Comment

1 October 2005

Hydrazine scene: It's a gas

A better methodology for calibrating hydrazine monitors.

By Kofi Korsah, William Murray, and Bruce Tomkins

The MDA 7100 is one of the detectors the U.S. Air Force uses for monitoring hydrazine. It determines the concentration of a test gas by exposing a chemically treated paper tape (Chemcassette) to a sample of the test gas. The Chemcassette reacts with the test gas and changes color in proportion to the concentration of the gas. The detector's built-in firmware has concentration versus optical density response curves used to calculate, record, and display the gas concentration.

Three methods exist to perform field calibrations of the MDA 7100: optics calibration, dynamic calibration, and manual calibration. Optics calibration returns the monitor's ability to interpret differences in stain intensity to the same level it was on the day the manufacturer calibrated the unit. It occurs during installation of a fresh Chemcassette detection tape. An optics calibration doesn't assume anything about the MDA 7100 (whether or not the system is in working condition). Dynamic calibration uses a test gas of known concentration. Unlike the optics calibration method, the Chemcassette must be in place when performing a dynamic gas calibration. Manual calibration of the hydrazine monitor is a dynamic calibration you perform when you don't know the concentration of the test gas.

Calibration method

We base our calibration method on generating hydrazine gas from a permeation tube and mixing it with humidified gas to provide a reference hydrazine concentration at the humidity we want. It employs a KinTek model 491MB standard gas generator, which houses the permeation tube, connected to a KinTek 491 HG humidifier and a KinTek 491 IM gas-sampling interface.

Flow diagram of the hydrazine gas generation/calibration system.

Typically, the permeation tube goes in an oven in the gas generator. The hydrazine permeation rate is a function only of the temperature of the oven. Then the nitrogen carrier gas source connects to the gas generator. Thus, we can obtain different gas concentrations from the same permeation tube by adjusting the oven temperature and the flow rate of the carrier gas. We immediately pressure regulate the carrier gas entering the gas generator to 50 psig. Part of this dry nitrogen carrier gas flows directly through the humidifier, generating humidified gas to the mixer. The other part becomes the carrier for the hydrazine permeating from the tube.

The dry hydrazine gas mixes with the humidified gas to generate the desired reference concentration. Next, connect an interface module containing a large Teflon cylindrical volume to the output of the humidifier module. The interface module is at atmospheric pressure and serves as a mixing and sampling chamber. It also ensures the detector under calibration will be sampling under pressure conditions for typical use. To prevent outside from getting into the mixing chamber and invalidating the results, make sure the flow rate of the input carrier gas is greater than the combined flow rate of any sampling pumps and monitors connected to the interface module.

Verify gas concentration

In most cases, calibration of the KinTek reference gas generator ensures generation of a known gas concentration for hydrazine detector calibrations. We perform this using coulometric titration of impinger samples to obtain a calibration curve that relates calculated gas concentration Cc (based on permeation oven temperature and carrier gas flow rate) to actual gas concentration Ca as calculated from the coulometric titration. The difference between Cc and Ca is due to adsorption of hydrazine on tubing surfaces.

In theory, this calibration curve should remain unchanged if the system setup and configuration does not change. Thus, you don't need to calibrate the reference gas generator every time you need to calibrate a 7100 hydrazine detector. In practice, develop control charts to determine a suitable time interval before obtaining a new calibration curve.

One advantage of using coulometric titration for verifying the concentration of the reference gas is it's a primary standard (for simple solutions), thereby guaranteeing, in principle, measurements will be traceable to SI units.

The coulometric titration system includes a reaction vessel, usually a 50 mL disposable plastic beaker, containing 30 mL of 0.1M sulfuric acid to which we've added 0.4-0.6 g of potassium bromide. The vessel also contains a bromine generator consisting of a glass tube (1 cm diameter) with a glass-fritted bottom (coarse porosity) to which electrolysis electrodes mount parallel to each side of the frit. Typically, the fritted glass tube is filled with 0.1 M sulfuric acid. Because we always kept the level of acid in the tube higher than that of the solution in the electrolysis cell, the latter could not diffuse into the bromine generator cell. The sensing electrode is a dual platinum foil electrode. We stirred the solution in the electrolysis cell briskly and consistently using a magnetic stir bar during all determinations. (See the schematic of the coulometric titration system for the reaction vessel and electrode assembly.)

The bromine generator and sensing electrodes connect to a constant current source and a conductivity measuring circuit. The constant current source is capable of generating a current in the range of 0-1 mA. However, in the calibration under consideration, we adjusted the current such that the coulometer applied exactly 0.1 mA to the bromine generator. The sensing circuit detects the point at which all the hydrazine in the solution is consumed and free bromine begins to build up in the solution. The output voltage of the sensing circuit, whose input connects to the sensing electrodes, is proportional to the increased resistance and voltage in the solution the accumulation of unreacted bromine causes.

The coulometric measurement method titrates hydrazine with electrochemically-generated bromine. A great advantage of the method is, theoretically, you can eliminate the need to prepare the usual calibration curve, involving multiple standards, for quantitating the analyte. Another significant advantage is as-needed electrochemical generation and immediate use of the bromine required from potassium bromide. There is no need to maintain a supply of chemically unstable and corrosive bromine. Since you can calculate the amount of produced bromine directly from the electrolysis time and current, you can use the following equation to calculate the amount m, in grams of hydrazine in the solution:

m = A*M*(t_s - t_b)/(n*F) (1)

Where:

• M is the molecular weight of hydrazine, here 32.05 grams/mole.
• A is the current applied to the bromine generator cell in amperes (coulombs/sec), here 0.1 mA.
• t_s and t_b are the times, in seconds, to the break point of the sample (solution containing the hydrazine) and blank (solution containing no hydrazine) respectively. (See the coulometric titration endpoint graphic.)
• n is the number of moles of bromine required to react with (oxidize) one mole of hydrazine, here 4.
• F is Faraday's constant, 96,500 coulombs/mole of electrons.

Daily preparation of normal calibration curves tested the hypothesis the theoretical values we calculated from the above equation were sufficient for quantitating hydrazine with acceptable accuracy. We typically analyzed five standards, ranging in mass between 0.1 and 3 µg hydrazine and a blank in quadruplicate prior to analyzing unknown samples. We plotted the calculated masses (found values) against the theoretical masses (true values) to determine the method bias. In all cases, we observed a linear bias curve with r² > 0.995. If the equation (1) were a perfect indicator of method performance, the slope of the calibration curve would be equal to 1.00. (See data for nine independent calibration curves in the "Evaluation of the bias and detection limit for hydrazine using coulometric titration" table on page 50.)

We calculated detection limits for hydrazine using a method approved by the U. S. Army Rocky Mountain Arsenal. This procedure uses not only the found and true values of hydrazine, but it accounts for the possibility of false positive and false negative errors (typically set equal to 5%).

The data in the table on page 50 show there is a clear high bias to the calibration curve, and the theoretical equation (1) is not sufficient for providing proper calibration data.

Schematic of the coulometric titration system. Displays the reaction cell and coulometer. A bromine generating electrode and sensing electrode are approximately the same size, but the bromine electrode is exaggerated to show the design.

There are at least three reasons for the apparent high bias. First, the bias is probably due to a delay time, which can be partly described by the time of diffusion of bromine from the generating electrode to the indicator electrode. You can partially, but not entirely, eliminate this delay time by stirring the solution briskly and placing the generating and sensor electrodes in close physical proximity to each other. The total amount of charge needed to reach the equivalence point will be larger than the theoretical value calculated using equation (1).

Second, the reaction kinetics between bromine and hydrazine are very rapid, but not infinitely fast. You may need a finite amount of time before the micromolar concentrations of analyte and reagent find each other and react. Third, a small quantity of bromine may be escaping into the air. The table indicates the greatest biases occurred during the warm summer months and the smallest biases during the early spring. The laboratory temperature during these months ranged between 22°C and 27°C. Although this is not a huge variation, comparatively, small changes in room temperature may influence the potential losses of microgram quantities of bromine to the headspace of the reaction vessel.

Uncertain calibration system

In studies evaluating the calibration of hydrazine detectors using the KinTek 491MB reference gas generation system, one set of experiments evaluated the main sources of uncertainty in the calibration method and possible methods to both decrease the calibration uncertainty and the amount of time required to perform calibrations. Results revealed the impinger sampling system was the best place to make improvements. Hydrazine is so reactive, relatively short exposures of tubing and connections to the air during manual impinger sampling is sufficient to make it necessary to recondition the previously conditioned tubing in order to prevent significant reductions in the observed hydrazine concentrations. After investigation, we determined using ball-and-socket joints would permit fast enough connection and disconnection of the tubing, impingers, and sampling pumps to substantially improve the repeatability and lower the uncertainty of the sampling procedure. This also decreases the amount of time required per impinger sample.

Relative humidity

A comparison of coulometric titration of the impinger samples to readings from the MDA 7100 monitor connected to the same test atmosphere showed a strong dependence of the MDA 7100 readings on the relative humidity (RH) of the reference gas used to generate the test atmosphere, and the RH of the environment (7100's Chemcassette). We plotted a correction factor (the ratio of the hydrazine concentration measured by coulometric titration to the concentration measured by the 7100) versus the difference in the relative humidity between the reference gas and the MDA 7100 Chemcassette's environment.

Permeation tube calibration

A comparison of KinTek-491 and readings on the 491 display panel of the oven temperatures indicated a difference of about 2°C in the operating temperature inside the KinTek 491's permeation chamber and the setpoint temperature of the permeation chamber oven. KinTek calibrates permeation tubes by using a weighing method based on putting the tubes in a permeation chamber oven of a different reference gas generating system called the Span Pac. Since the permeation rate of a permeation tube changes exponentially with temperature, even a small difference between the operating temperature inside the ORNL 491 system permeation chamber and the KinTek Span Pac system would cause the actual operational permeation rate in the ORNL system to be significantly different from the rate KinTek determined in their calibration process. Therefore, we recommend using coulometric titration to determine the needed calibration factor each time you use a new permeation tube. Once you obtain this factor for a particular permeation tube, you can measure it periodically using coulometric titration. To determine the time interval required before recalibration of a particular tube with coulometric titration, perform periodic recalibrations of a check permeation tube over a period of several months.

Furthermore, test results show it's possible to use the coulometric titration method to determine the permeation rate of permeation tubes at temperatures other than the temperature at which you calibrated them.

Behind the Byline

Kofi Korsah, William Murray, and Bruce Tomkins are with Oak Ridge National Laboratory in Oak Ridge, Tenn.