To limit the effect of ambient humidity on DMS, a gas-permeable membrane can be used to limit the water content of the sample while letting the compounds of interest through. In this post, we describe how humidity can effect the DMS response, and the process of adding a PDMS membrane to the sampling inlet of IonVision. The effectiveness of the membrane was tested with formic acid in different humidity conditions.
Humidity in DMS Measurements
Changes in humidity influence the sensitivity of DMS due to the increased proton affinity of water in high humidity conditions. Ambient humidity is often high indoors and in outdoor settings, and changes between days can be dramatic. This can cause problems for DMS, as the detection limit shifts, and the same concentration gives a different response depending on the humidity. This is especially troublesome in continuous monitoring situations, as the DMS response may shift during a time that no-one is actively following the process.
PMDS Membrane with IonVision
The PDMS membrane setup is shown in the image below. The membrane was squished between two aluminum blocks. Dimensions of the membrane were 5x8x0.25mm. The wet sample was cycled on one side of the membrane using the sample pump of IonVision. The sensor pump was used to cycle air trough the clean side of the membrane and to the IonVision sensor. As such, the only addition needed was the membrane block, otherwise the existing tubes and pumps in IonVision were used.
The system was first tested by using water in a headspace chamber as the sample. The dewpoint of the sample was 2°C, and the dewpoint on the other side of the membrane was -47°C. While using an empty headspace chamber, the corresponding dewpoints were 17°C and -47°C. Thus, the additional water in the headspace chamber did not increase the humidity of the sample going to the sensor from ambient conditions.
Measurement of Formic Acid in High Humidity
Formic acid was selected as the test chemical due to a customer interest. Formic acid has also been previously sampled with DMS. The response is quite characteristic and depends heavily on the sample concentration. A solution with 0.05ml formic acid in 1l of water was prepared for the test.
The typical response for formic acid was able to be seen in a spectra taken immediately after connecting the headspace chamber. The solution was sampled for six minutes. The response intensified between the first and second measurements, but no clear difference was seen between the second and third measurements, started 2 and 4 minutes after the connection, respectively.
The headspace chamber was disconnected from the setup, and four measurements were taken each 2 minutes. The measurement taken at 6 minutes after the disconnection was not clearly distinguishable from the reference taken before the formic acid sampling.
To conclude, the formic acid was immediately visible through the membrane, and the response stabilized after two minutes. The system cleared of sample remains six minutes after sample removal. Below are the spectra from before the measurement, after two minutes on sampling, and after six minutes of clearing.
The response through the membrane was compared to response straight from the headspace. The humidity of the headspace sample at the sensor was approximately -4°C. The response was more intense, so the membrane also diluted the sample going to the sensor. System clearing after sample removal took longer than with the membrane, and at six minutes the formic acid response was still visible.
PDMS membrane is effective in removing humidity from the sample. This comes at the cost of some signal intensity in low humidity conditions. Larger surface area of the membrane allows for more gas permeation. However, larger surface area means more molecules getting stuck and releasing over time, slowing down the system dynamics. This can be compensated by heating the membrane.
Have you been suffering from the scourge of humidity? Let us know with the below contact form and we can start working together on curing it with membrane sampling!