Questioning what “nontoxic” really means in the context of refrigerant safety.

I have been delving into the art of risk assessment for industrial refrigeration systems and unearthed some surprising and interesting quirks. Here are a few anomalies that help keep the daily workload interesting. I preface this column with a disclaimer that I am not suggesting that we abandon safe working practices—just sayin’.

The international standard ISO 817:2014 recognizes two classes of refrigerant toxicity, described in clause 6.1.2 of the standard as “lower chronic toxicity” (class A) and “higher chronic toxicity” (class B). ASHRAE Standard 34-2022 defines toxicity as “the ability of a refrigerant to be harmful or lethal due to acute or chronic exposure by contact, inhalation or ingestion” and goes on to say “For this standard, temporary discomfort at a level that is not impairing is excluded.” ISO 817 doesn’t have a definition of toxicity but it does define acute toxicity as “adverse health effect(s) from a single short-term exposure” and chronic toxicity as “adverse health effect(s) from long-term repeated exposures.”

The puzzle is that study of cause and effect in dozens of documented serious industrial refrigeration accidents showed that when fatal accidents occur with ammonia the casualties are almost always within a few feet of the point of release. However, the traditional methods of analyzing the consequences of ammonia releases don’t seem to acknowledge this and use calculation techniques more suited to other chemical hazards. Looking more deeply into references on ammonia toxicity, I found that there are no chronic toxic effects and the threshold level for serious permanent injury from inhalation was more than 1,500 ppm. So why would it be classed as “higher chronic toxicity” by ISO 817?

The threshold limit value (TLV) for long term exposure to ammonia is low because it has such an irritating reaction with moist body parts (eyes, nose, throat) that the occupational exposure limit (OEL) is therefore typically set in workplace regulations as 25 ppm. This value assigns it to class B in ISO 817, but this has nothing to do with toxicity; it is what Standard 34 calls “temporary discomfort at a level that is not impairing,” an effect supposedly excluded from the classification process.

This reminded me of my favorite refrigerant safety joke (credit to Professor Joachim “Joe” Paul, former head of IIR’s Science and Technology Council): Joe was a great proponent of the use of water as a refrigerant in vapor compression systems, but he always gave the stern warning that water was obviously the most hazardous refrigerant “because it’s the only one that contains sharks.” There is a serious point behind Joe’s quip: the assessment of risk has to be relevant to the circumstances. Suitable metrics and methods need to be used, otherwise the conclusions reached will be inappropriate and possibly even counter-productive.

In the case of ammonia risk assessments, the client needs to be clear about the objective. Worker safety should be paramount and leads to a number of key requirements in training, managing and equipping staff. Comfort and convenience for the surrounding community is also a relevant objective but requires a different set of measures and is not really a safety concern; it is more about being a good neighbor. Measures that ensure neighbor comfort will not guarantee worker safety and vice versa.

I worked on this puzzle with Alex Pachai, chair of the IIR’s Safety Working Group and Kent Anderson, former president of IIAR, following a safety workshop held a year ago at the IIR’s Cold Chain conference. We hope to present the findings in a series of papers in Paris later this year. Meanwhile, ISO 817 is currently being revised and the anomaly about chronic toxicity mentioned above will hopefully be resolved.