 38, 6425. (8) Kim, J.; Kim, T. Compact PEM Fuel Cell System Using Chemical Hydride Hydrogen Source for Portable Power Generators. Energy Procedia 2014, 61, 1992. (9) Pinto, P. J. R.; Sousa, T.; Fernandes, V. R.; Pinto, A. M. F. R.; Rangel, C. M. Simulation of a Stand-Alone Residential PEMFC Power System with Sodium Borohydride as Hydrogen Source. Int. J. Electric. Power and Energy Systems 2013, 49, 57. (10) Ferreira, M. J. F.; Coelho, F.; Rangel, C. M.; Pinto, A. M. F. R. Batch Sodium Borohydride Hydrolysis Systems: Effect of Sudden Valve Opening on Hydrogen Generation Rate. Int. J. Hydrogen Energy 2012, 37, 1947. (11) Brown, H. C.; Brown, C. A. New, Highly Active Metal Catalysts for the Hydrolysis of Borohydride. J. Am. Chem. Soc. 1962, 84, 1493. (12) Demirci, U. B.; Mielea, Ph. Cobalt-Based Catalysts for the Hydrolysis of NaBH4 and NH3NH3. Phys. Chem. Chem. Phys. 2014, 16, 6872. (13) Joydev, M.; Binayak, R.; Pratibha, S. Zeolite Supported Cobalt Catalysts for Sodium Boorohydride Hydrolysis. Appl. Mech. Mater. 2014, 490−491, 213. (14) Manna, J.; Roy, B.; Vashistha, M.; Sharm, P. Effect of Co+2/ BH4¯ Ratio in the Synthesis of Co-B Catalysts on Sodium Borhydride Hydrolysis. Int. J. Hydrogen Energy 2014, 39, 406. (15) Wu, C.; Wu, F.; Bai, Y.; Yi, B.; Zhang, H. Cobalt Boride Catalysts for Hydrogen Generation from Alkaline NaBH4 Solution. Mater. Lett. 2005, 59, 1748. (16) Gómez-Lahoz, C.; García-Herruza, F.; Rodríguez-Maroto, J. M.; Rodríguez, J. J. Cobalt (II) Removal f rom Water by Chemical Reduction with Sodium Borohydride. Water Res. 1993, 27, 985. (17) Patel, N.; Fernandes, R.; Miotello, A. Hydrogen Generation by Hydrolysis on NaBH4 with Efficient Co-P-B Catalyst: A Kinetic Study. J. Power Sources 2009, 188, 411. (18) Wolfe, R. L.; Ward, N. R.; Olson, B. H. Inorganic Chloramines as Drinking Water Disinfectants: a Review. J. Am. Water Works Assoc. 1984, 76, 74. (19) Ogawa, H.; Tanoue, E. Dissolved Organic Matter in Oceanic Waters. J. Oceanogr. 2003, 59, 129. (20) Steele, J. H.; Thorpe, S. A.; Turekian, K. K. Elements of Physical Oceanography; Academic Press: New York, 2009. (21) Höfer, R.; Jost, F.; Schwuger, M. J.; Scharf, R.; Geke, J.; Kresse, J.; Lingmann, H.; Veitenhansl, R.; Erwied, W. Foams and Foam Control. In Ullmann’s Encyclopedia of Industrial Chemistry; WileyVCH: Weinheim, 2000. (22) Joseph, D. D. Understanding Foams and Foaming. J. Fluids Eng. 1997, 119, 497.

I

DOI: 10.1021/acs.iecr.5b01647 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX