IdeaFest: Interdisciplinary Journal of Creative Works and Research from Cal Poly Humboldt
Abstract
Iron is an essential micronutrient required for the growth of phytoplankton, the base of the oceanic food chain. Iron exists in the ocean in extremely low concentrations, limiting primary productivity in ~⅓ of the surface ocean, and can be affected by physical processes such as upwelling of water masses, and by biological and chemical processes within the water. In order to further our understanding of iron availability in the ocean, here we refine and improve our understanding of an analytical method to measure nanomolar iron in seawater. The method used is an adaptation of Lohan et al. (2006), a flow injection analysis method with catalytic spectrophotometric detection. Here, we quantify the blank and its precision; we investigate the effect of different concentrations, pHs, and reagent qualities of the ammonium acetate buffers on the blank and baseline signal; and we investigate the limitation imposed by the data acquisition system (DAQ), which converts analog signal into digital readings. We found that our blank has an average concentration of 0.53 ± 0.07 nM; that the pH of the rinse solution does not have a significant impact on the peak shape, but that its concentration does; that reagent quality of the buffer and rinse solutions does significantly affect our blank and baseline; and that the two DAQ models tested have a lower limit on their detection range of ~0.03 nM.
Recommended Citation
Stafford, Brooke I.; Hodgson, Liam W.; Schuhmann, Noah S.; Vestal, Elijah J.; and Till, Claire P.
()
"Refining a Flow Injection Analysis Method for Trace Concentrations of Iron in Seawater,"
IdeaFest: Interdisciplinary Journal of Creative Works and Research from Cal Poly Humboldt: Vol. 9, Article 5.
Available at:
https://digitalcommons.humboldt.edu/ideafest/vol9/iss1/5
The method setup. The flow bench (left) is a trace metal clean space that contains the sample and reagent bottles. Plastic tubes containing the sample and reagents emerge from the back of the flow bench and reach a pump (red), which pushes the solutions through mixing coils, valves and chelating resin columns until the reaction stream reaches the spectrophotometer (orange). The DAQ (green) takes the analog signal from the back of the spectrophotometer and converts it into a digital signal. The valve controllers (blue) set the valves (yellow), which determine whether the system is rinsing, loading, or eluting. Ultimately, everything ends up in the waste (pink).
Figure 2.png (31 kB)
Schematic representing the preconcentration flow injection analysis setup for a) loading and b) eluting a sample.
Figure 3.png (33 kB)
Stacked absorbances of blanks ran with HPLC, Optima and quartz-distilled grade buffer and rinse solutions. Quartz-distilled reagents showed significantly lower blank absorbances with these initial tests.
Table 1.png (76 kB)
Blank concentrations and standard deviations for different load times of MilliQ. Load time does not appear to affect the blank measurement. We estimate our detection limit to be three times the standard deviation of the overall blank: 0.21 nM.
Table 2.png (14 kB)
Blank concentrations with varying buffer and rinse solutions. Note that the quartz distilled (Q) solutions result in a substantially lower blank than the HPLC and Optima solutions. As the Optima solution was undoubtedly trace metal clean when purchased, we suspect that the 3 nM blank with the Optima and HPLC solutions reflects something in our preparation of the solutions, rather than actual iron levels.
Figure 4.png (45 kB)
Zoomed in section of absorbance vs time for DAQ B showing the aforementioned data lines. Window in the top right depicts the portion of the peak. The difference between the red and green lines is the minimum absorbance change the DAQ can report. When converted to concentration using our standard addition curve, this corresponds to ~0.03 nM. This reflects limitations in our data sampling abilities
Figure 5.png (23 kB)
The effect of rinse solution pH on peak shape. Despite our expectations that the dip just before the peak would be affected by rinse solution pH, there did not appear to be any difference in the dip across the three pHs tested. We note that there was a small difference in the baseline between the three runs, however the rinse solution does not contribute to the reaction stream being measured by the spectrophotometer except in the moment just before the peak; further, the baseline routinely shifts throughout the day. Therefore, we do not expect that the baseline shift had to do with the rinse solution.
Figure 6.png (19 kB)
The effect of the rinse solution concentration on the peak shape. The seawater ran here gave an absorbance of about 0.7 regardless of pH, while under regular rinse conditions the same seawater on the same day gave an absorbance of around 0.5. This is a significant departure. Previously, we have observed similarly large peaks when the column is not rinsed for a sufficient time. So, we suspect the large peaks observed here with the low rinse concentration may also be due to insufficient rinsing of the column. Perhaps in both cases, some salts from the seawater remain on the column, are eluted into the reaction stream, and impact the reaction chemistry.