Progress 09/01/23 to 08/31/24

Outputs
Target Audience:From a scientific perspective the anticipated audience of the Phase II efforts will be the soil and agronomic research community. We will reach this audience via publications, presentations, and collaborations. From a commercial perspective, the target audience is also the soil and agronomic research community. We will reach this audience through presentations and exhibition booths at scientific conferences. During this period, we presented results at scientific conferences, and had a booth at both the European and American Geopphysical Union meetings. Atthe booth we discussed all Aerodyne products, including this upcoming products resulting from this project. This was especially fruitful at the meeting in Europe (Vienna), where there is intense interest in nitrogen use in farmland. Changes/Problems:Because of leadtime for some items - pumps and selector valves, tubing materials, and lasers, our progress has been slowed somewhat. In addition, we want to have extensive opportunities for graduate students from our academic research partner(Prof. Homyak, co-PI, UC Riverside) to work with the instrument and get experience obtaining and analyzing high quality data that is of publication quality. To ensure this happens, we will likely be requesting a 1-year no-cost extension. What opportunities for training and professional development has the project provided?In addition to the techincal achievements, we employed a young undergraduate scientist (Hamilton College) for the summer 2024. She spent the summer learning how to use the TILDAS instruments, then performed much of the testing described in teh achievements. She finished her summer work by giving a presentation about the results, and we look forward to having her as a coauthor on the first technical paper from this project. How have the results been disseminated to communities of interest?We have presented the initial results of this work at the annual Soil Science Society of America (SSSA). In addiiton we have maintained an exhibitor booth at the American Geophysical Union (AGU), European Geophysical Union (EGU), and SSSA meetings, where we discuss our results, and potential collaborations with academic and governmental researchers. What do you plan to do during the next reporting period to accomplish the goals?Despite the great advances discussed here, there is still more testing to do before moving to laboratory demonstration in real soils. These next steps are discussed below. Multiplexing The results shown above do not employ the multiplexing valve, as the testing was performed with a single MD probe. A next step is to test the multiplexing aspect by immersing multiple probes first in the same solution to test consistency between probes, then in solutions of different concentrations to test the ability to recover a wide range of concentrations. Testing in non-liquid media We would like further test the ability of the probes to recover the analytes in more real-world media. We will test this by first placing the MD probes into inert silica sand, where we can control the concentration if solute, and also the water content. It is known that MD probes do not quantitatively recover solute concentrations in soils with water content below ~50% water holding capacity, where there is insufficient contact between the probe and soil water for efficient solute transfer across the membrane. We will measure the recovered solute concentrations as we slowly drain the soil and dry it out, to confirm this threshold in our setup. In addition to these tests, we will perform standard additions of nitrate and nitrite into the silica sand, to demonstrate that the MD probes can recover controlled changes in soil N concentrations. This will be performed across a range of concentrations. Real soil testing In the laboratory we will make real-world soil measurements, specifically looking at the soil N concentrations as a function of discrete N addition (ammonia and nitrate additions). This will be performed on locally-available soils. Deployment at UC Riverside After these tests we will deploy the instrument at the UC Riverside for tests of the system with real-world pyrocosm soils. For this deployment we will focus on nitrate and nitrite measurements using a compact-TILDAS rather than a full dual-laser TILDAS. We will follow up several months after this with a dual-laser system that will allow for automated and multiplexed NO3-, NO2-, and NH2OH measurements, which is the ultimate goal of the project.

Impacts
What was accomplished under these goals? During this reporting period we have made progress on tasks 1, 2, and 4, as described below. Although we are making good progress, we will likely look to extend the project another year to ensure optimal performance and maximize potential success of the laboratory demonstration in collaboration with UC Riverside. ?Task 1. Interfacing mL liquid sample to the TILDAS: further modifying the TILDAS absorption cell design and choosing materials to best operate with liquid samples. In Phase I of this project we manually injected the sample of interest into a heated (100 C) 76-m multipass cell through a septum installed in the center of the cell. We found that although this successfully deposited the sample, there were occasions where a droplet would ballistically impact one on the absorption cell mirrors, reducing instrument performance. In this Phase II task we worked to mitigate this problem. Injection Design We created a modified injection design, where the injection occurs in a side port of the cell, which can be heated to a different temperature than the absorption cell path. Note that this is a part of the absorption cell, which is at vacuum during injection. Further, in Phase I we found that there was significant variability in the time signature of the HNO3 and HONO between injections, likely due to inconsistency in the sample droplet fragmentation in hot vacuum. In Phase II we have modified this such that the sample is not "squirted" into open vacuum as it was in Phase I. It is injected directly onto a hot surface, in this case a polished stainless steel head, to encourage consistency in the injection method. The optimization of this approach will be discussed in the Tasks 2 and 4 sections below. Materials Choosing materials for liquid injections is crucial to ensure that the sample meticulously extracted from an MD probe is quantitatively delivered to the analyzer. Although this aspect of the task is ongoing, we have tested polyetheretherketone (PEEK) tubing and found it to be suitable. This is advantageous because it is strong, rigid, inert, and importantly is easily available from multiple supplier with multiple inner diameters. Task 2. Flow design and sample preparation: designing and building the sample multiplexing system and optimizing automated sample preparation. During this task we have further defined the flow system to be amenable to automation and reliability. Here we describe the most recent design for automated sample extraction, preparation, and injection. Briefly, i) perfusate is pushed through the MD probe to extract the solutes; ii) the resulting dialysate is mixed with acid solution (acidification); iii) the acidified solution is passed into a sample loop; iv) the content of the sample loop are injected into the liquid injection absorption cell. Sample extraction We are interested in multiplexing the sample extraction such that one TILDAS can be used to measure samples form many MD probes. To achieve this, in Phase II we have used a PEEK multiselector valve (VICI Cheminert). The version we have can operated up to 16 MD probes, all controllable via the TILDAS software, TDLWintel. Further, although the typical perfusate flow is achieved using a syringe pump, here we explored using a continuous pump (VICI Valco). This pump can source from a large volume of perfusate water, allowing it to run for extended time without refilling. We find that it is easily controllable, and provides very stable flow down to <<1 uL/min. Acidification Critical to the detection of NO3- and NO2- is their protonation to HNO3 and HONO. We achieved this in Phase I by manually adding ~10 mM H2SO4 into the sample solutions. During Phase II we have automated this process by adding in a "tee" to the sample transfer line going from the multiplexer common to the sample loop. This was achieved by again using a high precision liquid pump (VICI), drawing from a high concentration acid solution. We used a pump with a stainless steel head to avoid corrosion of the pump parts. In this setup, we dilute a 100 mM acid solution 10:1 into a sample dialysate flow. For a dialysate flow of only 1 uL/min, this requires stable flow of the acid solution at 100 nL/min. The pump readily achieved this, as tested by timing how long it took the pump to fill microliter syringe with water. Sample injection Once the sample is loaded into the sample loop, it must be injected into the absorption cell. In the Phase II proposal, we had indicated that the best way to rapidly inject the sample into the absorption cell was to use high pressure air to back the sample "plug". After extensive testing we found that this produced an inconsistent injection. We hypothesize that the low viscosity of the air pushing on a high viscosity liquid was resulting in fragmentation of the liquid sample within the sample delivery tube, allowing the air to escape around the tube. The result was a sample delivery that frequently "spit" rather than delivered in intact sample plug, and required far more air flow than should be required to inject the sample. If viscosity is important, then using a liquid flow to push the sample would be more reproducible. Indeed, we found that pushing with a high flow (250 uL/min) of water did inject the sample much more reliably. We achieved this flow by again using a water flow pump (VICI) that could inject up to 5 mL/min. Task 4: Testing, refining, and challenging the system During this task, the goal was to test and optimize the system operational parameters of the compiles system, in order to achieve true automation. We have made significant progress toward this goal. There are several parameters to test: i) acid identity; ii) acid concentration; iii) sample injection rate; iv) sample hot zone temperature. Other parameters, such as MD perfusate flow rate, are also very important, but still being explored. Acid identity We measured the HNO3 and HONO response curves when using sulfuric acid (H2SO4) vs hydrochloric acid at acidify the solution. We find that there is no significant difference in response (to within fit uncertainty) depending upon the acid dose. This indicates that the acid identity is not nearly as relevant as the presence of the acid (in this case at concentrations of 10 mM). Acid Concentration We also tested the HNO3 recovery from a nitrate solution, when acidifying with either 5 or 10 mM acid. We find that the recovery is not dependent upon the concentration of the acid, as long as it is within 2.5x of the nitrate concentration. Of course, if the acid concentration is below the nitrate concentration, there is potential for weak acids in the solution to scavenge available protons, reducing neutralization of the nitrate. This result is highly relevant, as it indicates that the acid simply needs to be in excess in order to ensure neutralization. Future test will therefore use much higher acid concentrations (~50 mM), well in excess of reasonable nitrate concentrations in soil. Testing the compiled system We created a compiled system in which we can test the ability of MD probe and acidification step to quantify the true nitrate and nitrite concentration. For this system we use PEEK tubing throughout, with no-dead-volume compression fittings (VICI) that are easily thumb tightened. The pumps and sample loop valve are computer controlled. In these tests we have not included the multiplexing valve. Importantly, we observe nearly identical responses whether we are directly sampling form solution of through the MD probe. The MD probe is therefore quantitatively recovering the nitrate and nitrite concentrations. This a very promising milestone for the project, validating the analytical approach.

Publications