Counting bubbles… the saga continues

You might have read about the intriguing bubbly groundwater springs of southeast Belgium in a previous post in this very blog. Called by the local Walloon name ‘pouhons’, these springs have attracted economic and touristic interest for centuries – these waters have been bottled since the 1600s! The pouhons also attract a lot of scientific interest, because they store valuable information about geological and hydrological processes happening in the subsurface [1]. Our GeoEnergy research group at the Geological Survey of Belgium is particularly interested in studying the processes of generation of geogenic CO2 at depth, CO2 dissolution in groundwater and transport to the surface, so we can apply what we learn from natural examples to geologial CO2 storage sites. This has encouraged us to start developing a monitoring strategy targeting pouhons and streams of the area [2].

For almost one year, we have been carrying out a monthly survey at a ~300-metre-long stream that collects water from various CO2-rich sources, including the landmarked pouhon Delcor. Along the stream, we measure water pH, temperature, conductivity and flow, parameters that vary according to season and amount of rainfall. But measuring the CO2 dissolved in water is more complicated than you think. In the same way that the gas quickly escapes as soon as you open a bottle of a fizzy drink, naturally CO2-enriched groundwaters are degassing as soon as they reach the surface due to the change from higher underground pressure to the relatively lower atmospheric pressure. Besides that, the carbonic acid resulting from dissolution of CO2 in water dissociates into different species depending on the pH of the solution, making measurements more difficult. Although titration methods can be used to measure dissolved CO2, the oversaturated nature of the pouhon waters is hard to measure with precision. Just for comparison, these waters may have up to 4 g/L of dissolved CO2, while beer, which is industrially carbonated, has around 5 g/L dissolved CO2.

Monitoring work being carried out in southeast Belgium using a temperature-pH-conductivity meter (left) and titration (right).

With all that in mind, we spent a lot of time researching what is the best way to measure the amount of dissolved CO2 in water covering a wide concentration range, from pure stream waters to pouhon waters. We came across the work by Dr. Dorothy Vesper from West Virginia University and collaborators [3,4]. They successfully demonstrated that the use of a commercial carbonation meter, the CarboQC by Anton Paar, is suitable for measuring dissolved CO2 in natural waters. The CarboQC was ordered and first testing occurred in November 2019.

We were hoping that the job of counting CO2 bubbles in water would be simple once we had the equipment, but we can tell you it was not like this. Ideally operating with 100 ml of sample, our first challenge was to find large syringes to inject the sample in the measuring chamber. And remember that we mentioned earlier that solubility and speciation of CO2 varies with pH? This means that, besides the 100 ml of sample, we need to account for some 5-10 ml of a buffer solution that pushes the equilibrium towards complete H2CO3 speciation. Moreover, the correct injection pressure and pace is key for a good measurement, sufficient rinsing ensures there is no cross-contamination between samples, and CO2 tends to exsolve from an acidified and oversaturated solution. All of these variables needed to be optimised and checked for the intended field use, resulting in some adaptations to the previously developed and published method . This involved a lot of tests in the laboratory as part of the MSc internship of Diana Snyder from Ghent University .

After gaining confidence operating the equipment and with a well-developed methodology of our own, it was the time to test the setup in the field early December 2019. And we are happy to report that the test was a success! Besides portability and fairly quick sample processing times, the CarboQC gave satisfactory results when compared to the yearly measurements performed by the Spadel company in three pouhons – Delcor, Sauvenière and Groesbeeck.

Successfully testing the CarboQC in the field. Left: Diana Snyder filling the sample chamber with a syringe. Right: Diana Snyder and Kris Welkenhuysen happy with resulting pH of the buffered sample prior to measurement.
Comparison of results obtained for dissolved CO2 of three pouhons [1]

A first full field survey was successfully conducted end of December 2019; our monthly survey will be expanded with these direct dissolved CO2 measurements to investigate shorter-term variations.

Renata Barros & Kris Welkenhuysen
GSB – Geological Survey of Belgium


[1] Barros, R., Defourny, A., Collignon, A., Jobe, P., Dassargues, A., Piessens, K. & Welkenhuysen, K., submitted. A review of the geology and origin of CO2 in mineral water springs in southeast Belgium.

[2] Welkenhuysen, K., Snyder, D., Piessens, K. & Barros, R., 2020. Long-term monitoring of natural CO2-enriched sources in Spa, Belgium: measuring dissolved inorganic carbon. Abstract submitted to the 15th International Conference on Greenhouse Gas Control Technologies GHGT-15.

[3] Vesper, D. & Edenborn H., 2012. Determination of free CO2 in emergent groundwaters using a commercial beverage carbonation meter. Journal of Hydrology 438-439, 148-155.

[4] Vesper, D., Edenborn, H., Billings, A. & Moore, J., 2015. A field-based method for determination is dissolved inorganic carbon in water based on CO2 and carbonate equilibria. Water, Air, and Soil Pollutants 226:28, 12 p.

Note: This blog is optimized for viewing in Chrome or Firefox. Click on images to display in full resolution.