Determination of the Water Content of an Ionic Liquid

The moisture content is an important quality criterion of ionic liquids. In this article, we show how the water content of 1-ethyl-3-methylimidazolinum methyl sulfate can be determined using a thermogravimetric analyzer (TGA) interfaced to a mass spectrometer. The result was confirmed by titration measurements.

Introduction

Ionic liquids are salts that are liquid at room temperature. They exhibit many properties that make them extremely attractive for a number of applications: they are thermally stable, non-flammable, non-volatile, conduct electricity, have a relatively high heat capacity, and have very good dissolving properties for many substances.

It is therefore not surprising that they are used in numerous applications, for example as electrolytes in fuel cells or batteries; thermofluids; solvents, etc. An important quality parameter of an ionic liquid is its water content. In this article, we show how the water content of 1-ethyl-3-methylimidazolinum methyl sulfate (EMIMS, CAS 516474-01-4) can be determined using a thermogravimetric analyzer coupled to a mass spectrometer. The water content determined using this method was confirmed by titration measurements.

Experimental Details

The measurements were carried out using a TGA/DSC1 coupled to a ThermoStar mass spectrometer. The sample was measured in a 100 uL aluminum crucible from room temperature to 350 °C at a heating rate of 10 K/min. Nitrogen was used as purge gas at a flow rate of 50 mL/ min.

The titration measurements were performed using a METTLER TOLEDO V30 Compact Volumetric KF Titrator equipped with a DM143 electrode. For the analysis, about 2 g EMIMS was added to 30 mL HYDRANAL methanol. The water content was titrated with HYDRANALComposite 5.

Results

Figure 1 displays the TGA and first derivative curve (DTG curve) of EMIMS. From 50 °C onward, the sample loses mass in two steps. The first step between room temperature and about 270 °C could be due to the loss of moisture. Evaluation of the step height yields a value of 1.3%.

However, this step is overlapped by the second mass loss step, which begins at about 270 °C. This is evident because the DTG curve does not return to zero mg/K between the steps (see the dotted black line). The water content determined by the step evaluation of the TGA curve does not therefore correspond to the true water content.

If we assume that the peak on the DTG curve describes the loss of water, the water content can be determined by integrating the peak area as shown in Figure 1. This yields a mass loss of 0.21 mg which corresponds to a water content of 0.54%. This value is however strongly dependent on the type of baseline used to evaluate the DTG peak. Consequently, the water content can only be determined with an accuracy of about ±0.2%. 

Much more accurate information on the water content of the sample can be obtained by using a TGA coupled to a mass spectrometer (MS).In this method, the MS simultaneously measures the compounds evolved during the TGA measurement.

The results are summarized in Figure 2. They show that several other fragments and molecular ions besides water (m/z 18) are simultaneously produced in the first mass loss step, for example methyl (m/z 15), CO2 (m/z 44), SO and SO2 (m/z 48 and m/z 64).

This indicates that EMIMS is chemically no longer stable above about 180 °C. The amount of water released by the sample is proportional to the peak area of the MS curve for m/z 18. The proportionality factor is determined in a calibration measurement. To do this calcium oxalate monohydrate (COMH) was measured by TGA-MS under the same conditions (heating rate, gas flow) as the EMIMS sample. The result of this measurement is displayed in Figure 3.

It is well-known that COMH eliminates water of crystallization from about 100 °C onward. The height of the step on the TGA curve between 100 and 240 °C therefore corresponds to the amount of water released by the COMH sample. In the example, it is 0.189 mg.

During the release of the water of crystallization, the MS records an increased ion current for m/z 18 (water). The area of this peak, ∆Q, is directly proportional to the amount of water vapor released (in the example ∆Q = 591 nA s). The proportionality factor k = ∆m / ∆Q is therefore 0.32 μg/nA s.

Figure  4 displays the TGA and DTG curves of EMIMS again together with the MS ion curve for m/z 18. If the hatched peak area of the MS curve (322 nA s) is multiplied by the proportionality factor determined above, a water loss of 0.1 mg is obtained which corresponds to a moisture content of 0.27%. This of course assumes that the MS ion curve for m/z 18 in this temperature range originates solely from water.

Volumetric titration yielded a value of 0.25% for the water content of the EMIMS sample. Within the limits of measurement uncertainty, this value agrees well with the value of the water content determined from the TGA-MS measurements. 

The results of the different types of determination for the water content of EMIMS are summarized in Table 1.

Conclusions

The water content of materials that are thermally stable can be easily determined by TGA. In thermally unstable materials, the elimination of water and decomposition may however occur in the same temperature range. The water content can then no longer be reliably determined by TGA. If a mass spectrometer is available, the water content can be determined from the MS ion intensity curve. A suitable reference sample is needed (in this case calcium oxalate monohydrate) to determine the sensitivity of the mass spectrometer toward water (here for m/z 18). The method allows the determination of low contents of water from the MS ion curve. The procedure described here for the quantitative evaluation of MS ion curves for particular m/z values can in principle be used for the detection of any substances.

_{Determination of the Water Content of an Ionic Liquid | Thermal Analysis Application No. UC 407 | Application published in METTLER TOLEDO Thermal Analysis UserCom 40}