An
approach for nanoscale 3-D tomography of organic carbon (OC) and associated
mineral nanoparticles was developed to illustrate their spatial distribution
and boundary interplay, using synchrotron-based transmission X-ray microscopy
(TXM). The proposed 3-D tomography technique was first applied to in situ
observation of a laboratory-made consortium of black carbon (BC) and nanomineral
(TiO
Three main mechanisms for soil organic carbon (SOC) stabilization have been
proposed: (1) chemical stabilization as a result of chemical or
physiochemical binding between SOC and soil minerals (especially clay and
silt in current opinions), namely organo-mineral complexation; (2) physical
protection, which occurs predominantly at the microaggregate level and is
built on top of the chemical organo-mineral complexation; and (3) biochemical
stabilization in the form of recalcitrant SOC compounds (Six et al., 2002).
The accumulation and subsequent loss of OC have been found to be largely driven by
changes in the millennial-scale cycling of mineral-stabilized C, and a
positive correlation between non-crystalline minerals and OC has been found
in soils across a climate gradient (Torn et al., 1997). Metastable SRO
minerals, which may only exist in small portions in the fine and dense soil
fraction, are crucial for C stabilization (Cusack et al., 2012; Eusterhues et
al., 2005; Kaiser et al., 2002a; Mikutta et al., 2005, 2006, 2007; Rasmussen
et al., 2005). Even though SRO minerals are known for high reactivity
(Eusterhues et al., 2008, 2011), their development of
strong organo-mineral complexation with OC relies on immediate contact with
the reactive surface of OC. In contrast, although SRO normally accounts
for only a small portion of weight in the soil fraction, their distribution
could be prevalent due to small size. In real soil environments, the actual
key mineral players (crystalline micron-sized clay vs. nanoscale SRO minerals)
for in situ interaction with OC have not been identified, and their spatial
and temporal variations are not known (Vogel et al., 2014). To date, little
information is available on the in situ distribution of reactive minerals
below clay size, and on the subsequent organo-mineral micro-assemblage in
soils (Baldock and Skjemstad, 2000; Cusack et al., 2012; Mikutta et al.,
2006; Torn et al., 1997; Vogel et al., 2014). Traditional fractionation
methods based on size and external force for dissecting the association
strength between OC and minerals in soils are limited to bulk samples and
clay-sized mineral particles (Kaiser et al., 2002b; Kleber et al., 2007;
Sollins et al., 2009). Multi-dimensional perspectives call for in situ
observation of organo-mineral interplay and micro-assemblage at nanoscale
(Kinyangi et al., 2006; Lehmann et al., 2007, 2008; Solomon et al., 2012).
Advance of in situ interfacial details may lead to a breakthrough in mineral
physical protection mechanism for the long-term stabilization of OC. To
overcome the limitations of routinely used electron microscopic methods
(e.g.,
applicability limited to the surface layer or undesirable artifacts due to
pretreatments), a non-destructive high-resolution X-ray 3-D tomographic
technique is used in this study to explore the fine structure of OC and
boundary interplay with nano-sized mineral particles. It is worth noting that
while some 3-D tomography studies have been conducted on soil microstructure
and porosity using X-ray micro-computed tomography (CT), the best resolution
achieved only amounts to tens of microns (for example, Quin et al. (2014),
down to 70
The synchrotron-based TXM at the beamline BL01B1 of Taiwan Light Source
(TLS), which has been used in this study, provides two-dimensional imaging
and three-dimensional tomography at a spatial resolution of
High-resolution X-ray scanning and 3-D tomography is highly demanding in terms of technology, and multidisciplinarity and “big data” analysis (Lafond et al., 2015). Except for the adoption of high-resolution X-ray objective lenses, the fidelity of 3-D tomography relies on the accurate alignment of the 2-D projections in correct three-dimensional positions. However, non-negligible mechanical imperfection of the rotational stages at nanometer level, or the thermal effects may significantly degrade the spatial resolution of reconstructed tomography. We have developed a markless image auto-alignment algorithm for fast projection matching (Faproma, Wang et al., 2017) to overcome these challenges, and accomplished accurate reconstruction of 3-D tomography at the nanometer level.
Sample of BC was made in the laboratory using a leguminous plant (
Thin section of natural OC and mineral consortium (NH) was prepared using
micron- to millimeter-sized particulate sample from mountain soil.
Millimeter-sized particulate organic matter with minerals embedded inside was
taken from the lower dark layer at a depth of 72 to 93 cm in a “Typic
Humicryept” soil profile, located in Mt. Nanhua, Nantou County, Taiwan
(24
A superconducting wavelength shifter source provides a photon flux of
4
Under the most frequently used absorption contrast mode, 2-D images are recorded based on the projection of the different X-ray absorption coefficient integration along the optical pathway through the samples on a detector. The absorption mode is useful for materials of high absorption coefficient, such as minerals or high atomic number materials, but it performs poorly for the observation of low atomic number materials, such as organic or polymer materials. To recognize the OC structure more accurately, 2-D/3-D images for the same sample region are recorded using absorption contrast and phase contrast modes, respectively.
In the phase contrast mode, the gold-made phase ring positioned at the
back focal plane of the zone plate is used to retard or advance the phase of
the non-diffracted light by
Three-dimensional tomography reconstruction was performed using homemade
software, which was coded based on iterative image registration (Faproma)
(Wang et al., 2017) and filtered back projection (FBP) reconstruction
algorithms. Firstly, a series of single TXM images captured from
For correlated spatial distribution of selected elements (C, O, Fe, Al, etc.) in natural OC particles from mountain soil, a low-vacuum scanning electron microscope (SEM, JEOL W-LVSEM, JSM-6360LV) equipped with an energy dispersive X-ray spectrometer (Oxford EDS) and a cathodoluminescence (CL) image detector (Gatan mini-CL) was used for elemental mapping, at an accelerating voltage of 15 keV.
To analyze the forms of minerals associated with natural OC, particulate OC (with minerals on the surface and embedded inside) was ground and injected into capillary tubes (Special Glass 10, Hampton Research, CA) for synchrotron high-resolution X-ray diffraction analysis. Analysis took place at the 09A beamline at Taiwan Photon Source (TPS), which was equipped with a set of high-resolution monochromators (HRM). The wavelength was 0.8266 Å at an energy of 15 keV. The X-ray diffraction (XRD) spectra were recorded under room temperature for 240 s accumulation time. Specific X-ray diffraction peaks and patterns were assigned using the ICDD PDF-2/4 databases.
The 2-D X-ray images for BC and the mineral nanoparticle consortium. Two
images are taken of the same region using absorption contrast mode
For the FTIR analysis, mineral-bearing OC (NH) particles were ground, dried
(60
Three-directional orthogonal sections of laboratory-made BC and mineral
nanoparticle consortium. The upper row sections were extracted from
absorption contrast tomography
High-resolution 2-D X-ray photographs were captured for the identical regions
in laboratory-made BC and nanomineral consortium using dual-scan absorption
contrast and phase contrast modes (Fig. 1a and e). The cross-section views
exported from the reconstructed 3-D datasets reveal subtle details of BC and
mineral nanoparticles, and clearly outline the fine boundary of BC and the
distribution of TiO
3-D tomography illustration of laboratory-made BC and the mineral nanoparticle
consortium observed at
Cross-sectional views of the reconstructed 3-D tomography shared consistent
and comparable features of BC and nanominerals in multi-angles (Fig. 2).
According to the display of different slicing planes (
Three-dimensional tomography for visualization was computed and generated
based on the post-processing of the reconstructed 3-D datasets to illustrate
the spatial correlation between BC and minerals. Unprecedented details of 3-D
in situ distribution of BC and mineral nanoparticles were revealed in the
computed 3-D tomography (Fig. 3; Fig. SMOV1
Three-directional orthogonal sections of natural mineral-bearing OC
from absorption contrast tomography (
The laboratory-made consortium was successfully tested by the dual-scan methodology using both absorption contrast and phase contrast acquisition modes (Figs. 1, 2, and 3). Compared with BC made at high temperatures, low-temperature BC is more similar to pyrogenic OC exposed to the natural environment. Thus, low-temperature BC was specially made to test its applicability under absorption contrast mode. Results showed that the fine structure and boundary of low-temperature BC can be clearly observed under absorption contrast mode. Thus, for environmental OC samples, the use of absorption contrast mode is very likely sufficient for capturing organo-mineral features.
Unlike field samples, the minerals observed within the laboratory-made consortium are often distributed in clusters and are only sparsely in contact with BC surfaces. The preservation of plant-like structures in BC could play a role in carbon stabilization in the natural environment, as their porosity and reactive surface provide large areas and sites for mineral coating, which may contribute to their long residence and physical endurance (Eusterhues et al., 2008; Rasmussen et al., 2005; Rawal et al., 2016).
The X-ray diffraction pattern of minerals within OC particles from mountain soil. Highly reactive Fe oxyhydroxides are identified and
denoted with lines of different colors: ferrihydrite (ICDD 01-073-8408,
orange), goethite (ICDD 01-073-6522, blue), and lepidocrocite (ICDD
00-044-1415, green).
The nanoscale 3-D tomography in this study revealed a high heterogeneity
within the natural OC–mineral consortium, and most of the particulate OC
surface was coated by minerals. Natural OC exhibited strong organo-mineral
association on its surface at the nanoscale (Fig. 4; Fig. SMOV3
Elemental mapping by SEM-EDS for the mineral-bearing OC from the
mountain soil.
The FTIR spectra for the chemistry of organo-mineral association.
The aged OC is highly aromatic (1596 and 1386 cm
The high-resolution synchrotron-based X-ray diffraction confirmed the nature of associated minerals to be mainly SRO Fe oxyhydroxides, specifically ferrihydrite (ICDD 01-073-8408), goethite (ICDD 01-073-6522), lepidocrocite (ICDD 00-044-1415), and quartz (ICDD 00-033-1161) (Fig. 5; Table S1 in the Supplement). Quartz may be at most a minor component on the OC surface, considering their chemistry and particle size; however siliceous mineral surfaces may be coated with a veneer of hydrous Al- and Fe-oxides, which could confer a net positive charge and promote their reactivity in tropical environments (Chen et al., 2014a; Sposito, 1989).
Considering their large surface area and high reactivity, the abundant
nano-sized Fe oxyhydroxides could play a significant role in the long-term
stabilization of OC through chemical bonding and physical shielding
(Eusterhues et al., 2005; Kaiser et al., 2002b; Kiem and Kogel-Knabner, 2002;
Mikutta et al., 2006), and contribute to the longevity of OC in mountain
environments. According to elemental mapping, aluminosilicates may also be
present, however, their portion and crystallization levels should be low
judging from their minimal signal in the XRD spectra (Figs. 5, 6). The
primary minerals in mountain soil were quartz, amesite
(kaolin–serpentine), and muscovite (Fig. S4); minimal Fe oxyhydroxides signal
was observed. The mountain soil was not rich in Fe
oxyhydroxides; however, the studied particulate organic C was rich in Fe
oxyhydroxides. The key working minerals for physical protection and chemical
stabilization of OC could be nano-sized SRO minerals in a soil that is not
rich in Fe oxyhydroxides. The FTIR analyses demonstrated the chemistry of
organo-mineral association (Fig. 7; Table S2). The aged OC is highly aromatic
when it is highly reactive, similar to pyrogenic C that has been exposed to the
natural environment (Liang et al., 2006). Broad bands are observed at
1596 cm
Our in situ description of organo-mineral interplay at nanoscale provides direct evidence on the importance of mineral physical protection for the long term stabilization of OC. Large amounts of ferrihydrite and other Fe oxyhydroxides were also found associated with lignin-like OC in soil under an aquic moisture regime (Eusterhues et al., 2011). The abundance of mineral nanoparticles, and the nature of their high heterogeneity and short-range-order, could be common in a humid environment; however, they may have been severely underestimated by traditional analysis methods such as electron microscopy, X-ray diffraction and fractionation approaches, which mainly focus on clay-sized minerals (Mikutta et al., 2005). We recommend future research explores the following: (1) whether it is a general phenomenon that the minerals interacting with OC surface are essentially mineral nanoparticle and submicron-sized clay minerals; and (2) whether the minerals for C stabilization are primarily nano-sized SRO minerals instead of clay-sized minerals in soils. Perspectives on C stabilization and saturation may be revolutionized once the role of SRO minerals is considered in modeling soil C dynamics, in addition to parameters such as clay type and content. We suggest that the modeling of SOC turnover should also include BC and pyrogenic OC into the biochemically protected pool, as such organic C can persist over millennia under natural exposure (Liang et al., 2008).
In summary we have developed a high-resolution 3-D tomography approach using dual-scan modes and successfully applied it to study the in situ interplay of OC and minerals in a laboratory-made and natural OC–mineral consortium at the nanoscale. We discovered that the stabilization of the 3500-year-old natural OC was mainly attributed to the physical protection of nano-sized Fe-containing mineral (Fe oxyhydroxides) and to the strong organo-mineral complexation. We provided in situ evidence and revealed an abundance of mineral nanoparticles, in dense thin layers or nano-aggregates/clusters, instead of crystalline clay-sized minerals on or near OC surfaces. The key working minerals for C stabilization were reactive SRO mineral nanoparticles and poorly crystalline submicron-sized clay minerals. Nanoscale 3-D tomography provides new insight into the mineral physical protection of OC in soil. This high-resolution 3-D tomography approach is a promising technique for probing the multi-interfacial features between OC and minerals in lab and field samples. It is also potentially a powerful tool for tracking the fate of nanoparticles including heavy metals in the natural environments.
Videos of the 3-D tomography for the Supplement Figs. SMOV1 (
The supplement related to this article is available online at:
The authors declare that they have no conflict of interest.
We thank Dr. Chung-Ho Wang for his kind support; Ms. Hsueh-Chi Wang (TXM,
TLS-BL01B01), Dr. Yao-Chang Lee and Ms. Pei-Yu Huang (FTIR, TLS-BL14A1), and
Dr. Hwo-Shuenn Sheu and Dr. Yu-Chun Chuang (XRD, TPS-09A1) at the
end-stations of NSRRC (Taiwan) for their technical support; Yoshiyuki Iizuka (Academia Sinica) for helping with the SEM-EDS
analysis;
Dr. Chih-Hsin Cheng (National Taiwan University) for the SC specimen;
Dr. Yen-Hua Chen (NCKU, the department of Earth Sciences) for the TiO
Biqing Liang and Chun-Chieh Wang acknowledge the funding support from Taiwan Ministry of Science and Technology (MOST 102-2116-M-006-018-MY2, MOST 105-2116-M-006-010, and MOST 105-2112-M-213-001). Edited by: Sébastien Fontaine Reviewed by: two anonymous referees