A Lethal Concentration

The standard toxicity test for chemical compounds is called the LD50. LD stands for Lethal Dose and 50 indicates 50 percent. In other words, LD50 means the lowest dose at which a material kills half of the test subjects.

The results are usually given in milligrams of compound per kilograms of body weight. Many of these tests are conducted on laboratory rats. To give you a few rat results: the LD50 of table sugar (sucrose) is 29,700 mg/kg. For table salt (sodium chloride, NaCl) it's 3,000 mg/kg. Really poisonous substances, though, measure in the single digits: Sodium cyanide (NaCN), for instance, possesses an LD50 score of 6.4 mg/kg.

Basically, the lower the number, the deadlier the compound. Poisons in water and air are usually measured in lethal concentration rather than dose - in other words an LC50. Which got me wondering about the oil pouring into the Gulf of Mexico from BP's shattered oil rig. Not to mention the chemical dispersants being used in attempt to break down the spreading oil. What kind of lethal concentration might be building up in those waters?

The U.S. Environmental Protection Agency data on dispersants provide the LC50 in parts per million. Of course, these tests aren't done on rats but sea creatures, in this case Menidia, a small silvery fish that likes to hover near the water's edge and Mysidopsis, a tiny brine shrimp.

As has been earlier reported, Corexit, the compound chosen by BP, has some of the lowest LC50 numbers on the list, meaning that it's among the most poisonous. Also, it's among the least effective on Louisiana crude (the type flowing from the Deepwater break). Why the EPA went along with this choice remains a mystery to me - or maybe I just think the answer would depress me - but under public pressure the agency has now ordered BP to find an immediate alternative.

Nearly 700,000 gallons of Corexit have already been poured into gulf waters. But that pales, obviously, beside the amount of Louisana crude, now estimated at a minimum of 6 million gallons. So, I wondered, what is the LC50 of Louisiana crude on small salt water dwellers?

Of course, I realize, that comparing lethal concentrations is not straightforward. The results differ by species and by time as well as by amount of poison, The EPA numbers for Corexit 9500 (the formula used most heavily by BP) show that at 2.62 ppm, the dispersant kills half the silver fish in 96 hours/ four days. At a slightly higher concentration - 3.4 ppm - the compound kills half the little shrimp in two days.

As for crude oils, a very decent analysis by the American Petroleum Institute shows that all are toxic, but their effects vary with thickness and with the different chemistry seen in say, oil from the Gulf of Mexico and oil from Kuwait. The best estimate I've seen for South Louisiana Crude - after hours of exasperated research - comes from thesis work done at Louisiana State University several years ago. For instance, the study found that Louisiana crude had an LC50 of 4250 ppm for the warm-water loving killifish.

This suggests that crude oil is less acutely poisonous than chemical dispersants. But here's the really interesting finding in that terrific little study. Adding a dispersant - specifically Corexit 9500 - made the oil more poisonous. A lot more poisonous.

The "dispersed" oil had an LC50 of 317.7 ppm, making it more than 11 times more lethal in its effects. The study found a similar worsening for white shrimp, although not quite as dramatic. "Dispersed oils were more toxic than crude oils," noted the report.

Oh, definitely. Still, you might argue that this is only a master's thesis conclusion. But as it turns out there are plenty of other studies raising very similar warnings and they go back quite a ways. A report in the journal Environmental Toxicology a decade ago concluded that "LC50 values indicate that dispersed oil combinations were significantly more toxic to these organisms than .. crude oil." Another study, this time of snails and amphipods reached exactly the same conclusion.

To be fair, a study of the Australian octopus found no increased toxicity. But don't you wonder what we're doing out there in the fragile environment of the Gulf, whether we're reducing the spill damage or just turning the whole area into one ever-more poisonous bowl of toxic soup?

And don't you wish our officials gave any indication that they knew more about it than we do? I love doing this kind of research but in this case I'd much rather have our country's so-called regulators waving the LC50 red flag ahead of me.

More like this

In the past week, a lot of writers - and, yes, that includes me - have focused on the chemical dispersants being sprayed into the Gulf of Mexico to help manage the ever-expanding oil spill from BP's deepwater drilling rig. For instance, I recently pointed out while dispersants do help…
Thanks to fellow science blogger Ed Brayton for the link to this New York Times article, which suggests that because of ties to the company, BP chose to flood the Gulf with a dispersant that is both more toxic than many of the other options and also less effective. So far, BP has told federal…
Today, the U. S. Environmental Protection Agency announced that it would require BP to use a less toxic (and more effective) chemical dispersant than the brand used so far. I wish, I wish, I'm always wishing for these actions to sparkle with government intelligence and initiative. But it's…
The latest news from the Gulf of Mexico offers both relief (the "top kill" approach to ending the oil spill may be working) and dismay (the amount of oil pouring into the water is now thought to be closer to 20,000 barrels a day rather than the 5,000 barrels that BP has insisted on for weeks…

IME, three links is too many. This post with two, will try another with 3 in a moment.


I've heard some rumors that BP's lawyers are getting all Cuccinelli on scientists reporting on the gulf gusher.

To BP's lawyers, discrediting masters theses and the scientists who wrote them is just their duty to their shareholders.

That we're trying to fix an enormous chemical pollution problem by pumping 700,00 gallons of additional pollutants into the Gulf is staggering. For tiny spills they may have some utility, although I'd like to see the data first. But for this? It's like sending paper towels to New Orlean's 9th Ward after Katrina.

One point: "As has been earlier reported, Corexit, the compound chosen by BP, has some of the highest LC50 numbers on the list, meaning that it's among the most poisonous."

Shouldn't the most poisonous compounds have the lowest LC50 numbers, as with the LD50 numbers?

By Calming Influence (not verified) on 22 May 2010 #permalink

As far as I could see the thesis you linked to didn't try to control for oil-water surface area differences. Without some sort of mechanical dispersion of the oil to match the particle size created by the chemical dispersants how can the concentrations be directly compared? The portion of crude oil that will spread into the water column without dispersal of some sort is a small fraction of the total oil.

From the abstract of Another study

These LC50 and EC50 values indicated that dispersed oil combinations were significantly more toxic to these organisms than WAF of crude oil.

If you have access to the paper, can you find out if the concentrations were of water saturated with the WAF or the WAF itself? A LC50 of 19% (190,000 ppm again from the abstract) seems a bit high if it is referring to the hydrocarbons rather then saturated solution of hydrocarbons in water.

By Robert S. (not verified) on 22 May 2010 #permalink

A very interesting and useful post, but it would be good to fix the beginning of paragraph 5. After giving a nice explanation of LC50 and how lower numbers mean higher toxicity, you say that "Corexit has some of the highest LC50 numbers on the list". You meant to say that it has some of the LOWEST numbers, and hence highest toxicity, on the list.

By ecologist (not verified) on 23 May 2010 #permalink

Good point and I'll take another look at the paper. What's interesting to me is the consistency of the findings that dispersants do not do good things to oil. Here's a link - again an abstract but I'll try to do better - to a paper in which the authors actually recommend against using dispersants at all if the spill is near a coral reef because of the increased toxicity issue: http://bit.ly/anV66Q

If I understand correctly the dispersants are designed to break the oil into tiny droplets that will disperse throughout the water column, vastly increasing surface area, and allowing much faster degradation of the oil by microbes. This same effect would vastly increase the amount of interaction that an animal a few hundred feet under water would have with the oil, because without dispersant only the WAF would be in the water column with the rest of the oil at, or very near the surface. You can't increase the dispersal of the oil without increasing the amount of oil that interacts with everything in the water. There may very well be some nasty synergistic toxicity between the oil and Corexit, but I'd love to see some tests showing toxicity comparison between chemical dispersants and mechanical dispersion, say with a ultrasonic emulsifier.

Two links that may be interesting, first, how they prep a WAF sample the lab: http://video.google.com/videoplay?docid=-5344865627579685963#

Ultrasonic Emulsification /dispersion: http://www.youtube.com/watch?v=v8qHKwiBvhI

By Robert S. (not verified) on 23 May 2010 #permalink

Thanks for sending the links - they and you make a great point. Agreed that the toxicity data is only one part of the story. I've been fairly focused on Louisiana crude, since that's the oil in question here, but you may remember that the thesis I cited also looked at Alaskan crude. That obviously harks back to the Exxon Valdez spill and I'm now wondering if the kind of experiment that you're talking about might have been done there or if there might be some other interesting data that sheds some light here. Appreciate the shove in a good direction!

That's so odd. But I rechecked and it never showed up in the message queue. Do you want to test the theory - send it with two links? Or just try again. The Scienceblog folks never mentioned that as an issue to me.

well, this turns out to be more of an experiment than i could of wished. but try with two - we all know that one link works fine.

I found another paper on crude oil toxicity


This one also has a comparison with a known surfactant, sodium dodecyl sulfate (aka sodium lauryl sulfate). That is an extremely common surfactant used in just about every detergent, even in toothpaste. It has a LC50 comparable to the dispersant Corexit 9527.

here is the first post with one link

The acute toxicity is probably a generic property of strong surfactants. I found this nice paper which looks at a lot of different surfactants.


The increased toxicity associated with surfactant addition to crude oil comes from the dispersant activity. With the crude oil dispersed better in the water, there is greater access of the soluble hydrocarbons (which are toxic, especially the aromatics) to the water and so the hydrocarbon level dissolved in the water goes up, and so does the toxicity per dose of crude oil. The dispersed oil is more toxic, but is diluted faster and degrades faster. There is greater local toxicity, but there might be reduced toxicity over time. This is a very complicated thing to try and figure out.

My understanding is that the reason dispersants are used is so that the oil doesn't move as quickly to the surface and so doesn't move as quickly to shore. Because oil floats, there is a large tendency for it to collect at the water-air interface.

Yes, that's my understanding too, that it breaks the oil into smaller particles, which tend to settle, and which are more readily degraded by microbes. But I like your description of what happens. it makes a lot of sense in explaining what's going on. And all of this - including the paper you just sent over - reminds me that I should at some time do a blog on the topic of surfactants. Meanwhile, we seem to be caught in this zero-sum game that I don't think any of us are sure how to play out. But, I figure, the more we know the better.

Here is a reasonable discussion of a number of past oil spills.


My guess is that using oil dispersants off shore is probably a good thing, because it keeps the lighter and more toxic hydrocarbons off shore, even if they are in the water column. The heavier stuff will still float to the surface but without the lighter stuff it is more benign. I think once the oil is onshore or near shore, then dispersants are a bad thing and should not be used. Gathering the oil up in things like straw is probably the best thing to do once it hits shore.

I finally got around to checking this blog out, and this post is wonderful.

I think you're using LC50 numbers for dispersed oil with Corexit 9500 instead of the dispersant alone. Corexit 9500 alone has LC50s of 25.2 and 32.23 ppm for the silver fish and shrimp, respectively. The lower LC50s stem from experiments where they monitor the critters' response to water with oil and dispersant mixed in. And those numbers are about 10x lower.

That's average for modern dispersants. A 2005 National Research Council report (http://books.nap.edu/openbook.php?record_id=11283&page=207), they list dispersant LC50 ranges as 190-500 mg/L and dispersed oil LC50 ranges as 20-50 mg/L (why they don't use ppm, I don't know).

You're right. I went back and looked at that EPA data set, which is in ppm, and it's for a mix for Corexit and No. 2 Fuel OIl. Comparing the LC50s seems to reinforce the other studies I found which show that dispersant makes fuel oil up to ten times more poisonous. In other words, a very consistent finding for shrimp and silver fish species but not necessarily good news. I think I'll add that into the next blog post, which will be about crude oil.

Every membrane of every cell of every living thing is composed of lipids, so the toxicity of surfactants is no surprise. Plus, crude kills mostly by smothering, not toxicity.

The "data" on toxicity gives the impression of false precision. The adverse health effects are complex over time, space, species, tissue compartments, concentration and with multiple interactions between variable components and cannot be reduced to a single number even for a single species at a single life stage. All of these effects are non-linear and coupled which makes modeling them essentially impossible.

Roland is absolutely correct, every cell has a lipid membrane. Every lipid membrane has proteins that are partly in that membrane and partly in the cytoplasm and external to the cell. Hydrocarbons partition into that lipid membrane and change its properties. The proteins in the lipid membranes are there because there are hydrophobic regions in the proteins that are âattractedâ to the lipid membrane more strongly than they are attracted to the aqueous phase. When hydrocarbons get into the lipid membrane, they change the properties of the membrane and change the interactions of the membrane with the proteins in it. If that change exceeds a certain amount, then the proteins won't work properly. If enough proteins don't work properly, then the organism exhibits metabolic stress, physiological damage, and eventually death.

I don't think that differences in the âtoxicity numberâ by even an order of magnitude are necessarily important or significant. A lot of these âtoxicity numbersâ are going to be quite sensitive to the ratio of dispersant to oil, and to the detailed composition of that crude oil and to the idiosyncratic conditions of the test (O2, temperature, species, nutritional state, test duration, etc.) The toxicity effects of dispersant on diesel fuel and on this crude oil is expected to not be identical; the details which are unknown will have larger effects than the stuff that is known.

Getting rid of the insoluble but visible oil water emulsion on the surface may slow the degradation of the more soluble but invisible aromatic compounds which are more toxic by removing bacteria that can degrade them.

If you want to discuss âeffectivenessâ of clean-up, you really need to define what you mean and over what time frame; and appreciate that what ever you define âeffectiveâ to be will mean that the relative effectiveness of treatments will change as your definitions and time frames change. In many cases the order of ranking of methods will change.

One also has to consider the effects over time.
If the dispersal agent causes the oil to break down (through natural processes) much more quickly into non-toxic compounds (or less toxic), the toxic effects will be contained in a smaller area than would otherwise be the case, as would the smothering effect that causes marine animals (as well as marine birds and animals in lithoral zones) to be killed by suffocation or starvation by being unable to fly/swim/hunt/eat.

Thus while in a small area there might be an increased dieoff due to toxic materials (probably partially compensated by a smaller dieoff due to smothering, but that's nitpicking) the effects would be more contained.
Whether that's ever been studied I don't know. I do know similar studies have been done on nuclear waste dumping that showed conclusively that leakage would have a far smaller effect than had been assumed prior (effectively, dispersal of the nuclear waste material by sea currents would be so effective that even at relatively short distances from the spillage site there would be no adverse effects on the biosphere, the same may well be true if spilled oil were well dispersed).
I don't know if this study was ever published internationally (let alone online), I do know we had big trouble getting it published at all because of political pressure (this was in the mid 1990s, when the anti-nuclear movement was even stronger than it is now and (like the CO2/global warming fanatics now) didn't want anything published that would not support their ideas.

"Why the EPA went along with this choice remains a mystery to me - or maybe I just think the answer would depress me"

My understanding is that the EPA did not go along with it, and BP simply ignored them. De-list 'em and nationalize their assets.

I am wondering how Corexit interacts as moisture evaopration takes place after it has been used as a dispersant? oceanic circulation is not the only concern--though a big one, certainly! Air circulation and weather patterns, impacted such as they are by oceanic temperatures and currents, is another matter, altogether and I am curious if there are any studies as to the effects?

P.S. Does anyone know how much methane is being released or was released along with the crude that continues to spew? And, how might this add to the complexity of chemical reactions and maginified toxicity of crude and Corexit in the seawater?

Methane is probably the most benign of the oil constituents after CO2. Methane isn't toxic except at very high levels where it is probably narcotic, but those are very high levels. It is metabolized by bacteria into CO2 and biomass and that consumes O2, or (worse) generates H2S from sulfate. H2S is toxic to just about every multicellular life form. It is only an acute toxin, if the organism survives, it probably will do ok.

What would be considered a high level? Thanks for the info--I am also concerned about the amount of methane being released into the Gulf as that additional amount applies to global warming from greenhouse gases.

Also, I was recently reading an article by Peter Ward in Earth Science, regarding a hydrogen sulfide mass extincion theory postulated for the some of the mass extinctions not explained by asteroid theory. He cites the work of a couple of researchers, that formed the background for my initial question on methane. I quuote here--

"Calculations by geoscientists Lee R. Kump and Michael A. Arthur of Pennsylvania State University have shown that if oxygen levels drop in the oceans, conditions begin to favor the deep-sea anaerobic bacteria, which proliferate and produce greater amounts of hydrogen sulfide. In their models, if the deepwater H2S concentrations were to increase beyond a critical threshold during such an interval of oceanic anoxia, then the chemocline separating the H2S-rich deepwater from oxygenated surface water could have floated up to the top abruptly. The horrific result would be great bubbles of toxic H2S gas erupting into the atmosphere."

Now... first of all, since oxygen is present in oceans today in essentially equal concentrations from top to bottom because it dissolves from the atmosphere into the water and is carried downward by ocean circulation, only under unusual circumstances would anoxic conditions below the surface permit a wide variety of oxygen-hating organisms to thrive in the water column.
Now, would you say this is one of those unusual circumstances, or not?

Touched, I would not think that these circumstances are unusual. This oil spill is not going to change the ocean circulation.

What could change the ocean circulation would be global warming. Particularly when Greenland melts, the fresh water on top of the salt water will prevent the cold salt water from sinking at the North pole. That will stop the ocean circulation. Without cold water from the Arctic sinking, the ocean depths will heat up which could destabilize the methane hydrates there.

You compare the LCs of oil and dispersed oil, showing that the LC of dispersed oil is about 11 times lower. This means that dispersed oil is toxic at one eleventh of the concentration of oil.

However, I assume that dispersed oil is more dispersed than oil. The question is whether it is more than 11 times more dispersed.

I would find it hard to believe that in a large supply of water (eg the Gulf) dispersed oil would be many more than 11 times more dispersed than oil.

So your conclusion should be that adding even a highly toxic dispersal agent lowers the overall toxicity as the dispersal overcomes the increased toxicity.

What it did do is to make the oil more difficult (impossible) to see with the naked eye which presents a huge problem for desalinization plants (people's health) and nuclear power plants that can't take in oil. All by design in my humble opinion. My question is what happens when corexit mixes with frozen methane (in the case of a collapse of the seabed). Is that explosive mix of chemicals?

How does one detect the dispersants in their water? Considering its inevitable that it has or will reach us from the clouds and rained on most of North America by now how do we test for this dispersant in our water? I would like to check my tap and see if I need to setup water distillation to purify my water, will distillation process be able to exclude corexit from the water?

The components of Corexit 9500 are not as scary as the term "chemical dispersant" might suggest. Along with some light gasoline, it contains SPAN and TWEEN emulsifiers commonly found in foodstuffs, a common laxative, propanediol (a less toxic relative of antifreeze), and a hand cleaner that has a not-so-bad LC50 of 841. I'm surprised that its LC50 is as low as 25. Perhaps dispersing these materials in water for testing makes them more toxic than eating them does. Or perhaps any detergent-like compounds are tough on aquatic organisms.