HowTo – IC for nitrate, nitrite, chloride, sulfate analysis

Priority: low
Updating: mature

This HowTo describes how to prepare sample aliquots for anions analysis, operate the IC, and transcribe results into spreadsheet, then tabular database format.

Change log:

When Who Comments
2009 Sp17 originated based on oral tradition passed along from students Wei Zhang and Tony Salvucci. Refined based on advice from Shree K. Giri.
2013 04 15 Sp17 minor edits
2014 This script was forked by Sheila Saia around 2014 and that branch became the official lab documentation. sp17’s version has been evolved afterward to incorporate cleanout blanks, commercial ion standards, and other refinements but the essence is the same in Sheila’s and this version.
2017 12 18 Sp17 cumulative updates
2018 10 24 Sp17 cumulative updates.
2021 03 12 Sp17 adapted to OneNote for DEC HowTos collection
The Riley-Robb IC has been dead since 2019.
2022 09 29 Sp17 Introduces the USDA IC. This is very similar to the Riley-Robb IC. Most of the script is the same.
2023 06 29 Sp17 Recent conversion of this page to Markdown format. Restructuring to use Markdown headings. Minor edits of lab spreadsheet file conversion into database format.
2023 07 15 Sp17 Reminder to check dilutions in Chromeleon input sheet. Update for the USDA IC’s successful shutdown script.
2023 11 10 Sp17 Extensive refinements for the USDA IC. There is no sign of reviving the Riley-Robb ICs, which would cost $thousands. Rearranged section 3 to move sequence planning to first (3.2) after summary (3.1).
2023 12 22 Sp17 Minor refinements related to coding replication of analyses.
2024 09 17 Sp17 Minor edits.

ToDo:

1. Objectives

  • Determine anion concentrations in water samples or extracts.
  • Prepare data for database importing. (New section 2023 06 29.)

2. Supplies and equipment

2.1 Supplies

  • 0.5 mL Dionex IC vials. Available from Environmental Sampling Supply. (Provided by USDA.)
    • Thermo P/N 038010 vials.
    • Thermo P/N 038011 filter caps.
  • Calibration standard solutions (see separate script to make these up)
  • (optional) Ricca commercial anion standards or equivalent.
  • Pipettor tips 1000 uL (usually blue)
  • Gloves to protect from contaminating vials and samples
  • Deionized water
  • Fine tipped sharpie to label vials
  • Printout of this script to use as a checklist and reference while preparing and running.
  • Printout of your sequence indicating vial markings and full sample and calibrator IDs.
  • If using the USDA IC, make up 1L of eluent from AS22 concentrate:
    • 10 mL of concentrate, equal to 10.5g
    • 980mL or 980 g of deionized water
    • Final weight = 1000
    • Check the concentrate bottle for correct instructions.

A laptop computer may substitute for the printouts. Cornell Red Rover wifi reaches into the USDA Building and the signal is adequate in the IC room.

2.2 Equipment

  • Eppendorf or VWR 100-1000 uL adjustable pipettor
  • Dionex ion chromatograph with autosampler and control computer
    • USDA model is Aquion RFIC, running a carbonate once-through eluent. The autosampler is AS-DV. Chromeleon version 7.
  • USB memory stick to save your results
  • Old Dionex autosampler cassettes to hold vials; use cassettes to carry vials to USDA, cassettes are not used there.
  • Black cylinder used to press caps into vials (kept in R-R B72A drawer with cassettes).
  • Lab logbook and pen.

Never use the IC for analysis of samples containing high concentrations of any anion or organic matter, such as bioreactor samples and extracts from solids using KCl. This contaminates the chromatography column for subsequent users. Such samples can be analyzed only after deep dilution.

Also to protect the column, it can be a good idea to vacuum or syringe filter samples at 0.45 microns before analysis. However, the autosampler vials used also provide particulate filtration thus this is a conservative extra step. CRC stormwater samples are all filtered at 0.7 microns shortly after collection which is adequately close to 0.45 (which is specified since we stock those filters). Drinking water samples, such as from household wells, do not need to be filtered.

20220922: The USDA IC has been charged at $8/sample including vials and eluent concentrate they provide. We operate the machine ourselves. Sometimes we provide parts or supplies instead of paying the full $8.

3. Operation

3.1 Overview

  1. Reserve the IC machine a few days in advance, by emailing USDA staffer Eric Craft ejc35@cornell.edu . Make sure with Eric that there are enough supplies: vials, vial caps, eluent concentrate.
  2. Plan a sequence. Prepare vials with blanks, calibration standards, (possibly diluted) samples, and checks.
  3. Thaw frozen samples. Samples need to be completely liquid to draw aliquots.
  4. Fill IC eluent container with enough to last your full sequence. Use software to indicate initial level of eluent. Warm up the IC. Prime if necessary.
  5. Create your sequence while the machine is warming up, and fill autosampler cassettes with vials.
  6. Run the sequence, inspecting periodically for correct operation.
  7. Browse the graphical results for all standards and samples. Note where the time peaks are for each target ion.
  8. Adjust time bands for calibration of each ion. Verify calibration fit, and verify that all samples have their target ion peaks labeled.
  9. Export data from IC software to an Excel spreadsheet.

Repeat 2-8 in any additional run.

3.2 Planning a sequence

A possible order is:

  1. At least one deionized water blank. These are to clean out residues in the column from the last sequence to run, and also give additional warmup time. You may see a declining baseline conductance. That is okay as long as the baseline flattens before the end of the initial blank(s).
  2. Standards from lowest to highest. It is OK to have a zero standard first, i.e. another blank. A zero standard is optional since the machine response is very linear and the calibration fit will almost always go through zero without forcing that in the regression.
  3. One (cleanout) blank. We sometimes see some carryover between consecutive vials (not yet in the USDA IC), and the highest standard just run will have the most carryover.
  4. Your samples, including any spiked versions which probably should be after their companion unspiked version. Insert a blank after the last sample.
    • Optional: Intersperse cleanout blanks to check for carryover between consecutive samples. These have not been needed for the USDA IC.
  5. Checks. Replicate of one of the calibrators, single ion check, of certified standards from an independent source. If you are using any known or suspected high concentrations among these, include a cleanout blank after it so the following vial has an independent result.
    • Checks can be anywhere in the sequence, not just after the natural samples.
  6. One to three cleanout blanks. These are to clean out the column before the next sequence runs. If you are the next user shortly afterward, you can use a single blank here because your sequence starts with one to three blanks that will do cleanout as well as allow the machine to stabilize. If this is your last sequence, use at least two cleanout blanks here to be courteous to the next user.
  7. Shutdown blank. Has a position in sequence. As of 20231110 a blank vial is not needed in the corresponding autosampler position.

3.3a. Preparing standards, blanks, and checks

Wear gloves to avoid contaminating vials from your skin.

Label vials on lower half with fine-tipped Sharpie, before filling. You want the label to be visible when the vial is in the cassette which blocks the upper part of the vial from view.

Make up your standard solutions following the separate script. Recommend 4 or 5 multi-ion standards, one check-duplicate of a mid-range standard, and (occasionally) 2-10 ppm single ion standard for at least one ion. The single ion standards and the replica multi-ion standard are used for performance evaluation, i.e. they are “checks”. A check can be anything with known concentrations of the target ions, including a sample tested in a previous run, or a diluted commercial standard.

Single ion checks have several purposes:

  1. Scanning for contaminants in crystal reagents. For example, the NaNO3 reagent could contain traces of sulfate or chloride. (The IC is probably not sensitive enough to detect these traces … but we sometimes check.) This does not need to be done frequently.
  2. Determining which time-peak of a chromatogram corresponds to which ion in a multi-ion run. This again is not needed to be done frequently since the order of elution remains the same.
  3. Evaluating detection limits. Run several replicate vials of the same low concentration, and several blanks, and observe the standard deviations. (Needs a separate script; one was written by a SWL student.) This can be matrix specific. Consider making a spike series instead, starting from a sample that has tested at a very low concentration previously

Fill vials with one squirt of pipettor set for 550 uL. Insert cap with white filter downward. Press cap into vial so top is level with vial rim.

A blank is simply a vial containing deionized water. Blanks are used to clean out the column between groups of samples. There can be ions carried over between consecutive samples, the maximum carryover happening after a sample with high concentration.

Change pipettor tips between calibrators and checks of different concentrations. One tip can be used for all blanks, or for making multiple check or calibrator vials of the same concentration. One user makes the entire series of blanks then calibrators with a single pipette tip, working strictly in order of increasing concentration. This reduces precision very slightly by diluting the successive vials with carryover micro-droplets from the prior step that remain within the tip after ejection – probably at 600 uL this effect adds very little noise. The advantage is that this reduces plastic waste and cost.

Pacenka uses a unique tip per unique source, to reinforce the habit of changing tips which is essential between samples having unknown concentrations. The greatest precision risk is that a high concentration followed by a low concentration using the same tip will raise the low concentration. We observe the same effect within the chromatographic column.

3.3b. Making some samples spiked with nitrate (optional)

If you want to add +X mg/L of nitrate to a field sample, to check for a matrix effect perhaps, you will mix together some sample and a small amount of 1000 ppm (or farther diluted with deionized water) nitrate-nitrogen standard. The concentration in the mixture is related to the relative volumes and concentrations in the standard and the sample. Let Vo and Vs be the volumes of original sample and spike, and C0 and Cs be the concentrations in both. The concentration in the mixture should come out as

Cx = (C0*V0 + Cs*Vs)/(V0 + Vs)

You need to test the sample separately without spiking to determine the C0 value.

It can also be interesting to analyze a series of two or more spike levels all starting with the same underlying sample. This can be called a “matrix spike series” and is commonly done in “recovery” tests for synthetic organics. This should start with an unspiked version of the base sample to get a baseline. It is best done based from a sample that has tested at a non-detect level in a previous run. This can even provide an alternate calibration series that is more realistic than using calibrators made from deionized water.

Spiking can be done using stock standards made from crystals or commercial liquid standards.

Spiking can be done for more than one ion. Nitrate is only one example of a single ion spike.

3.3c. Preparing sample vials

If samples need to be diluted for concentrations to be within calibration range, dilute them using deionized water. Use clean intermediate containers (such as 15 mL centrifuge tubes) to hold the diluted samples, labeling them as diluted. Mix via inverting a few times with cap on.

If you are spiking a sample that needs dilution, be aware that the math of interpreting the result includes a multiplicative dilution factor. You spike after dilution. So if you spike +1 in a sample diluted by a factor of 5, your calculated result should be +5 higher not +1 higher. This may appear to take you above calibration range; however the machine is seeing the +1 on top of the diluted value thus the range is expanded. It is best to spike low concentration samples rather than high concentration ones.

Label vials on lower half with fine tipped Sharpie. It is best to number the samples 1..N, standards as ST1 through ST5, blanks as BL, etc. Keep a log or a spreadsheet (printed in advance) with the full sample IDs, vial marking, and sequence position.

The vials should be filled with 0.55 mL of sample or standard or blank water. Set adjustable pipettor for 550 uL.

Change pipettor tips between samples.

Note: different samples of a sequence can have different dilution ratios. Different dilutions of the same sample can be included in the same sequence.

Note: filled vials may be refrigerated for one or two days before analysis. Do not freeze! The samples do not need to be warm to analyze since the IC will warm them to a constant temperature.

3.4. Warming up the IC.

The IC needs to reach operating pressure and low specific conductance before analysis.

  1. The control computer is always running and the IC with autosampler are usually running. Make sure they are powered on and quiescent. Link the control computer to the autosampler if needed.

  2. Top up the eluent bottle with bicarbonate solution if it initially contains less volume than will be needed for your sequence. Inform the Chromeleon software how much eluent is in the bottle (up to 2L).

  3. Empty the waste receiving tank (under bench) if it has more than 1L in it. A 38-vial sequence adds around 800 mL to waste, and the tank can hold around 3L.

  4. Prime the IC if it has not been used recently or if it has been out of eluent. This is done via a syringe and Chomeleon software.

  5. Set the flow rate to 1.2 mL/minute and current to 41. The IC will begin to run when the flow rate is set above zero. The pressure should reach at least 1800, and be stable, or else the machine needs priming.

  6. The Dionex manual says the maximum target conductance is 30 uS/cm when using an anion column. This is reached quickly. Let it drop below 20 before proceeding.

3.5. Entering a sequence into Chromeleon

Entry can be done while the IC warms up.

Copy then paste one of your own earlier Chromeleon sequences. Then adjust the contents to match your new sequence. Make sure to “fill” in the new vial number column if you have inserted or deleted anything. The copied sequence will include raw data and indications of completion of earlier vials. Delete all raw data using a menu command, and set all vial statuses to Idle instead of Finished or Interrupted.

Fill in dilution values for your diluted vials, if any. Blanks, checks, and calibrators should always have dilution = 1. The software will multiply the concentration result for this vial by this dilution value. Spiked samples require extra care in interpreting results unless they have true dilution = 1. Changing standards values from the earlier run … (possible but I forget where it is in the Chromeleon menu structure)

Remember that if you begin with a copy of a prior sequence to adjust the per-vial dilution values. These are not visible on the screen unless you scroll to the right.

Save your sequence.

3.6. Running the sequence

The USDA autosampler is a carousel. Press the button near the front to release it from gearing and move the carousel to expose your range of positions to fill.

Pull down the batch or Queue menu or button in Chromeleon software. Add your sequence to the queue list; delete any others. Press the “ready check” button. There may be warnings. One warning will be an estimate of how much eluent is needed; this is your last chance to top up the container and inform the Chromeleon software of the new starting volume.

Click the start button in Chromeleon’s batch (Queue) dialog. Warnings will trigger a dialog to confirm proceeding.

A long time passes, 16 minutes per vial. (Thus 50 vials take over 13 hours.) Watch the first one or two vials load and be analyzed.

Leave a note with your email address on the computer or autosampler, in case someone else inspects and sees a problem.

Visit the running machine every few hours to verify that eluent has not run out in the USDA machine. If you observe that the eluent is low, you can refill it by loosening the cap and pouring in more eluent mix slowly (minimizing bubbling) while the machine is running, making sure to keep the intake hoses in the bottle completely immersed so no air gets sucked in.

Visit the running machine within one hour after it is scheduled to end, to verify that the shutdown has run. If it has not run, run shutdown manually via the Chromeleon software. The USDA system has a reliable shutdown script, but if the sequence crashes before the shutdown position the shutdown script will not run.

If you fail to remove your spent vials, the next user has the right to dump them. It is discourteous to leave your vials for someone else to dump. Some spent vials will have their caps high, as if they were never drawn from. Probably this is a bounceback due to pressure rather than a missed injection.

3.7. Browse results and note times of ion peaks

From sequence browser (“Data” tab on the USDA system) with your sequence selected: Double click on the first vial’s entry. Scroll through the vial plots watching for the following: Has the baseline become flat by the third blank? If not, you needed more warmup time or the device needs maintenance.

Write down the times of your ion peaks in each standard. After all standards are visited, find the range of each ion peak and compute its midpoint.

Did any peaks overlap? That is a sign that you need to use lower maximum standard values. You can disable use of the highest standard in calibration, per ion, if necessary. Check the blank following the highest calibrator. Is its plot similar to the earlier blanks, or are you seeing some carryover from the highest calibrator? Also check the blanks between groups of samples.

Browse your samples. Look for your target ions in each. Are there any peaks that are close to but outside the time ranges for your target ions? Consider widening your time ranges a little if so. Also look for overlapping peaks for your target ions. These indicate a need to dilute and rerun. Check your replicate standard and single ion standards, if used. The replicate should have a similar plot to the earlier standard. Single ion standards should have peaks within the time ranges of their respective ions from the calibration standards.

3.8. Adjust time bands to reflect browsing results

In the Chromeleon software, re-express the observed ranges as midpoints and half-widths. This is done by double clicking on the Nitrate.qnt method line toward the upper right of the sequence screen. This is also the mode to verify that the concentrations in the calibrator vials are entered correctly and to observe the calibration linearity.

Verify good calibration fits. Disable selected standards if they are problematic, document why. Re-browse your analytical results now noting that the peaks are labeled with ion names and noting calculated concentrations. (no need to write down unless you see a problem)

(needs to identify where in Chromeleon 7 to find this; some screenshots would be nice)

3.9. Save your results

Create a new file in wordpad. This is under the Windows Accessories part of the Windows applications list.

For each anion in turn:

  • Navigate to that anion in the middle window. This will reveal that ion’s results for all vials in the lower windows.
  • Select all columns in the lower window. Copy them to clipboard.
  • Paste into wordpad.

Save the wordpad file under a suitable name; RTF format is OK. Move the file to USB memory stick, and transfer file to a computer having Excel. Copy from Wordpad and paste into a new Excel file, one sheet per anion.

3.10 Evaluate results

Scan the results to decide on lower and upper censoring bounds (detection limits) for each target anion. The concentration results from blanks, lowest calibrators, highest calibrators, and low checks should be your guide. Also consider inter-ion effects by looking for superposed peaks in the chromatograms. Unless you have evaluated detection limits formally for this matrix via manifold replication (separate script), you should set your lower detection limit conservatively high and your upper detection limit conservatively low.

Decide which samples need to be reanalyzed with dilution, and how deep to dilute. For example, you might choose to dilute any samples that have calculated values above double the highest calibrator for any ion. If you dilute by a factor of 5 you raise the low and high detection limits of every anion by a factor of 5.

  • Special approach #1: If you want to observe the masking of nitrate by sulfate, make up several check samples with constant high sulfate value and increasing nitrate values starting at a low level. High nitrate will also come out of the column earlier, thus high nitrate and high sulfate should still result in distinct peaks. It is only the combination of latest (lowest) nitrate with earliest (highest) sulfate that cause interference.

  • Special approach #2: Make a matrix nitrate spike series based on a sample exhibiting interference. (This is called the “method of standards addition”.) You may be able to extrapolate a single-sample calibration curve that properly separates the superimposed nitrate pulse (by moving it earlier) from the early sulfate pulse. d

  • Special approach #3: Make several dilution levels of a sample exhibiting interference. Use the result for nitrate from the sample with least dilution that provides a reliable separation of peaks. The software should be able to separate a visually distinct pulse on top of the rising limb of the sulfate peak. The software cannot separate peaks when there is just one asymmetric sulfate peak visible.

There have been no superposed peaks in the USDA IC, and we use the eluent strength specified on the concentrate bottle.

For other reasons for dilution than interference, you can decide to report a result as above your high detection limit instead of rerunning with dilution.

To your Excel file for the run, add a summary sheet that copies via formulae all computed ion concentrations from their individual tabs (as pasted from Chromeleon via Wordpad) onto the summary tab. Note your intended calibrator, replicate calibrator, and single ion check values next to their corresponding analytical results and include comments about the quality of the calibrator and check results. Also remark about the cleanout blank results, noting any carryover. Use background colors green for good result, yellow for suspect but usable, and red for “don’t use this”. Samples needing to be rerun with dilution should also be noted in red.

If useful, you can include in your summary tab of the spreadsheet some columns with formulae that reinterpret the reported results by applying sample-by sample or default detection limits. This would substitute “<0.1” for a non-detect (for example), copy the value for an in-range result, and “>100” for a censored high result.

Annotate your file to identify your sample batch, dilution levels, or other batch notes. For pesticide project work, you need the date that the analysis was done; this can be in the file name.

Save spreadsheet to your memory stick.

4. Transcribing results data from spreadsheet to database

For DEC pesticides work, we must prepare for submission of most data to EQUIS, thus we have a database format derived from EQUIS requirements. To transcribe the data from machine-oriented spreadsheet into database format involves the following:

  • Expression of results as number, qualifier string, and detection limit in separate columns.
  • One row per sample per parameter. Spreadsheet contains one row per sample, multiple columns for multiple anions.
  • Sample identifiers in multiple columns matching project. Typically site ID, sampling point ID within site, date sampled, time sampled. Replicate analyses including different dilutions should have different replicate numbers starting with zero (which means the original) for the earliest run.
  • Any QC samples such as checks and spikes need special encoding of their identity to link to the natural samples that they apply to.

The process for this conversion is described in HowTos for DEC pesticide project 2021-2025, specifically the DEC pesticides results transcription.