Conjugation of artificial humic acids with inorganic soil matter to restore land for improved conservation of water and nutrients

Artificial soil (AS) with a dark–brown appearance and high soil organic matter (SOM) was manufactured under hydrothermal conditions from poor sandy landfill using the 'hydrothermal humification process.' This approach may retrosynthesize the natural clay–humin complex, which is the main contributor for water and ion binding in fertile, natural soils. The structure and morphology of as‐created organic–inorganic composites was examined, and it was shown that the as‐created artificial SOM (A‐SOM) indeed tightly binds to the mineral surfaces, thus creating remediated soil or more general AS. A‐SOM does not change the bulk structure of the involved inorganic minerals but activates their surface. Depending on the biomass used as the starting product, the high effectivity of synthetic humification improves the organic carbon and nitrogen content when compared with the employed sandy soil (SS). The composition was adjusted to be comparable with a reference sample of cultivated soil (2.92% in C content and 7.8% in SOM) to enable a fair referencing. We then analyzed the most frequently used soil quality indicators for agricultural land use and found strong water retention and nutrient conservation, which reflects the successful restoration of mineral–humus conjugation. Pure ASs provide superior performance in the analyzed series, whereas simulated soils (mixture of SS and AS) still exhibit satisfactory capacities of water and ion bonding. The values were found to be very similar to cultivated soils sampled from Germany and Harbin, China.

is covered with a layer of dark-colored soils, that is, BS, chernozem, meadow soil, dark-brown forest soil, brown soil and so on. Black soil in Northeast China is characterized by high organic matter and is a fertile soil in China, with a total area approximately 1.03 million km 2 (Xu, Xu, Chen, Xu, & Zhang, 2010). The three BS regions are distributed in the north temperate zone with four distinct seasons. These three BS areas are most suitable for farming, which store thick humic substances and organic carbon from the incomplete decomposition of biomass partly condensed to BS matter through thousands of years. To feed an increasing world population, severe degradation of BSs took place over the past several decades (Yao et al., 2017). Humic matter is thereby not a lasting gift but can be degraded, depending on conditions even rather fast (Lützow et al., 2006;Schmidt et al., 2011). Unlike the BS in North- In general, base soil is composed of minerals, organic matter, gases, liquids and organisms, and organic matter and minerals form the main body of the soil. Notably, soil organic matter (SOM) composition resembles typical BS more than other major soil units (Thiele-Bruhn, Leinweber, Eckhardt, Siem, & Blume, 2014), which is a driver for important soil functions like carbon storage and nutrient release (Franko & Merbach, 2017). This can be quantitatively expressed as soil organic carbon contents, and the high (~5-8 wt %) organic carbon content in natural BSs (Yao et al., 2017) make them more fertile (Lal, Negassa, & Lorenz, 2015). Of similar importance, the storage of C in SOM is three to four times higher than the carbon stored in terrestrial vegetation and atmosphere (Lehmann & Kleber, 2015), indicating that it has a crucial role in the global carbon cycle. The climate change thereby could be moderated through storing enormous amounts of C in the soil while decreasing CO 2 concentration in the atmosphere. We consider this as the most feasible tool to control a warming world while using the excess carbon for food production at the same time (Cohen, Pfeiffer, & Francis, 2018;Shepherd, 2016). However, SOM is affected by global change due to the interactions with climate conditions and changes of land management (Franko & Merbach, 2017). Therefore, the stability of SOM in the environment is particularly important for soil quality assessment. The SOC (soil organic carbon) accumulated in mineral soil (SOC Min ) exists in soils as labile and stabilized fractions, especially in the topsoil horizon (Grüneberg, Schöning, Hessenmöller, Schulze, & Weisser, 2013), whereas subsoil SOM is more stable (Chabbi, Kögel-Knabner, & Rumpel, 2009 showed that fire-derived aromatic C in SOMs-A is more stable than microbially derived aliphatic C. The higher decomposition of SOMs-B fractions weakened the relationship of their Δ 14 C values with the alkyl and aryl C contents. Natural processes can provide us the leading inspiration to generate thick layers of SOM as in BSs, but they are inefficient and slow as an immediate tool to counteract destructive human interference. Geochemical knowledge is more than sufficient to enable a retrosynthetic approach to mimic humin formation by faster and more efficient chemical processes. As well known, SOM is composed of the humification products of plant residues both above and below ground (primary sources), microbial residues (secondary sources), and various products from chemical decomposition and biochemical activity (Haberhauer, Rafferty, Strebl, & Gerzabek, 1998). Our previous work demonstrated the feasibility of carbon sequestration simulating natural coalification by hydrothermal chemical approaches (Hu et al., 2010;Titirici, Thomas, & Antonietti, 2007), and concurrent work has realized successful fabrication of humic substances from crude biomass (F. Yang, Zhang, Cheng, & Antonietti, 2019). Under the applied mild condition, the hardly convertible lignin in biomass reacts together with organic acids and dehydrated sugars, and most analytical parameters of the products are practically indistinguishable from organic matter isolated from real soil. This follows elder knowledge where scientists described interrelation of natural organic acids and lignins in the natural soil formation process (Marshall & Page, 1927).
In the present paper, we describe an attempt to transform poor sandy soil (SS) from a landfill operation into 'new BS' through a onestep conversion process, in which the physical and chemical properties of color change and mineral composition and structure are similar to real BS. The soil carbon content in the new BS is increased, which will facilitate the activation and integration of soil minerals. Worth noting, soil fertility aligns to crop yields represents agricultural productivity (Bünemann et al., 2018;Patzel, Sticher, & Karlen, 2000), which is described as 'the ability of the soil to supply essential plant nutrients and soil water in adequate amounts and proportions for plant growth and reproduction, in the absence of toxic substances which may inhibit plant growth' (www.fao.org). The ability to bind water and nutrients thereby are two important indicators for soil quality. These quantities are determined for the model artificial soils (AS) and compared with the corresponding values or cultivated Germany soil and Chinese BS.
This contribution is part of a series where we aim to prepare humic substances and SOM only using an easy hydrothermal process (hydrothermal humification [HTH]) on the base of crude biomass leftovers or agricultural 'waste' for large-scale land remediation. It is clear that we can only present a first set or scientific fundamentals, and many other aspects require the involvement of the agricultural stakeholders and end users to a much larger degree than practiced to date. Before experiments, the rude biomass should be cut into a suitable size. Typically, 1.20-g rude biomass powders, 5-g SS, and various KOH qualities (ensuring the molar rations between degradable cellulose and alkali close to 1) were put into the bottom of the glass tube in 50-ml autoclave in the oven to heat to 200 C and then keep for 24 hr to ensure the sufficient coalification. After the temperature naturally cooled down to room temperature, the AS with brown color can be obtained after subsequent drying procedure. All solutions were prepared with ultrapure water (Millipore, 18 MΩcm) for both material preparation and adsorption experiments. Typically, SOM examination of SS, AS, and cultivated soil is also carried out following the procedures: first, the open space was placed in a muffle furnace, burned at 95 C for 30 min, taken out and cooled in a desiccator for 20 min, and the mass was weighed; then, it was burned at the same temperature for 30 min, taken out, cooled, and weighed. Repeated the above steps until the mass difference between the two times is less than 0.5 mg. Weigh about 0.50 g of dry sample in a known quality crucible, put it in an oven at 105 C for 12 h, then take it out, put it into the desiccator, cooled it, weighted it, and recorded the quality; next, transfer the crucible to a muffle furnace and heat up to 550 C, burn for 5 hr, cool in a desiccator, weigh, and record the mass and calculate. Basic mineralogical properties of soil samples were listed in Table S1. Additionally, a glucose model as well as two different models of crude biomass were humified in the presence of a controlled amount of base, and the resulting humic-mineral complexes were characterized by scanning electron microscopy (SEM), elemental analysis (EA), energy dispersive X-ray (EDX), and Fourier-transform infrared spectroscopy (FTIR) analyses.

| RESULTS AND DISCUSSIONS
Our starting point is a rather low-quality SS sampled at the location of our institute, as it is typical for land refilling after building operations and 10 years of spontaneous recultivation. Figure 1 depicts the sampling location just below the thin weak new cultivation layer (0-20 cm). One can see the typical yellow to yellow-gray fine powder sand with uniform texture, rich in calcium or nodules with prominent vertical joints. These characteristics in structure and components cause that water seepage is easy, as it is to be eroded by water to form gullies, and also to cause subsidence and collapse. Left-alone recultivation thus provided only thin humus horizons, poor soil carbon content, and a very low ability of water and nutrient conservation (Cui F I G U R E 1 A photograph of the actual location for sampling poor sandy soil at the MPI Golm campus, Germany, as left after building operations. The insert depicts the appearance of sandy soil and this soil after conjugation with artificial humic acids F I G U R E 2 Low-and high-magnification scanning electron microscopy images of the sandy soil sample from a refilled construction site In HTH technology, crude plant biomass is added to the soil minerals, and the mixture is 'cooked' with a certain amount of KOH in an autoclave under anaerobic conditions. Applying temperatures of the order of 200 C and autogenous pressure for a few hours greatly accelerate the humification process. The amount of biomass is directly related to the SOM content and contrary to composting, most of the carbon indeed is kept in the soil. Doing the humification in the presence of the minerals enables mineral activation with abundant OH functionalities by the starting base, endowing them a more hydrophilic character, and the organic acids formed later in the process promote a tight conjugation of organic matter with the surface of the minerals, a structure mimicking the mineral-humus complex (Varadachari, Mondal, Dulal C., & Ghosh, 1994). This also avoids separation of the organic matter by liquid water towards floating macroscopic organic patches, as it generally found for mixed-in compost (Liu et al., 2018).  Figures S4 and S5), along with fibers and other particles with biological texture, structural reminders of the natural biomass precursors due to the presence of lignin, that cannot be degraded under such mild hydrothermal reaction conditions. However, a majority of humic substances is found to be tightly bound to the soil minerals, constituting the synthetic SOM (A-SOM). This indeed is very similar to the appearance of organic matter in real soil.
Literature reports that natural organic acids and lignin derivatives occur in close relation and that the natural humification process obviously results in coexistence of weakly modified lignins with condensed organic acids (Marshall & Page, 1927).
The natural huminogenesis involves biological, physical, and chemical processes to convert dead plant materials into organic soil matter that are able to form intimate associations with soil minerals (Lehmann & Kleber, 2015). However, due to the predominant metabolization of the biomass as such, the efficiency of this cascade is only low. The whole hydrothermal reaction is purely chemical and very efficient with respect to the carbon yield, as practically all carbon of the waste biomass is contained in the final material. Worth noting, in spite of the rather high amounts of added base, we have examined the pH in SS and BS before and after the hydrothermal reaction and found 7.22 versus 7.10 and 6.85 versus 6.76, respectively, indicating that the transformation of biomass into organic acids and biochar in the presence of the added base is essentially pH buffered. It progresses until neutrality is reached, and the soil after manufacturing can be used as such. We assume that the larger fulvic and humic acids are strongly bonded to the soil minerals to form intimate associations F I G U R E 3 Low-and high-scanning electron microscopy images of glucose-based artificial soil with soil minerals (e.g., the many clay minerals contain cations including Ca and Mg at their surface), thus keeping the nature of near pH neutrality as an important property of soil.
The corresponding EDX results ( Figures S6 and S7) confirm that the carbon in AS is mainly bound to the mineral surface. We determine 30.5 wt % for wood-based artificial soil (W-AS) and 27.6 wt % for leaf-based artificial soil (L-AS), much higher than in the parental SS (6.7 wt %) and even higher than in cultivated soils (17.8 wt %). These values are higher than the corresponding values determined by chemical combustion analysis (discussed next), reflecting that EDX is a nearsurface sensitive technique and that the carbon is indeed localized at the surface of the mineral particles but not within. The conjugation of A-SOM with the minerals is thereby quantified, thus complementing the qualitative visualization from the photographs ( Figure S8). The content of the common elements of soil (O, Si, Fe, Al, K, Ca, and Mg) remains nearly constant (only relatively low due to the increased carbon content) before and after hydrothermal treatment. This underlines our assumption that the minerals are possibly surface etched for better coating but not changed in their structure and composition.
The data of EA to quantify the carbon content are displayed in Table 1. The carbon contents at the surface and in the bulk are expectedly very different, but the trend is the same, that is, AS shows a highly increased organic carbon content (0.55% in G-AS, 3.44% in W-AS, and 3.89% in L-AS) as compared with the SS precursor F I G U R E 4 Low-and high-scanning electron microscopy images of wood-based artificial soil F I G U R E 5 Low-and high-scanning electron microscopy images of leaf-based artificial soil (0.02%). The enrichment at the surface is in all experiments of the order of a factor of 100. Figure S9 presents the relationships of carbon contents obtained from EA and EDX analyses with a classical SOM determination, indicating that EA results indeed agree with the SOM determination.
To our deep satisfaction, the synthesis of artificial SOM via HTH indeed increased SOM content in a depleted mineral soil, without having the ability to wash it away. The numbers are 1.2%, 7.8%, 2.9%, 16.6%, and 12.0% for SS, cultivated soil, G-AS, W-AS, and L-AS, respectively. We have chosen the biomass content in the primary synthesis in a way that the SOM is even larger than in the cultivated soil sample to allow subsequent dilution. Worth noting, this general approach gives us a promising method to prepare retrosynthesized BSs with a variety of compositions just through regulating the ratio of soils and biomass as well as the biomass type, and currently, samples were created with an amount of humic components, which exceeds natural BSs. Interestingly, AS can improve soil very well and not only increase SOM but also change soil structure, making them close to BSs, as concluded in Table S1. Another point that needs to be emphasized is that the as-prepared AS is fabricated under medium-high temperatures and pressure, that is, the product is sterile and free of microorganisms but also prions, peptides, DNA, or more complex organic molecules such as agricultural chemicals are decomposed and converted.
A use in plant breeding is favorably based on a mixture with real soils to establish an again active microbiome, that is, an also biologically fertile soil system. SEM images of simulated soils (mixtures of AS and SS), cultivated soil (Brandenburg, Germany black soil, GBS, Germany), and BS (Northeast China; Figures S10-S12) depict indeed similar rough surfaces so that mixing should occur without size depletion or differences in wettability.
An initial approximation to determine potential mineralogical variations was done with X-ray diffraction (XRD) on dried soil powders.
As seen from Figure 6, nine peaks can be attributed to SiO 2 (Joint Committee on Powder Diffraction Standards no.78-2315) in the natural sand refill (the main component), centered at 20.7 , 26.6 , 36.5 , 39.5 , 40.3 , 42.4 , 45.8 , 50.1 , 54.9 , and 57.3 , respectively Besides, the XRD profiles of five studied soils have some additional small peaks located at 27.3 /27.8 , which may be attributed to iron or calcium phosphates or iron, calcium, and magnesium silicates.
The FTIR curves exhibit some differences among five studied soils ( Figure 7a), with three strong peaks centered at 3,627, 1,637 and 1,537 cm −1 , respectively, for BS caused by stretching vibration of OH, C C, and C N groups, different to the same regions on the other samples: two weak absorption peaks (3,357 and 1,640-1,620 cm −1 ) are observed in the spectra of cultivated and AS. For all the curves, the existence of bands located at 1,100-1,000, 782, 692, and 600 cm −1 is attributed to the vibrations of C O, C H, OH, and/or NH 2 groups. Generally, FTIR curves in the range of 1,000-2,000 cm −1 allow for a better understanding of the chemical structure of SOM. Figure 7b-f shows more details in this region with the second derivative curves (inserts). This is a well-known tool to effectively separate the broadband absorption peaks of complex dense matter with the otherwise nearly invisible changes through added peaks. As expected, five main peaks at 1,717, 1,650, 1,616, 1,540, and 1,508 cm −1 , respectively, are identified in every curve, corresponding to vibration of C O ( COOH), C O (amides I), asymmetric COO , C N (amides II), and C C (aromatics).
In order to examine the changes in different functional groups of the studied soils, peak-fitting results in two representative regions with semiquantitatively analysis are also presented in Figure 8; the peak areas come from FTIR fitting quantified in Table S2. In range of 2,800-3000 cm −1 representing aliphatic T A B L E 1 Comparative carbon contents from elemental analysis and energy dispersive X-ray analysis and soil organic matter Abbreviations: EA, elemental analysis; EDX, energy dispersive X-ray; G-AS, glucose-based AS; SOM, soil organic matter.
In addition, the area ratios of A 1,650-1,800 to A 1,500-1,650 and  (Table S3). The BS is out of this comparison due to the different soil origin. Worth noting, artificial fulvic and humic acids obtained from glucose and crude biomass respectively exhibit rather big differences, including in the values of the C index (0.36 vs. 0.22), the ratios of A 1,650-1,800 to A 1,500-1,650 (0.55 vs. 0.28) and A 1,040-1,260 to A 1,500-1,650 (5.15 vs. 3.67); the as-formed AS by addition of lignin containing crude biomass is indeed closer to the data of real soil, reflecting the increase and match of humic substances. Therefore, we have chosen L-AS as a reference for AS for the following systematic soil property studies.
As discussed above, water and nutrition retention are two important evaluation criteria for determining soil fertility. Therefore, water retention and nutrient conservation are briefly tested in order to preliminarily evaluate the AS performance in comparison with real BS. Figure S13 depicts the experimental equipment for water and nutrient binding examinations with an initial 200-ml solution F I G U R E 7 Fourier-transform infrared spectroscopy spectrum of the five studied soils (a) sandy soil, (b) cultivated soil, (c) glucose soil, (d) leaf soil, (e) and black soil (f), and the second derivative of the spectra at 1,000-2,000 cm −1 Fourier-transform infrared spectroscopy region  In general, the formation of natural humic substances in geological soil needs quite a long time, up to a few thousand years; it is based both on chemical as well as on biological processes, and the yield is F I G U R E 8 Curve-fitting Fourier-transform infrared spectroscopy spectra for five studied soils at 2,800-3,000 and 800-1,800 cm −1 Fourier-transform infrared spectroscopy region The mixture of sandy soil and artificial soil with the mass ratio of 1:1.
[Colour figure can be viewed at wileyonlinelibrary.com] comparably low (Visser, 1962). This report aims to accelerate the underlying carbon fixation processes simply by chemical engineering means while keeping the carbon fixation yield close to complete. At the same time, both process and chemicals must be tuned that the artificial product is matching as much as possible with the composition and structure of real soils to find a suitable and sustainable approach for soil modification. Ideally this will access the process of 'terraforming' of weak or used-up soil matter to give fertile plating media.