Fixing on the nitrogen fixation problem

Haber Bosch fixes nitrogen.  The process is considered by many to be one of the most important technological advances of the 20th century.  Indeed, it is estimated that about a third of the people born since 1909 have been sustained by fertilizer produced by the Haber Bosch process (Erisman et al., Nature, 2008).  It must also be said that the process has been responsible for the death of millions as well, because it also fixes nitrogen for use in explosives, but it is its use in agriculture that we are focusing on today.

Nitrogen is an essential component of all living cells.  It is crucial to metabolic processes, to growth and development, and in plants it is a component of chlorophyll, so it's essential for photosynthesis as well as seed and fruit production.  But nitrogen can be hard to come by if you're a plant.  It's a major constituent of our atmosphere – 78% -- but plants can't use it in its atmospheric form because as such it is chemically inert, an unreactive molecule that must be primed for further chemical reactions.  The bonds between the two nitrogen atoms (N₂) in atmospheric nitrogen are the strongest bonds known, in fact, so it must be 'fixed' by some natural or man-made process, the bonds broken, and the nitrogen otherwise made ‘reactive, or available to plants.*

Fritz Haber, 1918

Plants can use nitrogen from either nitrate (NO₃^-) or ammonium ions (NH₄⁺).  They absorb it through their roots, but neither form is abundant in soil.  Fritz Haber developed a process to 'fix' nitrogen about a century ago, to split the nitrogen bonds and prime the molecules for further chemical reactions, linking them to hydrogen to make ammonium or to oxygen to make nitrates, and thus make nitrogen bioavailable.  Now known as the Haber Bosch process, because Carl Bosch later scaled it up for commercial use, agriculture has been dependent on it ever since.

There are up sides and down sides to this.  Haber Bosch converts atmospheric N's into ammonia molecules at high temperature and pressure.  Wikipedia tells us that "This conversion is typically conducted at 15–25 MPa (2,200–3,600 psi) or 150–250 bar and between 300–550 °C (572–1,022 °F)…"  As such, it is not an energy efficient procedure.  Indeed, just this one chemical reaction uses 1-2% of the world's energy supply.  And according to a recent BBC radio program hosted by UCL chemistry professor Andrea Sella about the Haber Bosch process, only 20% of the nitrogen fixed in this way goes into final products; the bulk gets lost in the environment.

The nitrogen cascade, from an EPA report on reactive nitrogen

The Haber Bosch process now produces about 100 million tons of ammonia a year, according to the same BBC program, which then is used to make 454 million tons of nitrogen fertilizer, usually in the form of anhydrous ammonia, ammonium nitrate, or urea.  The Green Revolution of the 1960's brought fertilizer-dependent agriculture to much of the world, and along with it, a more productive way to feed the world, but there have been consequences. As Aneja et al. put it in a Nature Geosciences article in 2008 ("Farm Pollution", paywalled):

Nitrogen emissions in various forms (nitrogen oxides (NOx), nitrous oxide (N2O), ammonia (NH3), and organic nitrogen (Norg)) are one of the two main classes of pollutants that are emitted by modern agriculture. Although produced naturally in soils through microbial denitrification and nitrification processes, nitrous oxide —a greenhouse gas that is much more effective than carbon dioxide in trapping heat in the atmosphere — arises from animal production in large quantities, depending on the nitrogen input and management of manure. In order to increase yields, agricultural operations often directly add reactive nitrogen to soils, either through the application of fertilizer or livestock manure to fields, or by growing nitrogen-fixing crops. These measures increase nitrous oxide emissions via microbial reactions, especially enhanced nitrification. Indirect additions of reactive nitrogen exacerbate the problem. For example, nitrogen from fertilizer or manure volatilizes as ammonia and oxides of nitrogen are redeposited in downwind regions as ammonia, particulate ammonium, nitric acid and nitrate.

Haber Bosch has led directly to water and air pollution, reduced biodiversity, acidification of the soil, ocean dead zones, negative effects on human health, and global warming.  So, while biological fixation of nitrogen also has consequences, people are beginning to think about replacements for the Haber Bosch process primarily because of its environmental effects.  There are two angles from which to approach this -- one is to engineer plants to fix nitrogen themselves, and the other is to make the mechanics of fixing nitrogen more efficient.  

In theory, it's possible to produce ammonia at low pressure and temperatures, unlike the Haber Bosch process.  The trick is to push electrons into the nitrogen molecule, N₂, to weaken the bonds, priming it to accept additional protons, positively charged hydrogen, to eventually produce ammonia.  This is being tried in the lab, but there are a lot of ways this can and does go wrong, so that the hydrogen leaks rather than makes ammonia and the process is decades away from being commercially viable.

And even if it does become viable, it still means farmers will be spreading tons of synthetic fertilizer on their fields, still polluting soil, water and air.  So another approach is to engineer plants to fix their own nitrogen, as legumes -- peas and beans -- do naturally with the aid of symbiotic bacteria that produce an enzyme called nitrogenase.

The interaction between legumes and these nitrogen-fixing bacteria happens in nodules that the plant forms on its roots.  These are essentially nitrogen organs, and they fill with bacteria that can break the nitrogen bonds of atmospheric N₂ so that it becomes available to the plant.  It's not a one-way deal, though, as the plants are a source of carbon for the bacteria.

Most major cash crops don't have the ability to interact with nitrogen-fixing bacteria, though, which is why, given the widespread practice of monocropping, they need to be intensively fertilized.  But, the genetic architecture by which legumes interact with bacteria is understood and the pathway is already present in grasses, just doing something else.  So plant breeders are trying to coax the genes to recognize signals from bacteria so that these plants can create their own interactions with nitrogen fixing bacteria.

The process in legumes, which uses rhizobial bacteria, happens in the absence of oxygen, however, and an anaerobic setting can't always be arranged.  But there are bacteria that can fix nitrogen in the presence of oxygen.  Cyanobacteria do it, for example.  Gunnera tinctoria, or giant rhubarb, has a symbiotic relationship with nitrogen-fixing cyanobacteria, and they are also photosynthetic, producing oxygen.  They are actually inside the cells of the plant, meaning that it's possible that a plant wouldn't need the root nodules that legumes use in their nitrogen fixing process.

Gunnera tinctoria: Wikipedia

Sugar cane, too, fixes nitrogen, either with or without oxygen.  It doesn’t exploit the same rhizobia that legumes do, nor does it use cyanobacteria.  Instead, it interacts with Glucoacetobacter diazotrophicus, an organism that lives in intercellular spaces in the stem of the plant.  Other plant/N-fixing bacterial relationships are still being found, and plant breeders are experimenting with coating seeds with bacteria as a way to introduce the symbiotic relationship between plants and bacteria.

Would this be efficient?  That is, can it replace a significant amount of the inefficient fertilizer produced by the Haber Bosch process?  The same BBC radio program we mention above suggests that it should be possible to replace at least 20% of it, and that plants would still give good crop yields even though there's an energy cost to fixing nitrogen.  Twenty percent is not insignificant, given the total amount of fertilizer used in the world today, and given that demand for it is rising in China and India, as they try to  feed their ever-growing populations.  But, it would still leave us heavily dependent on Haber Bosch.

The topic of nitrogen fixation was raised recently in an exchange of emails among a group of agronomists, farmers, ag economists, geneticists and interested onlookers including me.  The question of whether engineering corn to fix nitrogen would be the path to lower reliance on synthetic fertilizer was summarily dismissed as silly.  Kendall Lamkey, chair of Agronomy at Iowa State, did the numbers.  He pointed out that soybean is a legume.  It fixes nitrogen, and yet:

In the corn-soybean system in Iowa, soybean (the legume) by far has the largest total N needs, particularly at high yield levels.  Total N needs are dependent primarily on seed protein content and yield.  ...a 65 bushel/acre soybean crop has a seed N need of 240 lbs/acre. Soybean has a N harvest index (the proportion of the total above ground N that is in the grain) of about 0.7, so the total N requirement of a 65 bushel/acre soybean crop is about 350 lbs/acre.  Soybeans depend primarily on soil N until about the start of pod-fill when they switchover to depending primarily on N-fixation. We talk about soybean fixing about 50% of its total N needs on the average, but the reality is that the range is large and dependent on many factors, but with our levels of soil nitrate and yield, 50% fixation captures a lot of the acres. That means a 65 bushel soybean crop needs about 175 lb N/acre from the soil supply. 

Kendall went on to say that a recent record soybean yield in Iowa (160 bu/acre) he bets required on the order of 600 lbs of nitrogen per acre.  However,

A 200 bushel/acre corn crop tops out at about 178 lbs/acre at physiological maturity.  The N-harvest index is around 0.64 as well, so 114 lbs/acre of the 178 is in the grain.

So, a soybean harvest at these yield levels removes much more N from field than a corn harvest and both require about the same amount of soil N, which comes from fertilizer and manure inputs and soil organic matter decomposition.

If you measure change in soil N content over time in the corn-soybean rotation, you will find there is a net loss of N from the system, which translates into a net loss of soil organic matter.  If you do the same thing with a corn-corn rotation, you will find a net increase of N in the system.  This difference is driven primarily by the low crop residue inputs in soybean compared to corn. 

Unless we change the cropping system, N fixation in corn is not a cure.  Neither is reduced fertilizer N inputs.  It would help, but not as much as you might think, unless it caused farmers to change their cropping system.  Changing the cropping system is what caused the problem in the first place. 

Before 1913, when fertilizer made by Haber Bosch became commercially available, plants got their nitrogen naturally.  Farmers rotated crops, planting nitrogen-fixing legumes in fields where they also grew cereals, they applied manure from animals that had eaten nitrogen-rich feed, and/or the composted nitrogen-rich remains of other crops. Where available, farmers used guano as fertilizer, shipped far and wide from Pacific Islands, or nitrates from South America, but by the turn of the 20^th century, these sources of reactive nitrogen were not enough to feed the growing world population. 

Cropping practices have changed dramatically in the last century, largely because farmers can rely on synthetic fertilizer to enhance their crop yields.  A sharp decrease in the diversity of crops planted in the American midwest occurred over that 50 or 60 years, and now corn and soybeans are the predominant crops there, with heavy reliance on synthetic fertilizer.  Haber Bosch has changed the way agriculture is done, with all its consequences.  

And, Haber Bosch has also thus indirectly enabled dramatic population growth, making possible an increase in the number of people on earth from close to 2 billion in 1900 to 7 1/2 billion today.  Erisman et al. write that "...the number of humans supported per hectare of arable land has increased from 1.9 to 4.3 persons between 1908 and 2008", and they estimate that probably half of us wouldn't be alive today without Haber Bosch.

Of the total world population (solid line), an estimate is made of the number of people that could be sustained without reactive nitrogen from the Haber–Bosch process (long dashed line), also expressed as a percentage of the global population (short dashed line). The recorded increase in average fertilizer use per hectare of agricultural land (blue symbols) and the increase in per capita meat production (green symbols) is also shown. (Erisman, Nature Geoscience, 2009)

Would growing food that fixed nitrogen itself enable further population growth?  Not if what it would change is the way that nitrogen is made available to plants, not the amount of food that can be grown.  If plants can be engineered to fix nitrogen, this could have beneficial environmental effects, although biologically fixed nitrogen is also a pollution source.  But, if we’re still spreading as much reactive nitrogen on fields as we do now, regardless of how it’s made, this can’t attenuate the environmental effects.

So, perhaps a law of nature that we must face is that every solution raises its own problems.


----------------------------

*Reactive nitrogen (Nr) includes inorganic chemically reduced forms of N (NHx) [e.g., ammonia (NH3) and ammonium ion (NH4+)], inorganic chemically oxidized forms of N [e.g., nitrogen oxides (NOx), nitric acid (HNO3), nitrous oxide (N2O), N2O5, HONO, peroxy acetyl compounds such as peroxyacytyl nitrate (PAN), and nitrate ion (N3-)], as well as organic compounds (e.g., urea, amines, amino acids, and proteins).  From a 2011 EPA report, “Reactive Nitrogen in the United States: An Analysis of Inputs, Flows, Consequences, and Management Options”. 

This post was inspired and enhanced by a recent discussion among agronomists, ag economists and geneticists, and comments, and improved with assistance from Matt Liebman.  Any remaining errors are my own.