Soil is a huge component of the global carbon cycle. As shown in figure 1, the world’s soils contain more carbon than the atmosphere and all living things combined, and the flux of carbon into and out of the soil dwarfs the rate of anthropogenic carbon emission from fossil fuels.
Figure 1. Global carbon pools, in gigatons, and the major fluxes between them. (Adapted from ref. 1.)
Figure 1. Global carbon pools, in gigatons, and the major fluxes between them. (Adapted from ref. 1.)
About two-thirds of the soil’s carbon is contained in organic molecules. (The rest forms inorganic minerals such as carbonates.) All organic molecules are thermodynamically unstable; left to seek their free-energy minima, they decompose into carbon dioxide, water, and other small molecules. Despite that instability, organic matter entering the soil remains there for an average of 25 years, and can even stay for centuries or millennia without decomposing.
An accurate treatment of that capacity to store and stabilize carbon is an essential ingredient in climate models. At present, soil erosion and changes in land use each release an additional gigaton or so of carbon into the atmosphere per year.1 Better soil management could slow that trend or even reverse it, so that natural carbon sequestration in the soil might offset a portion of fossil-fuel emissions.2
The problem is that we don’t really know how the stabilization happens. It’s known that interactions between organic matter and mineral grains are responsible in some way. But soils are complicated systems—composed of many different minerals, organic compounds, liquids, gases, and microbes—and their composition and dynamics vary from place to place. Models have long assumed that all mineral surfaces are equally good at stabilizing carbon, so a soil’s carbon-storage capacity is determined by its total surface area—or, equivalently, its clay content. (Soil mineral particles are classified by size: micron-sized clay, larger silt, and even larger sand.) Evidence has been accumulating that the approximation is overly simplistic: Organic material doesn’t cover soil mineral surfaces evenly, and clay content isn’t always well correlated with stored carbon.3 But a better description has been hard to come by.
Now Ingrid Kögel-Knabner, Cordula Vogel, and colleagues at the Technical University of Munich have taken a microscopic approach.4 Using nanoscale secondary-ion mass spectrometry (nanoSIMS), a technique borrowed from materials science, they’ve shown conclusively that organic matter binds to just a fraction of total surface area. They also found that the rough surfaces of clustered mineral particles bind far more carbon than the smooth surfaces of single grains and that new organic matter binds almost exclusively to surfaces where old organic matter is already present.
Dishing the dirt
NanoSIMS is a method for making spatially resolved maps of chemical identity. An ion beam, focused to about 100 nm, is shot at a surface. The primary ions from the beam dislodge secondary ions from the surface, which are then collected with ion optics and sorted with a mass spectrometer. Screening for the mass of a particular chemical species, such as carbon-12, gives a map of where that species is located on the surface. Available instruments allow the imaging of up to seven masses simultaneously—a useful capability, because imaging with nanoSIMS destroys the sample.
Imaging the clay-sized particles from a soil sample showed that organic carbon was confined to about 20% of the mineral surface area. Comparing nanoSIMS images with scanning electron microscope images, which allow mineral clusters to be distinguished from single clay particles by visual inspection, showed that organic matter was almost never present on single particles.
But Kögel-Knabner and colleagues wanted to go beyond that static picture to look at the process by which new organic matter—in their case, litter made from leaves of corn and potato plants—is incorporated into the soil. They used an established soil-research technique of tracking isotopically labeled litter, made from plants grown in a closed environment and fed with CO2 enriched in 13C and fertilizer enriched in nitrogen-15. That’s expensive, but it works: The resulting plants have 13C and 15N concentrations several times higher than those found in nature. Isotopic analysis can then reveal the fate of the leaf material as it interacts with the soil.
Some isotopic tracking experiments are performed in situ and last for years. But the Munich researchers did theirs in a lab over just six weeks. They combined 50 g of natural soil with 0.5 g of the isotopically labeled litter and incubated the mixture under controlled temperature and moisture conditions.
Control systems, some with unlabeled litter and some with no litter, were incubated in parallel. The researchers sampled and analyzed the soil after two hours; one day; and one, three, and six weeks. They separated the clay-sized particles of interest from free organic matter and other mineral components by size and density fractionation. Then they took a drop of a dilute suspension of clay particles and let it dry on a silica wafer for nanoSIMS imaging.
Oxygen-16 images showed the outlines of the clay particles, and 12C and 12C14N, as shown in figure 2a, revealed the overall distribution of organic matter. To see the newly incorporated organic material, the researchers looked for regions with unusually high fractions of 13C or 12C15N, as shown in figure 2b. (The mass spectrometer is sensitive enough to distinguish between 12C15N and 13C14N.)
Figure 2. Nanoscale secondary-ion mass spectrometry images of clay-sized soil particles after three weeks of incubation with isotopically labeled plant litter. The oxygen-16 signal in both images shows the outlines of the mineral particles. (a) Organic matter, indicated by 12C and 12C14N signals, covers about 20% of the total mineral surface. That fraction remained roughly constant over the course of the six-week experiment. (b) Newly incorporated organic matter, indicated by enrichment in 13C and 15N, covers a smaller but growing fraction of the area. (Adapted from ref. 4.)
Figure 2. Nanoscale secondary-ion mass spectrometry images of clay-sized soil particles after three weeks of incubation with isotopically labeled plant litter. The oxygen-16 signal in both images shows the outlines of the mineral particles. (a) Organic matter, indicated by 12C and 12C14N signals, covers about 20% of the total mineral surface. That fraction remained roughly constant over the course of the six-week experiment. (b) Newly incorporated organic matter, indicated by enrichment in 13C and 15N, covers a smaller but growing fraction of the area. (Adapted from ref. 4.)
Images taken at different times necessarily showed different mineral surfaces. Even if nanoSIMS were a nondestructive technique, it would be impossible to retrieve the same clay particles from the incubator more than once. So tracking the labeled litter’s progress in binding to the clay had to be done statistically. For each nanoSIMS image, the researchers determined the fraction of the total mineral area that contained 12C and the fraction of the organic area enriched in 13C. The former fraction did not change over the course of the six weeks, but the latter increased: At two hours, one-third of the organic area was isotopically enriched; at six weeks, more than half of it was. Clearly, the new organic matter was attaching to the clay, but it wasn’t seeking out new mineral surfaces to bind to.
What makes rough mineral surfaces more hospitable to organic matter than smooth ones, and what makes some rough surfaces better than others? Kögel-Knabner and colleagues attribute the difference to microbial activity. The submicron nooks and crannies of mineral clusters make good homes for single-celled organisms, but not every suitable surface is populated by microbes.
So far, the researchers’ nanoSIMS work has focused on just one type of soil—a topsoil taken from near their home in Germany and typical of soils in central Europe, the US, and parts of Australia and Asia. As a next step, they plan to extend their analysis to other types of soil and to study the effect of the soil particles’ composition in addition to their size and shape. They hope that that work will paint a clearer picture of how much carbon soil can hold and how best to exploit that capacity.
