[OLD] Creation of nitrate from nowhere

[webmaster]

From: Larry Hull

Subject: Creation of Nitrate from nowhere

I am attempting to simulate the leaching of gypsum and nitrate from soil during infiltration. I have suction lysimeters at two levels in the soil, and have input and output concentrations. I am treating the soil as a well mixed batch reactor to simulate a LiBr tracer test. Admittedly not the best approach, but a simple first approximation. There is a net increase in nitrate in the water as it infiltrates. I am attempting to simulate this as a kinetic dissolution reaction of Ca(NO3)2. However, even if I remove the mineral Ca(NO3)2 from the system, I get an increase in nitrate. I don't know where the nitrate is coming from. The input (ie the reaction moles of NO3-) are also lower than the output NO3-. The script and the databases needed to run the script are attached. Note- thermo.dat has been modified by adding Ca(NO3)2 as a mineral.

 

From: Craig Bethke

Subject: Re: Creation of Nitrate from nowhere

That's an interesting question with an answer that may surprise you. You set up a flush model that includes a number of sorbing surfaces. Nitrate accumulates in the system because as fluid passes through it, nitrate becomes associated with charged surfaces. Although nitrate doesn't specifically sorb on any of the surfaces, it accumulates in the ionic diffuse layer that counterbalances surface charge. In this case, net surface charge is positive, and anions -- nitrate in this case -- accumulate in the diffuse layer. With time, therefore, the amount of nitrate in the system increases, even though the concentration in the input fluid is less than the concentration in the initial fluid. Is this result realistic? That's hard to say. Construction of the diffuse layer is a problem that has long vexed reactive transport modelers: it must be included in the model in order to maintain charge balance as fluid passes through the system, but surface complexation theory gives no guidance as to the layer's specific composition. Given this ambiguity, all reactive transport (and reaction models in a flush configuration) must make up a diffuse layer in a somewhat ad hoc manner. React does this simply by using abundant ions of charge opposite that of the surface.