Liquid nitrogen storage tank failure: Can we improve the current system?

Consider This
Liquid nitrogen storage tank failure: Can we improve the current system?


Kimball O. Pomeroy, Ph. D., H.C.L.D.
The World Egg Bank
Phoenix, Arizona

Consider This:

On March 4th two separate fertility centers reported malfunctioning equipment involved in keeping frozen eggs and embryos safe. University Hospitals Cleveland Medical Center estimated 4,000 eggs and embryos may have been damaged by a malfunctioning storage tank. The tank had been having problems with its automatic refill and the clinic had been working with the manufacturer to remedy the problem. Apparently, an alarm system that was designed to detect a temperature rise was turned off by an employee (1). Another tank failure occurred at Pacific Coast Fertility where it was reported that the level of a liquid nitrogen storage tank was very low and may have damaged some specimens (2). In order to help initiate a rational solution to these recent failures, this article will present the current status of tanks and alarms used for storing reproductive tissue, weaknesses of this current system will be discussed and finally, suggestions on how to proceed to reduce the frequency of these tragic events.

These recent accidents bring several important factors to light, the most important being, “How safe are specimens stored in liquid nitrogen containers?” First, it should be noted that the tanks used for the storage of human embryos, sperm, and eggs were developed mainly for the livestock industry. They were designed for the storage of bull semen and cattle embryos. Manufacturers have most likely designed these tanks for the much larger market of the livestock industry, consequently there are few systems that will prevent these tanks from failing. These storage tanks have remained relatively unchanged for the last 20 to 30 years.


Two major types of tanks are used for storing human eggs, semen and embryos. First, the large high-capacity tanks that usually have automatic refill functions. These tanks can hold 100 to 500 liters of liquid nitrogen and hold tens of thousands of specimens. They are attached to a supply tank that provides liquid nitrogen to the tank when the internal sensor detects that liquid nitrogen level is low. These tanks also often can detect nitrogen levels and temperature and will trigger an alarm when the temperature rises above a specified level.


Small tanks (30 to 60 liters of nitrogen) are the second type used for tissue storage. They have fewer functions. Usually they have no auto-refill functions, do not measure temperature and only alarm when nitrogen falls below a specified level. They can often last over 100 days if filled properly and not opened. These tanks are also mobile and have the benefit to not “put all of one’s eggs in the same basket”, but they are also more susceptible to failure due to them being moved, lifted and banged around. They can hold from 500 to 2000 specimens, depending on the rack system used and the type of specimen container (vial, straw, etc.).

Figure 1. Diagram of a small-capacity (47 liter) liquid nitrogen storage tank. The vacuum is between the inner storage chamber and the outer shell. Specimens are stored in the inner storage chamber. 

Tanks are made of two sandwiched layers of aluminum or aluminum alloy that have a vacuum between them (See Figure 1). Air is removed from the outer chamber, and a partial vacuum is created in the area between the two chambers (3). It is this vacuum (like in a thermos bottle) that insulates the tank and reduces the amount of conductive energy loss and allows the liquid nitrogen to last longer than otherwise. The inner storage chamber of the tank is suspended from the outer shell of the tank by the neck tube. Stress on the neck tube, which can be caused by jarring or excessive swinging motion, can crack the neck tube, resulting in vacuum loss and tank failure.  Another important item to note is that these tanks are Dewars that conduct the cold from the bottom of the tank to the top of the neck. This means that even a small amount of liquid nitrogen in the tank will protect the entire tank from top to bottom. An intact tank can lose most of its liquid nitrogen yet still keep the entire contents at the appropriate temperature. How well this Dewar works when vacuum is lost is not known.


Some tanks are stored off-site, in isolated rooms, or even near the busy workings of the laboratory. The benefit of the latter is that workers will pass by the tanks many times every day. Most laboratories follow CAP recommendations to check nitrogen levels at least weekly on every tank. This should be sufficient for a tank that can last over 100 days without refilling. If a catastrophic failure occurs though, as may have happened in the Cleveland case, the only way to save the tank may be to be on-site at the time of the failure.


Catastrophic failure can occur when a tank is refilled many times and it is lifted, banged into other tanks or has aged. The welds in the metal, especially at the site where the original vacuum is pulled, deteriorate and air enters this vacuum space. Alternatively, the neck may separate from the outer shell due to repetitive abuse. It has been proposed that the inadvertent spilling of liquid nitrogen around the site of where the vacuum was pulled may cause repeated contraction and expansion, thus weakening the integrity of the site. This leak of the vacuum may occur slowly such that the tank gradually fails.


Theoretically, a tank can fail by developing a hole that goes through both sheets of aluminum. If the leak is near the bottom of the tank, within minutes the temperature will rapidly rise, rendering the specimens worthless. Responding to the alarm in this case may not even be enough of a notice to save the contents. There have been no reports of this type of tank failure.


To detect the slow failure of a tank due to vacuum loss, embryologists track the amount of liquid nitrogen required to refill the tank each week. When this volume starts to slowly creep up, it is time to replace the tank before it fails completely. Another hallmark of a tank that has lost its vacuum is the present of frost around the tank’s neck or vacuum fitting (3). It is common protocol for most IVF laboratories to discard tanks after 10 years of use.


The two recent failures may prove costly for both IVF clinics. One suit is seeking $5 million in damages. Even if most suits were to seek just the cost of new IVF cycle for each patient, the costs of a failed tank could cost from 1.2 to 6.0 million dollars. (Using a conservative estimate of $12,000 per IVF cycle, 100 patients in a small tank or 500 in a large-capacity tank.) Clearly, with these potential risks all IVF clinics should mitigate these risks by investing in state-of-the-art storage tanks and alarm systems, maybe even shift these risks through storage by an off-site commercial storage bank.


Costs of tanks for an IVF clinic can be substantial. A good 250 liter high-capacity tank costs about $10,000 while several small 35-liter tanks may cost about $1000. A large clinic might own several high-capacity tanks, or they might have 20 to 30 of smaller Dewars. These smaller tanks usually have a single temperature probe that monitors the tank’s status. If this probe is placed such that is near the bottom of the tank, there will be little warming time if the temperature rises. Ideally, it should be placed near the top of the tank. Many clinics have gotten by with inexpensive alarm monitors that will call embryologists when the alarm is triggered, but the recent failures of these tanks will most likely result in the purchase of more advanced tanks that not only monitor temperature, but also monito nitrogen levels. There will also be a push to purchase even more advanced alarms.


There are several things that can be done now with current storage tanks to increase the safety of stored tissues. First, wherever possible these tanks should be located in well-trafficked areas. This will allow for a more continuous surveillance of the tanks. The storage of tanks in some secluded tank room should be avoided. In addition, specimens from one patient should be split into more than one tank. This will mean that if one tank fails, not all of a single patient’s specimens will be destroyed. This may result though in more liability for the laboratory as there will most likely be more patients in each tank as opposed to when all of a patient’s specimens are placed into one tank.


Next, all monitoring of the tanks should include weekly alarm testing, manual measuring of liquid nitrogen levels, plotting of the nitrogen usage of each tank and the discarding of any malfunctioning tank with a replacement tank. By plotting the levels of liquid nitrogen prior to refilling of the tank, one can detect when the vacuum of the tank begins to fail. Refilling then must be on about a weekly basis. More frequent refilling will not allow one to see the drop in liquid nitrogen level as accurately. If a tank is found to be questionable, it should be immediately removed and replaced with a new tank while the problem is undergoing troubleshooting.  The last thing that should be done in the weekly tank check is the verification that the alarm system is working. Because the alarm system must be usually turned off during tank maintenance, this must be the last test. This alarm test should test the entire alarm system such that each tank is induced to fail such that a phone number is dialed and a response must be received.


Ideally, tanks should monitor both temperature and nitrogen level. This does not happen on most small tank systems where usually only the nitrogen level is monitored. A drop in a tank’s nitrogen level should precede that of an increase in tank temperature. The nitrogen sensor must be placed such that it will trigger well-before a minimal level of nitrogen is present. This often means using a short probe so that the sensor is just above the lowest level where tissue is located.


The movement of tanks in the laboratory can often result in them being banged around and potentially weakening welds. For this reason, movements of these tanks should be restricted, and tanks should be handled delicately. Consideration should be given to bringing the nitrogen to the tanks or even installing an auto-refill system.

There are many types of alarm systems for tanks (See Figure 2) but they all function similarly. The alarm system consists of a probe, the alarm and a response system that can be programmed to call a phone number (or series of phone numbers) when an alarm is detected.

Figure 2. A typical hard-wired alarm system for storage tanks. Two small-capacity tanks have a nitrogen level probe attached to the tank. This probe then is attached to the wall-mounted alarm module which is attached to the response system.


The alarm probe is usually a long wire with a capacitor affixed to the bottom. When the level is below this capacitor (the end of the probe), the capacitance is reduced, and an alarm is sounded. As long as there is liquid nitrogen contacting the capacitor though, the alarm does not sound. Liquid nitrogen probes can be developed that do not just detect the level at a point, but that can determine the liquid level along the entire length of the probe. This type of probe is currently only found in high-capacity tanks.


The probe is attached to an alarm that is programmed to produce low voltage at the alarm output when the capacitance of the probe drops below a specific value. In most alarms there is also an audible sound at the tank. The alarm output is attached to a response system that follows a program telling it what phone number to dial, how long to continue trying to dial that number and when to go on to the next phone number. More advanced response systems will also send emails or texts that might report the problem and which piece of equipment is alarming. The connection from the alarm to the dial out system can be either hard wired or can be wireless. These response systems often can provide an alarm and dial-out to notify the embryologist that the power is out. In this case, a battery back-up kicks in. Regardless, these systems still must rely on either a phone line or an internet connection (or both). If the phone line is down or occupied (if it is a shared line), then the dial-out will not occur until communication is reestablished. Response systems that work purely through the internet may be more susceptible to not being able to establish communication if the power is out, especially when the outage encompasses a wider area. All systems must be constantly monitored and tested to ensure they are working.


The long-term solution to tank failures should involve the gathering of information from all known instances of tank failures so that the cause of these failures can be determined. An improvement of this system then should address the past instances of tank failure. Were there failures due to negligence where alarms were not used, alarms were not monitored, or tanks were abandoned? Or were there instances where tank failure was not detected due to faulty phone lines, complex dial out procedures or too many false alarms? What was the time frame of when nitrogen evaporation occurred? Was it a chronic leak or did it occur in just a minute or two? Answers to these questions will dictate whether minor changes will fix the problems or whether the tank and alarm systems need to be redesigned from scratch. It may be impossible to obtain this information especially with the looming litigation. To attempt to fix these problems though without first identifying the root causes could act as little more than a knee jerk reaction. The fertility industry should work together to identify the sources of all past tank failures and design procedures and systems to address these “real” issues.


In designing a new, safer storage system, consideration should be given to the following:

  1. Have the ability to determine which tank is alarming.
  2. Provide the temperatures and nitrogen levels over the internet or phone.
  3. Constant monitoring of the vacuum status of the tank.
  4. Intelligent programming that will take information from the auto-refill system, temperatures, nitrogen levels and vacuum status and be able to predict imminent tank failure.
  5. When tank failure occurs, be able to provide a quick supply of liquid nitrogen.
  6.  Design tanks so that there is a backup to the vacuum layer – maybe by making a tank with three layers of aluminum and a vacuum between each layer.


As we demand more security and safety in our industry, the costs for doing this will increase. Embryos and gametes used to be shipped mainly by national couriers throughout the US. With failures of these systems that were not designed for the specialized needs of shipping one-of-a-kind valuable human specimens, many have resulted to using specialized shipping services that provide insurance, personal couriers or having the patient perform the transportation. What used to cost $500 then, now costs thousands of dollars to ensure the safety of the tissue. Some clinics even refuse to be involved in the shipping of these priceless tissues. Expect to see the same once storage, specific for human embryos and gametes, is developed. Costs for storage will increase dramatically to pay for this increased safety. Storage of these tissue may also shift from storage at the clinic to companies that provide off-site storage at secure facilities. Unfortunately, this will result in an increased liability as the tissue will need to be shipped to the off-site facility and then to the IVF clinic when it is needed.


Once data from known tank failures is gathered it will take the combined resources of embryologists and manufacturers to design a system that is less prone to failure. This system will most likely include changes to monitoring protocols and changes in the manufacture of tanks so that these terrible accidents can be avoided.


(1)Buduson, Sarah, (2018, March 27), Retrieved from

(2) Ho, Catherine, (2018, March 14) Retrieved from

(3)Nebel, L. Raymond, 2007, “Techniques for Artificial Insemination of Cattle With Frozen-Thawed Semen”, In: Current Therapies in Large Animal Theriogenology. Robert S. Youngquist and Robert R. Threlfall eds., Saunders Elsevier, St Louis Missouri,  pp 253-258.