HowTo – Storing samples at Cornell

Priority: medium
Updating: very mature from prior projects

Change log:

When	Who	Comment
2021 03 04	Sp17	First version based on earlier DEC sample handling.
2021 08 03	Sp17	Fleshing out toward reviewable. Updated to indicate that fridge monitoring alarm includes too-low as well as too-high temperatures.
2021 10 01	Sp17	Reformatted for export in DOCX.
2022 08 18	Sp17	Minor update. Bumped to 90%. DEC lab confirmed that we do not need to filter samples, they will filter as needed.
2022 05 03	Sp17	Revised the storage protocol to include acidified, refrigerated aliquots for cations analysis. Revised the storage protocol to include discarding of samples to live within limited freezer space.
2023 06 21	Sp17	Converted to Markdown, minor updates.
2023 06 22	Sp17	Added principle for sample handling – original container is sacrosanct, minimize volume removed, never put anything back into it.
2023 07 30	Sp17	Minor updates, adpated to Quarto.

1. Objectives

Maintain chemical integrity of samples while they are stored at Cornell. These are intended to be stable reflections of chemical conditions at the time the sample was collected.
Preserve remaining sample volume in case it is needed later for an unforeseen opportunity.

2. Quality assurance aspects

This includes preservation of volume for unforeseen opportunities.

Minimizing time out of freezer for samples during processing.
Detecting freezer failures early. Having backup freezer capacity in case of extended failure.
- There are temperature monitoring devices in one fridge and four freezers. These report to internet. There are SMS alarms routed to Pacenka if freezer temperature exceeds 35.6F for more than 10 minutes (allowing for auto-defrost cycles). There are alarms routed to Pacenka for the fridge if it exceeds 40 degrees F or goes below freezing.
Preserving power during AC outages.
Recordkeeping of failures.
- Subsamples affected by freezer and fridge failures should be noted in the HowTo: Cornell lab log
- Subsamples left in fridge for more than two days following thawing should be noted in the lab log.
In general we divide samples across two or more containers in case there is a loss or storage failure involving one container.
In general we collect extra volume just in case there are useful analyses that we don’t think of early.
In general never put any sample water back into an original field container.
In general use only a minimum of sample water for analysis. We do almost no rinsing of containers with sample water except while sampling or filtering.
Organization in storage: Finding samples for analysis and making aliquots.
Purging samples when they have expired in utility.
- There is a spectrum of opinion about when to purge. BKR keeps samples forever (no time machine available), sp17 likes to have open space in freezers and fridges to aid in sorting and retrieval. BKR adds freezers, sp17 tries to make do with what we have unless it is failure prone.

3. How we store samples

Guiding principle based on DEC lab experience: In general samples should be frozen from shortly after they first reach Cornell until they are disposed of as no longer needed. It is accepted to thaw a sample for making an aliquot for the DEC lab, or for analysis of the sample at Cornell for auxiliary parameters. Samples that are temporarily thawed should generally be refrigerated, for no longer than four days. Samples that need to be at room temperature for analysis should be out for only a few hours, and returned to fridge as soon as possible after sample analysis begins.

Life cycle of samples in storage:

Initial freezing. Samples are carried on ice from the field to Riley-Robb. In a few cases, such as a late night return from field, they will be stored in fridge until the following morning.
Organization within freezer.
Removal from and return to freezer.
Disposal of older samples.

The four freezers have emergency backup AC power that kicks in quickly during AC power outages. The freezers and fridge have Ambient Weather temperature sensors that report to internet and there are alarms if there are temperatures outside of programmed limits. The alarms come via SMS.

The four freezers are all less than 8 years old and should have at least a decade of useful life remaining. We have one weaker freezer in reserve.

The project freezers (DEC and USDA funds) are expensive commercial models (restaurant industry?) that are well insulated, and should be able to keep samples frozen for days if backup power fails and the doors remain closed.

Our double-wide fridge is less than 2 years old as of mid 2023. It is also a robust commercial model for the restaurant industry.

We avoid using the other freezers and fridges in the SWL facilities. Those are mostly used by Todd Walter’s students. They are more residential grade models, older, and contents are minimally managed. They are mostly full with departed students’ items.

3.1 Initial handling upon return from road trip

The initial storage upon arrival depends mostly on if there is any in-lab processing before the first freezing. Samples would be refrigerated before in-lab processing. This will depend partly on the selection of analytes. Initial processing of stormwater samples for glyphosate, for example, emphasized the early separation of particles by vacuum filtration through GF/F filters. (The DEC lab indicated that filtering is not necessary for their purposes, thus we have not specified how to do filtering.)

As of early 2023, we began to make acidified 40 mL aliquots for cations analysis before first freezing. Otherwise we pour off any extra water in the container, and put it in freezer. Multiple bottle samples are in gallon ziplock bags.

Lake samples are already frozen by the volunteers before shipment to Cornell. These are transferred to a Cornell freezer within a few hours after express shipping brings them to Cornell. The earliest samples thawed partly, and acceptably, during shipment.

3.2 Organizing samples in freezer

Clustered organization within the freezer facilitates retrieval of the intended subsamples. Typically geographically and topically related samples are stored together in bags, with bags in boxes with their “peers”, usually boxes for location and bags for similar sampling times with the given location. All subsamples of the same sample should be in the same bag. The bags and boxes are labeled with group identification; bags may be unmarked if the bottle labels are legible through the bag. Most samples are divided across multiple containers, partly for extra volume, and partly to have a second or third container in case a container is lost or breached.

We use boxes from Uline, 10 inch cubes, with the top flaps cut off. Freezer shelves are positioned to fit boxes with a little extra space above.

3.3 Thawing a sample for analytical purposes

Removal from a freezer for sample processing should be as brief as possible, within the constraint of not thawing the same sample container too many times.

It is preferable to thaw samples in a fridge overnight over thawing at room temperature; in exceptional circumstances, thawing may be accelerated under warm tap water.

A sample must be thawed completely for making an aliquot, doing a more direct test, or putting it into a new container in case the container is damaged. It needs to be completely mixed before drawing liquid out. Some analytical procedures require the sample to be at room temperature at the start of a test. It is preferable to begin with completely thawed but still cold samples, quickly draw portions into the test containers, and return the original sample to fridge. Let the portions to be analyzed warm in the test containers. Discard these warmed portions of samples after analysis. After returning the original container to fridge, it may be preferable to hold the samples in fridge for a day in case they might be needed again for retesting; then return them from fridge to freezer. When they return to freezer, restore to their original grouping bags and boxes so they can be found again.

When we divide a field sample across more than one container, there is a letter suffix on the container label after the sample ID to indicate which container of a sample is being worked with. Different containers of a sample should be assumed to have the same sample in them. However it is best to use up subsample A for early thawing and processing, then proceed to B and then C. This lets the longer stored subsamples have fewer cumulative thawings.

When making duplicates for parallel analysis, it is preferable that they come from the same subsample so that different prior thawing counts do not introduce another reason for different analytical results besides variation in the analytical stage. This also applies to unfortified and fortified (spiked and unspiked) recovery test duplicates.

The above reasoning also applies when preparing multiple aliquots of the same sample for DEC - all aliquots from the same subsample when possible. Through 2020 we would make two 50 mL centrifuge tubes of aliquot per sample, each containing no more than 45 mL (to allow for refreezing). The letter suffix rule also applies to the parallel aliquots – but the letter suffices on the aliquots are unrelated to the suffices on sample containers.

4. Culling samples to recover freezer space

Freezers and room are expensive to archive samples for years, and the representativeness of samples declines with storage time. Thus samples should be discarded after their probability of being needed becomes small. We begin with three 250 mL bottles in most cases, discarding one bottle after the first 2x45mL aliquots are made for DEC and aliquots are made for cations, alkalinity, and anions analysis. After DEC returns analytical data OR June 30 of the following year, whichever comes later, we reduce to a single bottle. Make sure that this bottle is nearly full if possible.

5. Historical problems and reactions:

Inexperience has led to samples not having enough airspace to allow for freezing, thus the ice will break out of the container during freezing. As a general rule, we allow at least 10% airspace. Thus a 50 mL centrifuge tube should not contain more than 45 mL of liquid water.

Another aspect of the 50 mL tubes in freezers: If they are stored on their sides in freezer while liquid, ice expansion laterally during the freezing process may push off the cap even if there is airspace above the ice. Thus they should be frozen while held vertically in a rack. This also affects the DEC aliquots. Once a container is frozen, it may be removed from a rack. (This problem has not affected nalgene wide mouth containers thus we don’t worry about container orientation during initial freezing.)

Refrigerator failure 2020 12 22 vicinity: The fridge’s thermostat failed and it went below -10C for at least a few days. The contents froze hard. No past DEC samples were affected. Freezing destroyed some ELISA kit reagents and damaged some lab experiment samples (glyphosate washoff from plant residues) that had inadequate air space to accommodate freezing. In general any containers that are intentionally frozen must have a roughly 10% allowance for expansion during freezing. The unexpectedly frozen samples whose glass test tubes cracked had almost no air space and had semi-tight fitting plastic caps. Some of the force was exerted against the cap, but some of the test tubes cracked.

Response: The samples were salvaged by partial thawing, then pouring off their liquid and remaining ice into new containers. Freezing violated protocol for these samples slightly, because they were being held in fridge for early ELISA testing just after they were created. Most had already been analyzed but a few needed reanalysis, thus the experiment had some results from frozen samples and some from samples that were never frozen.

Freezer failure 2020 12 22 vicinity: (no past DEC samples affected). This freezer was left behind by a faculty member who left for another university. It tested OK initially. It was later discovered to have a weak door latch. Someone had opened it and not closed it tightly so the door opened and stayed open unnoticed for at least a couple of days. The contents all rose to room temperature as the freezer could not cope with having an open door. This was discovered on the same day as the December 2020 fridge failure. This freezer was not yet highly trusted so was not used for recent higher value samples; we had finished sorting 2020 samples just before and had not previously used this freezer until then, thus were unaware of its door problem.

Response: A few of the samples in the freezer had no owner identified and were dumped. The remainder were refrozen and left in the reactivated freezer, with the boxes marked about the container failure. The entire contents are now considered untrustworthy.

Improvement: We replaced this freezer in 2023.

During 2020 and 2021 we could not allow undergraduates students to use our lab rooms due to Covid-19 protocol conservatism. We decided to sacrifice some subsamples by sending them home with students for additional parameter testing. The samples had already been tested for glyphosate. The test methods for chloride and specific conductance in the students’ homes did not rely on preservation via freezing or even refrigeration. The samples left Riley-Robb in heavily insulated chests and with ice packs. They came back at room temperature a week later. The students were unable to refresh the ice or re-freeze the ice packs. We decided to keep the returned aliquots; they are refrozen and their group bags marked that these subsamples are only for additional testing of parameters unaffected by room temperature heat or use as a last resort. Improvement: Mark the individual containers, perhaps with red stick-ons, not only the grouping bags.

As in the 2020 12 22 fridge incident, there have been other times when we have needed to put a sample into a different container if we drop it while frozen and the container fractures or if it was frozen with inadequate expansion space. To salvage a frozen sample:

Wash off the exterior of the fractured container with deionized water, taking care not to further fracture the container or let DI water into the container via a crack.
Place the fractured container into a clean larger beaker and allow it to thaw in fridge. It may leak a little through cracks.
Prepare a fresh sample container with label identical to the original. This should be large enough to hold the entire remaining sample. Indicate on the label
Pour off the thawed sample into the new container, omitting what has leaked through cracks into the encompassing larger beaker.

Samples in Nalgene wide mouth containers, our field standard for several years, can fracture their caps without losing sample integrity. The outside part of the cap fractures but the part of the cap inside the neck is intact. Thaw as above and carefully tease out the remains of the cap, perhaps using a knife blade to pierce the cap remains to make it turnable in the threads.