Structural stability and natural ¹³C abundance of humic acids in buried volcanic ash soils

Abstract

To examine the relationships between soil age and the structural stability and natural ¹³C abundance (δ¹³C) of humic acids, buried humic layers in volcanic ash soil profiles near Mt Fuji, Japan, namely Yubunebara (present to 10,000 years before present (ybp); n = 5) and Ashitaka-Onoue (<1,000–30,000 ybp; n = 6), were analyzed. The contents and δ¹³C values of the fulvic acids (humic substances) and non-humic substances in the fulvic acid fractions were also determined. The degree of humification of the humic acids differed remarkably between the surface and buried layers, and the A ₆₀₀/C value (absorbance at 600 nm mg⁻¹ C mL⁻¹ of humic acids in 0.1 mol L⁻¹ NaOH) increased toward the older buried layers. The proportion of aromatic C in total C of the humic acids, estimated using cross polarization/magic angle spinning ¹³C nuclear magnetic resonance spectroscopy, also increased, whereas the proportions of alkyl and O-alkyl C decreased from the surface toward the older layers. These observations suggested higher stability in the aromatic than aliphatic components and the contribution of aromatic C to the increase in the A ₆₀₀/C value. The composition of the C functional groups did not vary appreciably among layers older than 10,000 ybp. The δ¹³C value of bulk soil organic matter (SOM) fluctuated in both profiles, suggesting that the proportions of C₃ and C₄ plants grown when each layer was the surface layer differed among the layers. Although, in general, variation in the δ¹³C value of each humus fraction corresponded to that of bulk SOM, there were significant differences in the δ¹³C value between the humus fractions and bulk SOM, and the differences were more conspicuous in the younger layers and less conspicuous in the older layers. A negative correlation (with an exception) was also observed between the difference in the δ¹³C values of humic acids from that of bulk SOM and the relative content of aromatic C in the humic acids. Possible causes of the differences and changes in the δ¹³C values are discussed.

Kew words:

• ¹³C NMR
• degree of humification
• fulvic acids
• humic acids
• stable carbon isotopic ratio

INTRODUCTION

Humic substances are defined as dark-colored amorphous polymers that were synthesized from biomass constituents or their metabolites biochemically and/or chemically in the environment (Stevenson 1994). It has been considered that initial humic substances lose their components that are relatively more easily decomposed, while condensation–polymerization and oxidative reactions proceed, which results in a more stable structure (Kumada 1987). These processes correspond to the progression of humification and are frequently evaluated based on an increase in the degree of darkening, particularly for soil humic acids, the alkali-soluble and acid-insoluble fraction. The degree of humification may enable us to predict the stability of humic acids in soil. However, detailed structural changes throughout the humification process have not been sufficiently elucidated, partly because of the long period of time required to trace them. For example, Otsuka et al. (1994) estimated that the duration of the period required for the chemical formation of Type A humic acids, the highest class in terms of the degree of humification, was approximately 150–2,000 years, based on the incubation of plant residues with volcanic ash at 75°C and 90°C. Although it is difficult to determine the applicability to natural soil of the results they obtained at high temperatures in the absence of microbial activity, other data also indicated that humic acids in volcanic ash soils (Melanudands) approximately 50 and 240 years after ash fall belonged to Types P₀ and B, intermediate classes in terms of the degree of humification (Kumada 1987).

A comparison of humic acids in soils of different ages has been used to analyze the long-term process of humification (Kumada 1987; Otsuka and Kumada 1978; Sakai and Kumada 1982). In Japan, more than two buried humic layers that have developed in deposits of volcanic materials are sometimes observed in a soil profile. The soil age could be estimated from the presence of age-known volcanic ash or ¹⁴C measurement. Direct comparison among the layers is impossible if there is no evidence showing that climatic conditions and the type and amount of vegetation are identical among the layers. However, such soil profiles are better, as samples, to assume the diagenesis of humic acids than soils collected from different sites, because the influence of geological differences could be ignored, particularly in plains. Based on the analysis of published data on humic acids in several volcanic ash soil profiles, Kumada (1987) suggested that the increase in the degree of humification, the changes in elemental composition and the formation of a graphitic structure were likely to be almost complete within 1,000–3,000 years. Variations in more detailed structural characteristics, such as C functional groups, have not been determined.

Natural abundance of ¹³C, which is generally expressed as δ¹³C in unit of per mil () using Eq. 1, is a signature providing information about the origin and production process of each organic material (Barrie and Prosser 1996):

where R is the molar ratio ¹³C/¹²C and standard refers to the Cretaceous carbonate fossil Bellemnitella americana from the Peedee Formations in South Carolina, USA. Variations in δ¹³C values are derived from differences in the rates between isotopes during the physical, chemical and biological processes. The δ¹³C value of plant cells varies depending on their photosynthetic pathway, C₃ or C₄ (Bender 1971; Smith and Epstein 1971), which is reflected on the δ¹³C value of soil organic matter (SOM) where respective types of plants are grown (Balesdent et al. 1987; Schwartz et al. 1986). Even in soil profiles under constant vegetation historically, the increase in the δ¹³C value of bulk SOM (up to 5) with increasing depth or ¹⁴C age has been observed through horizons (Agnelli et al. 2002; Bol et al. 1999; Huang et al. 1996), suggesting a positive effect of diagenetic processes on ¹³C enrichment. Although the cause of the increase in the δ¹³C value has not been elucidated and there are few data indicating the chronological changes in δ¹³C values of humic acids (Agnelli et al. 2002), the δ¹³C value could become a key index in the elucidation of long-term humification processes.

In the present study, the diagenesis of humic acids toward the development of a more stable structure was investigated by comparing the composition of C functional groups, estimated by solid state ¹³C nuclear magnetic resonance (NMR) spectroscopy, and the degree of humification among buried humic layers with different soil ages using volcanic ash soil profiles near Mt Fuji, Japan. In addition, the relationships between soil age or the structural properties of the humic acids and their δ¹³C values were analyzed. To obtain supplementary information, the content and δ¹³C values were also determined for bulk SOM and for the humic and non-humic substances in the fulvic acid fractions.

MATERIALS AND METHODS

Soil samples

Soil samples collected from one Yubunebara (Gotenba, Shizuoka, Japan) and two Ashitaka-Onoue (Numazu, Shizuoka, Japan) soil profiles were used. Detailed descriptions of the soil profiles were presented in the reports of Yoshida et al. (1978) and Tsutsuki and Kuwatsuka (1989), respectively. Layer designations were also referred to in these reports. Five humic layers in the Yubunebara profile, including the surface layer (Yu-1), a sub-layer aged 1140–1200 years before present (ybp; Yu-2), and three sub-layers that were assumed to have accumulated at ca. 3,000 (Yu-3), 5,000 (Yu-5) and 10,000 ybp (Yu-7) were used. From the Ashitaka-Onoue profile, six humic layers with approximate ages of <1,000 (At-A), 8,000 (At-D), 18,000 (At-4), 27,200 (At-8), 28,100 (At-15) and 30,000 (At-20) ybp were selected.

Analysis of humus composition

Total C content in the soil samples was determined using a NC analyzer, Sumigraph NC-800 (Sumika Chemical Analysis Service, Osaka, Japan). Humic and fulvic acid fractions were extracted according to the NAGOYA method (Ikeya and Watanabe 2003; Kuwatsuka et al. 1992) by continuous shaking with 0.1 mol L⁻¹ NaOH (300 mL g⁻¹ soil C) at 25^oC for 24 h under N₂. Purification procedures to obtain powder samples of humic acids also followed the NAGOYA method.

The fulvic acid fraction obtained as an acidic supernatant after precipitation of humic acids in the extract by acidification (pH < 1.5) was then fractionated, using a column packed with cross-linked insoluble polyvinylpyrrolidone (PVP; “Polyclar SB100”, Gokyo Sangyo, Osaka, Japan), into adsorbed and non-adsorbed fractions according to the method of Watanabe and Kuwatsuka (1991). Because the humic and non-humic substances in the fulvic acid fractions were concentrated in PVP-adsorbed and non-adsorbed fractions, respectively (Watanabe and Kuwatsuka 1991, 1992), these two fractions were designated as fulvic acids and water-soluble non-humic substances (WS-NHS) in the present study. The organic C concentration was determined for a portion of these two fractions using a dissolved C analyzer, TOC-V_CPH (Shimadzu, Kyoto, Japan), under acidic conditions after bubbling with N₂. Another portion of each fraction was allowed to dry on a hot plate at pH 3–4 and was mixed thoroughly in an agate mortar.

The logarithmic value of the ratio of the absorbances at 400 and 600 nm of humic acids in 0.1 mol L⁻¹ NaOH, log(A ₄₀₀/A ₆₀₀), and the absorbance at 600 nm mg⁻¹ C mL⁻¹, A ₆₀₀/C, were determined to express the degree of humification. The organic C concentration was determined using the TOC-V_CPH (Ikeya and Watanabe 2003), and the absorbances were measured with a spectrophotometer, UV-2200 (Shimadzu).

All the above procedures were conducted in duplicate.

Measurement of solid state ¹³C NMR spectra

Solid state ¹³C NMR spectra of the humic acids were recorded at 75.6 MHz on a Chemagnetic CMX-300 spectrometer (Varian, Palo Alto, CA, USA) using the cross polarization/magic angle spinning (CPMAS) method under the following conditions: spinning rate, 5000 Hz; contact time, 1 msec; pulse delay time, 0.7 s; line broadening, 50 Hz. Accumulation was repeated 1,000–2,000 times. The spectra were divided into four regions, 0–45, 45–110, 110–160 and 160–190 ppm, and the relative area of each region in the total spectral area was determined. These four regions were represented by the major C species appearing in each region (i.e. saturated alkyl C, alkyl C substituted with hetero atom [O-alkyl C], aromatic C and carbonyl C from higher toward lower magnetic fields). The spinning side bands of the signals with maximums at approximately 130, 150 and 170 ppm were detected at approximately 200, 220 and 240 ppm, respectively. The counter side bands occurring at approximately 60, 80 and 100 ppm overlapped with the signals of O-alkyl C. For the correction, the sum of the areas for the peaks at 200, 220 and 240 ppm was subtracted from the area of O-alkyl C, while twice the area for the former two peaks and for the latter peak were added to the area of aromatic C and carbonyl C, respectively (Monteil-Rivera et al. 2000).

Measurement of δ¹³C

The δ¹³C values of the powder samples of the three humus fractions as well as the pulverized soil samples were measured using a stable isotopic ratio mass spectrometer connected to an elemental analyzer (DELTA Plus-NC2500, Thermo Finnigan, San Jose, CA, USA). The experimental error was ±0.1 . The relationship and significance of the difference between the δ¹³C value of bulk SOM and each humus fraction were analyzed using a regression analysis and t-test, respectively.

RESULTS

Humus composition

Figure 1 shows the total C content in the soil samples and the proportions of the three humus fractions in total SOM on a C basis. Large decreases in the total C content were observed from the Yu-1 to Yu-3 samples in the Yubunebara profile (Fig. 1a) and between the At-A and At-D samples in the Ashitaka-Onoue profiles (Fig. 1b). According to Tsutsuki and Kuwatsuka (1989), there were two humic layers designated as At-B and At-C (2,100 and 3,500 ybp of age, respectively) between At-A and At-D, and total C content decreased conspicuously between the two layers.

The proportion of humic acids in total SOM was in the range of 16.2–31.9% (Fig. 1c,d) and did not show a relationship with soil age. The proportion of fulvic acids in total SOM tended to be greater in the upper two layers (7.0–7.2% or 8.3–10.5%) than in the lower layers (2.0–2.8% or 5.1–6.3%) within the respective Yubunebara and Ashitaka-Onoue profiles. A similar trend was also observed for WS-NHS, namely 16.7–18.0% versus 5.6–7.8% in the Yubunebara and 16.5–17.2% versus 5.6–10.4% in the Ashitaka-Onoue profiles.

The difference in the degree of humification of the humic acids (Fig. 2) was remarkable between the Yu-1 and Yu-2 samples. The Yu-1 humic acids were classified into Type B and the others belonged to Type A (Ikeya and Watanabe 2003; Kumada 1987). The A ₆₀₀/C value of the humic acids in the buried layers appeared to increase gently with increasing soil age up to ca. 10,000 ybp.

¹³C CPMAS NMR spectra of humic acids

Changes in the ¹³C CPMAS NMR spectra were distinct from the Yu-1 to Yu-3 humic acids and between the At-A and At-D humic acids, respectively (Figs 3, 4). The resolution of the signals assigned to the methoxyl groups (54.0–55.9 ppm) and aromatic C substituted with hetero atom (151.5–153.8 ppm) was higher in the spectra of the younger samples, Yu-1, Yu-2 and At-A humic acids. The relative intensity of the signals of alkyl C with the maximum at 22.6–32.5 ppm, secondary alcohol (71.9–73.1 ppm) and acetal (103.3–105.7 ppm) was also greater. The difference in the signal intensity in each region between the older samples could be distinguished after correction of the influence of spinning side bands (Fig. 5).

Figure 1  Total carbon (C) content in (a) Yubunabara and (b) Ashitaka-Onoue soils and the proportions of the three humus fractions in total soil organic matter on a C basis in (c) Yubunebara and (d) Ashitaka-Onoue soils. (•), humic acids; (▴), fulvic acids (PVP-adsorbed fraction of fulvic acid fraction); (▵), water-soluble non-humic substances (PVP-non-adsorbed fraction of fulvic acid fraction).

The proportion of aromatic C in total C estimated from the ¹³C CPMAS NMR spectra (Fig. 5) increased from the Yu-1 (31.1%) to Yu-7 humic acids (66.7%) and from the At-A (45.4%) to the At-8 humic acids (68.5%), which corresponded to the decrease in the proportions of alkyl and O-alkyl C. A proportion of aromatic C above 63% was observed for all the buried layers aged 10,000 ybp (Yu-7) and older. The variations in the proportions of carbonyl C were small among all the humic acids and ranged between 15.7% (Yu-1) and 19.9% (At-D).

δ¹³C value of each humus fraction

The δ¹³C value of bulk SOM fluctuated both in the Yubunebara and Ashitaka-Onoue profiles (Fig. 6), suggesting that the proportions of C₃ and C₄ plants grown when each layer was the surface layer differed among the layers. In general, variations in the δ¹³C value of each humus fraction corresponded to that of bulk SOM (Fig. 6). In the regression analysis, the δ¹³C values of the three humus fractions were positively correlated with that of bulk SOM (n = 11; r = 0.835–0.896; P < 0.005). The δ¹³C values of the humic acids, fulvic acids and WS-NHS ranged between −17.4 and −23.2, −19.8 and −25.0 and −16.1 and −20.8, respectively. The δ¹³C values of the humus fractions in the surface soils were higher in the order, WS-NHS, humic acids and fulvic acids, with >0.9 difference between them. This was also applicable to younger buried layers from Yu-2 to Yu-5, whereas the δ¹³C values of the fulvic acids were higher than those of the humic acids, with >0.5 differences in the Yu-7, At-4, At-8 and At-15 layers.

DISCUSSION

The lower C content in the older buried layers does not necessarily indicate that C decreased with time because the initial C content before burial may have differed among the layers. However, the smaller proportions of WS-NHS, which frequently contain polysaccharides and peptides/proteinaceous materials as major components (Anderson et al. 1989; Cheshire et al. 1992; Watanabe and Kuwatsuka 1992), in total SOM and those of alkyl and O-alkyl C in humic acid C in the older layers, suggested the loss of components that are relatively more easily decomposed in humus after burial. The proportions of alkyl C and O-alkyl C in total C for the humic acids in the buried layers aged 10,000 ybp and older, 2.3–3.1% and 9.9–14.7%, respectively, also tended to be smaller than the values for Type A humic

Figure 2  Degree of humification of humic acids in (a) Yubunebara and (b) Ashitaka-Onoue soils. (▪), A600/C; (□), log(A400/A600).

acids in surface volcanic ash soils, namely 10.8–20.9% and 15.6–32.0% (n = 5; Maie et al. 2002), estimated using methods similar to those used in the present study.

It is generally difficult to detect significant differences in the values of log(A ₄₀₀/A ₆₀₀) among Type A humic acids because of the small range, approximately 0.5–0.6 (Hiradate et al. 2004; Ikeya and Watanabe 2003; Kumada 1987). The distribution of the A ₆₀₀/C value (Fig. 2) and of the relative aromatic C content (Fig. 5) was also narrow for the humic acids in the buried layer soils. However, regression analysis revealed a significant positive correlation between both values for all 11 humic acids (r = 0.926; P < 0.005) and for 10 samples lacking Yu-1 Type B humic acids (r = 0.929; P < 0.005), as previously observed for various soil humic acids including A, B, P and Rp types (Watanabe et al. 2005). Thus, aromatic components in humic acids were more stable than aliphatic ones in soil, and their concentration probably contributed to the increase in the A ₆₀₀/C value from the surface to the buried layers. The present results did not provide information about the progression of polycondensation of aromatic components.

Figure 3  13C cross polarization/magic angle spinning nuclear magnetic resonance spectra of Yubunabara humic acids.

Fluctuations in the δ ¹³ C values, which were probably caused by the difference in the relative occupations by C ₃ and C ₄ plants, may have masked chronological changes in the δ ¹³ C values of the humus fractions. However, the differences in the δ ¹³ C values between each humus fraction and bulk SOM showed some characteristics of the δ¹³C values of the respective humus fractions. The δ¹³C values of WS-NHS were statistically higher (n = 11; P < 0.05) than that of bulk SOM, which may result from the large amount of carbohydrates and peptides/proteinaceous materials, because hemicellulose, cellulose and amino acids in plant cells tend to show higher δ¹³C values than whole plant cells (Benner et al. 1987; Deines 1980; Gleixner et al. 1993), and the carbohydrates and amino acids synthesized by microorganisms are expected to be enriched in ¹³C compared with the substrates (Macko and Estep 1984; Šantrůčkováet al. 2000). The differences in the δ¹³C values between WS-NHS and bulk SOM were smaller in the older layers. Possible interpretations include the microbial synthesis of WS-NHS using ¹³C-depleted organic matter (such as fulvic acids) as a C source while the

Figure 4  13C cross polarization/magic angle spinning nuclear magnetic resonance spectra of Ashitaka-Onoue humic acids.

decomposition of existing WS-NHS proceeded, and/or increases in the δ¹³C values of the other humus fractions with time. The difference in the δ¹³C values between WS-NHS and the fulvic acids was negatively correlated with soil age (r = −0.86, P < 0.005; data not shown), which is compatible with either of the above hypotheses. The δ¹³C value of humin, which was estimated by balancing the amount of δ¹³C between bulk SOM and the three humus fractions, was also likely to be higher than the δ¹³C value of bulk SOM in the buried layers, although the difference was small (<1; data not shown).

Lignin tends to show a lower δ¹³C value compared with whole plant cells (Benner et al. 1987; Gleixner et al. 1993) and has been considered to be the major source of aromatic components in humic and fulvic acids. In general, the δ¹³C values of the humic and fulvic acids were lower (n = 11; P < 0.005) than the values of bulk SOM (maximum −3.3), which was possibly related to the contribution of lignin-derived C to humic and fulvic C. Lichtfouse et al. (1995) also reported higher and lower δ¹³C values of humin and humic acids than that of bulk SOM, respectively, and Hiradate et al. (2004) showed that the average δ¹³C values were significantly lower for humic and fulvic acids than for bulk SOM in surface volcanic ash soils. The higher δ¹³C values of the humic acids compared with those of the fulvic acids in the surface and younger buried humic layers might be due, in part, to the higher amino acid contents in the humic acids than in the fulvic acids (Watanabe et al. 1994). The δ¹³C values of the humic acids after the loss of their aliphatic components in the older layers (Fig. 5) were lower than those of the fulvic acids (Fig. 6).

Figure 7 shows the presence of a negative correlation (r = −0.921; P < 0.005) between the difference in the δ¹³C values of the humic acids and bulk SOM [δ¹³C_HAs− δ¹³C_{bulk SOM}] and the proportion of aromatic C in total C of the humic acids for all samples except for the At-D layer. This observation could be explained by

Figure 5  Proportions of carbon (C) functional groups in total C of (a) Yubunebara and (b) Ashitaka-Onoue humic acids estimated from 13C cross polarization/magic angle spinning nuclear magnetic resonance. (○), alkyl C; (□), O-alkyl C; (•), aromatic C; (▴), carbonyl C.

Figure 6  δ13C values of bulk soil organic matter (SOM) and humus fractions in (a) Yubunebara and (b) Ashitaka-Onoue soils. (□), bulk SOM; (•), humic acids; (▴), fulvic acids (PVP-adsorbed fraction of fulvic acid fraction); (▵), water-soluble non-humic substances (PVP-non-adsorbed fraction of fulvic acid fraction).

Figure 7  Relationship between the difference in the δ13C value of humic acids and bulk soil organic matter [δ13CHAs −δ13Cbulk SOM] and the proportion of aromatic carbon (C) in the total C of humic acids (n = 10).

assuming that, in general, aliphatic and aromatic components in humic acids show relatively higher and lower δ¹³C values, respectively, like those in plant cells (Deines 1980; Gleixner et al. 1993), indicating the selective preservation of humified-lignin or humified-tannin as core structures of humic acids. In contrast, the formation of humic acids from charred plant materials, which could be an important source of humic substances in volcanic ash soils (Kramer et al. 2004; Shindo et al. 2004), and subsequent preservation may not have resulted in smaller δ¹³C values of humic acids relative to that of bulk SOM because the δ¹³C values of charred plant materials were reported to be closer to that of bulk SOM than to those of humic acids (Shindo et al. 2005). However, the difference in the δ¹³C values between humic acids and bulk SOM could also become greater if the δ¹³C values of fulvic acids and/or humin increased in their diagenesis, even if the δ¹³C values of humic acids did not decrease. Thus, it is still possible that the accumulation of microbe-derived and/or char-derived aromatic C contributes to the high aromatic C content in humic acids with a high degree of humification.

CONCLUSIONS

The degree of humification of humic acids increased in soils within 10,000 ybp, and the corresponding increase in the relative content of aromatic C reflected the enhancement of the structural stability. The C composition of the humic acids did not vary appreciably among layers older than 10,000 ybp.

Although variation in the δ¹³C value of each humus fraction generally corresponded to that of bulk SOM, the difference in the δ¹³C value of bulk SOM varied among the fractions and among the layers.

There was a negative correlation between the relative content of aromatic C in the humic acids and [δ¹³C_HAs −δ¹³C_{bulk SOM}].