A Simple, Low-cost, and Robust System to Measure the Volume of Hydrogen Evolved by Chemical Reactions with Aqueous Solutions

There is a growing research interest in the development of portable systems which can deliver hydrogen on-demand to proton exchange membrane (PEM) hydrogen fuel cells. Researchers seeking to develop such systems require a method of measuring the generated hydrogen. Herein, we describe a simple, low-cost, and robust method to measure the hydrogen generated from the reaction of solids with aqueous solutions. The reactions are conducted in a conventional one-necked round-bottomed flask placed in a temperature controlled water bath. The hydrogen generated from the reaction in the flask is channeled through tubing into a water-filled inverted measuring cylinder. The water displaced from the measuring cylinder by the incoming gas is diverted into a beaker on a balance. The balance is connected to a computer, and the change in the mass reading of the balance over time is recorded using data collection and spreadsheet software programs. The data can then be approximately corrected for water vapor using the method described herein, and parameters such as the total hydrogen yield, the hydrogen generation rate, and the induction period can also be deduced. The size of the measuring cylinder and the resolution of the balance can be changed to adapt the setup to different hydrogen volumes and flow rates.


Introduction
Due to their high energy density, lithium-ion batteries are currently one of the most popular power sources for portable consumer electronics. However, the amount of energy that can be delivered by a battery is limited. There is thus currently much interest in developing alternative methods of providing portable power. One of the more promising methods is the use of proton exchange membrane (PEM) fuel cells, which generate electricity and water by combining hydrogen and oxygen. PEM fuel cells have two main advantages over batteries. Firstly, PEM fuel cells can provide power for a much longer period of time (as long as a flow of hydrogen is maintained). Secondly, depending on the fuel source, PEM fuel cells can have a much greater energy density than batteries, meaning that a smaller system can provide more energy. 1,2 As a result of this, there is a currently a large amount of research directed at developing portable, on-demand hydrogen sources. [2][3][4][5][6][7] One method which is currently receiving much attention is the generation of hydrogen by reacting chemicals with water. 8,9 One of the most important parameters which must be measured in these reactions is the evolution of hydrogen. For simple reactions, such as the evolution of hydrogen by the addition of chemical hydrogen storage materials to aqueous solutions, it is advantageous to have a simple, low cost measurement system. An example of such a system is the water displacement method, in which the volume of gas generated in a chemical reaction is measured simply by tracking the volume of water displaced from an inverted water-filled measuring cylinder. This technique originated in the pneumatic trough, which was developed by the botanist Stephen Hales and then adapted and put to its most famous use by Joseph Priestley to isolate several gases, including oxygen, in the 18 th century. 10,11 The water displacement method is applicable to any gas which is not particularly soluble in water, including hydrogen, and is still widely used to record the volume of hydrogen generated from the reactions of various chemicals, such as sodium borohydride, aluminum, and ferrosilicon, with water.
4. Before the equilibration period ends, open a new spreadsheet in the spreadsheet software package and then open the data collection software. Load the method created in Step 1 by going to 'File' on the data collection software start menu, and then 'Open method'. 5. Just before the 10 min equilibration period is due to end, go to 'Activate' and then click on 'Normal mode'. Data will start being logged in the spreadsheet software package. 6. At the end of the 10 min equilibration period, add the silicon by rapidly inverting the glass vial and depositing the silicon into the sodium hydroxide solution. 7. Rapidly place the ground glass joint of the adapter which is attached to the tubing into the neck of the round bottomed flask. Zero the balance. The moment at which the balance is zeroed will be taken as time (t) = 0 in the data analysis. 8. After 10 min have elapsed, stop the data logging by pressing the backspace key and then selecting the 'Quit' option on the data collection software menu. Save the file in the spreadsheet software package. 9. Remove the adapter which is attached to the tubing from the round-bottomed flask and add water to quench the reaction. 10. Isolate the solid residue in the flask for further analysis by centrifugation or gravity filtration, or transfer the entire reaction mixture to a beaker and neutralize with hydrochloric acid (1 M) and dispose of the waste appropriately.

Data Analysis
1. Ensure that the data is loaded into an appropriate spreadsheet software package. 2. Find the point at which the balance is zeroed; this is considered to be the (t) = 0 point of the reaction. 3. Delete the data which precedes this. 4. Insert a column to the left of this data. This will contain the time. 5. Add appropriate time intervals, starting from zero, to the column which has just been inserted. The balance used in these studies logged 8.5 data points per second, and thus time intervals of 0.117647 (=1/8.5) sec were used. 6. Consider gas which has been collected over water to be saturated with water vapor. During the collection process, the water level in the measuring cylinder adjusts to maintain the internal pressure in the measuring cylinder at atmospheric pressure. 7. Apply an approximate correction factor using Dalton's Law, which states that the sum of the individual partial pressures of the gases in a mixture (P 1 …P n ) is equal to the total pressure (P tot ). As, if the room temperature is 298 K, the partial pressure of water vapor is 31,69.9 Pa, and the total pressure of gas in the measuring cylinder is atmospheric pressure (101,325 Pa), it can be calculated that there is approximately 3.08% water vapor by volume in the collected gas. Estimate the amount of water vapor in the hydrogen at other temperatures by using the partial pressure of water vapor at the temperature in question. 8. To obtain an estimate of the amount of hydrogen generated (if the room temperature is 298 K), multiply the gas volume by 0.97. 9. Estimate the initial hydrogen generation rate by fitting a linear trend line to the initial steep slope of the hydrogen generation curve. 10. Take the induction period as the time taken for water to be displaced from the measuring cylinder. These estimates of induction period are not absolute; the actual hydrogen generation reaction starts before the end of the 'induction period' estimated in these experiments as a certain amount of hydrogen must be generated to be able to begin displacing water. However, these values do allow for an assessment of the relative change in induction period between experiments.

Representative Results
To investigate the reproducibility of the experimental set-up, varying masses of silicon were reacted with aqueous sodium hydroxide solutions to generate hydrogen. Each reaction was performed in triplicate. The average hydrogen generation curves are shown in Figure 1. Average total hydrogen yields, hydrogen generation rates, and induction periods for each mass of silicon were also calculated and are plotted with error bars representing one standard deviation in Figures 2, 3, and 4, respectively. There was very little deviation in the total hydrogen yields and hydrogen generation rates between reactions, and a greater level of deviation in the induction periods.   shows some representative results from a sub-optimal experiment. In this case, the low flow of hydrogen between 200 and 800 sec results in the build-up of drips due to the surface tension of the water, which fell at approximately 400 and 710 sec. Though these drips do not affect the calculation of the maximum hydrogen generation rate, they could have an effect on the total hydrogen yield if, for example, the measurement was stopped before the drip fell. It is thus necessary to either alter the reaction conditions (in this case, for example, by adding a greater mass of aluminum-silicon alloy or using a higher concentration of sodium hydroxide) to ensure a higher flow of gas or the reaction setup to prevent the buildup of drips.

Discussion
The most critical steps of the protocol are those which occur at the beginning of an experiment. The large temperature dependence of the rate of these hydrolysis reactions means that great care must be taken to ensure that the solution temperature has reached equilibrium before the addition of the solid. The solid must be added rapidly and completely, the ground glass joint of the adapter must be properly inserted into the neck of the round-bottomed flask, and the balance must then be zeroed as rapidly as possible. An incorrect measurement of start time and reaction temperature will generate incorrect results.
The method does have some limitations. It is imperative that the beaker into which the measuring cylinder is inserted is as narrow as practicable to ensure that the water displaced from the measuring cylinder is rapidly channeled down the plastic bridge onto the balance. Otherwise, the surface tension of the water allows for a slow buildup of the water level at low flow rates (see Figure 5) until the point at which all of the water is released in a large drip.
The error of the balance also limits the resolution of the data. In these experiments, a balance with an error of ± 0.05 g was used, which is adequate when generating several hundred milliliters of hydrogen, but a balance with a smaller error would be required if smaller volumes were being measured.
By automatically logging the data in a spreadsheet, this method offers a significant improvement in accuracy and temporal resolution with respect to water displacement methods which rely on recording the volume of gas evolved manually. However, though it is considerably lower in cost than methods which use cameras and image analysis software to track gas evolution, it is generally lower in temporal resolution, and such camera-based methods also avoid the problem of oscillating mass-balance readings due to water forming drops and therefore produce data which can be more easily processed by differentiation.
The water displacement method is applicable to the collection of any gas that has low solubility in water. Thus, this experimental protocol could be modified for the measurement of rates of gas generation from other chemical reactions which evolve poorly water soluble gases.

Disclosures
The authors have nothing to disclose.