IV. Chemical Structure of Humic Substances

22



M. SCHNITZER



FAs, humins, and whole soils using a variety of methods. These included the oxidation of methylated and unmethylated humic substances with an alkaline KMnO4

solution (Schnitzer and Khan, 1978; Griffith and Schnitzer, 1989). The somewhat

milder oxidation with alkaline CuO, as well as the sequential oxidation with CuONaOH ϩ KMnO4 and with CuO-NaOH ϩ KMnO4 ϩ H2O2 solutions, has also

been employed (Schnitzer and Khan, 1978; Schnitzer, 1978). Humic substances

have also been degraded under acidic conditions with peracetic acid and nitric acid

(Schnitzer, 1978). Other oxidants used include alkaline nitrobenzene, sodium

hypochlorite, and H2O2 solutions (Schnitzer and Khan, 1978). Degradations with

Na2S and phenol have also been carried out (Hayes and O’Callaghan, 1989).

Major compounds produced by the oxidation of methylated and unmethylated

humic substances from widely differing pedological and geographical origins under alkaline as well as under acidic conditions are aliphatic carboxylic, phenolic,

and benzenecarboxylic acids (Schnitzer and Khan, 1978; Schnitzer, 1978; Griffith

and Schnitzer, 1989).

Among aliphatic oxidation products are mono-, di-, tri-, and tetracarboxylic

acids. Major aromatic oxidation products are benzenecarboxylic acids such as the

tri, tetra, penta, and hexa forms (Fig. 10), whereas phenolic acids include compounds containing between one and three OH groups and between one and five

CO2H groups per aromatic ring (Fig. 11).

From the oxidation products identified and from 13C NMR spectra of humic

substances, it appears that aromatic rings are cross-linked by paraffinic chains (Fig.

12). On oxidation, the aliphatic carbons closest to the rings become the C of CO2H

groups and remain bonded to the rings whereas the other carbons in the aliphatic

chains are oxidized to either aliphatic acids or CO2. The formation of CO2 from



Figure 10



Major benzenecarboxylic oxidation products.



A LIFETIME PERSPECTIVE



Figure 11



23



Major phenolic oxidation products.



the oxidation of side chains may explain the low oxidation yields of aliphatic acids

compared to benzenecarboxylic and phenolic acids. Several conclusions can be

drawn from the oxidative degradation of humic substances extracted from hundreds of soils of diverse origins: (a) isolated aromatic rings are important structural units of all humic substances, (b) aliphatic chains are linking aromatic rings

to form alkyl aromatic networks, (c) the model structure shown in Fig. 12 has an



Figure 12 Chemical structure for humic substances based on oxidation products.



24



M. SCHNITZER



aromaticity of 50% if we exclude functional groups, and (d) the structure in Fig.

12 also contains voids of various dimensions that can trap organic and inorganic

soil constituents. These characteristics are typical of soil humic substances (Schulten and Schnitzer, 1997).



B. REDUCTIVE DEGRADATION

Reductive degradation is another approach to obtaining structural information

on humic substances. Essentially, the methods used most widely for this purpose

are Na-amalgam reduction and Zn-dust distillation and fusion (Stevenson, 1994).

Reduction with Na-amalgam produces phenols and phenolic acids that are thought

to be released through the cleavage of other linkages present in humic substances.

Zn-dust distillation and Zn-dust fusion are harsh methods that have been used for

the structural analysis of alkaloids and other complex organic molecules. These

methods yield polycyclic hydrocarbons and may provide useful information on the

“core” of humic substances. Major products formed by the Zn-dust distillation of

HA and FA are methyl-substituted naphthalene, anthracene, phenanthrene, pyrene,

and perylene (Hansen and Schnitzer, 1969). Methyl groups on the polycyclic rings

are probably the remains of longer alkyl chains linking the polycyclics in HA and

FA structures.



C. PY-FIMS OF HA, FA, AND HUMIN

The Py-FI mass spectrum of a HA extracted from the Armadale horizon (a Spodosol) (Fig. 13a) (Schnitzer and Schulten, 1992) shows the presence of four major components: carbohydrates, phenols, lignins, and n-fatty acids. Noteworthy is

the prominence of the n-C24 (m/z 368), n-C26 (m/z 396), n-C27 (m/z 410), n-C28

(m/z 424), and n-C30 (m/z 452) fatty acids. The whole range of n-fatty acids extends from C16 to C34. Other components present in smaller amounts are

monomeric lignins, n-C10 to n-C20 diesters, and n-C44 to n-C50 alkyl monoesters,

of which the n-C45 monoester (m/z 662) is the most abundant. Relative weak signals characteristic of N components are m/z 59 (acetamide), 79 (pyridine), 81

(methylindole), 93 (methylpyridine), 117 (indole), 131 (methylindole), and 167.

The Py-FI mass spectrum of a FA extracted from the Armadale Bh horizon (Fig.

13b) is dominated by carbohydrates, phenols, and lignins. The most intense signals

are m/z 58 (acetone) and m/z 60 (acetic acid). Both compounds are emitted thermally

from methylethylketones, carbohydrates, and fatty acids at temperatures Ͼ300ЊC.

The spectrum also shows the presence of smaller amounts of n-fatty acids (m/z 256,

284, 312, and 382), sterols (m/z 414), n-alkyl diesters, and monomeric and dimeric

lignins. No distinct signals due to N-containing compounds can be detected.

The Py-FI spectrum of a humin separated from the Armadale Ah horizon (Fig.



A LIFETIME PERSPECTIVE



25



Figure 13 Py-FI mass spectrum of (a) a HA extracted from the Armadale Ah horizon and (b) a FA

extracted from the Armadale Bh horizon. From Schnitzer and Schulten (1992), with permission of the

publisher.



14) shows the presence of carbohydrates, phenols, monomeric and dimeric lignins,

alkylbenzenes, and alkyl esters. The presence of a homologous series of n-fatty

acids, ranging from n-C16 to n-C27, is indicated. Of special interest is the series of

n-alkylbenzenes with signals at m/z 316, 330, 344, 358, 372, 386, 400, 414, and

428, which appear to indicate the presence of C6H5иC17H35 to C6H5иC25H51 nalkylbenzenes, respectively. Molecular ions m/z 206 and 220 appear to arise from

di- and trimethyl phenanthrene. Intense signals probably due to n-C10 to n-C20 nalkyl diesters are observed from m/z 202 to 342. Except for weak signals for pyrrole (m/z 67) and methylpyrrole (m/z 81), no signals due to N-containing compounds appear in this spectrum.

Table III summarizes the compounds identified in the Py-FI mass spectra of HA,

FA, humin, and soil. The most abundant compounds identified in the humic fractions are carbohydrates, phenols, lignin monomers, lignin dimers, n-fatty acids, nalkyldiesters, and n-alkylbenzenes. Minor components include n-alkyl mono- and

diesters, n-alkylbenzenes, methylnaphthalenes, methylphenanthrenes, and N-containing compounds. HA tends to be enriched in n-fatty acids and the humin in nalkylbenzenes.



26



M. SCHNITZER



Figure 14 Py-FI mass spectrum of a humin separated from the Armadale Ah horizon. From

Schnitzer and Schulten (1992), with permission of the publisher.



Assignments of the major signals in the presented mass spectra were made as

described in considerable detail by Schnitzer and Schulten (1995).



D. CURIE-POINT PYROLYSIS GAS CHROMATOGRAPHY /

MASS SPECTROMETRY (GC/MS)

Detailed descriptions of the experimental details of Py-FIMS and of Curie-point

pyrolysis GC/MS of humic substances and whole soils have been published pre-



Table III

Compounds Identified in the Initial Armadale Soil and in HA, FA,

and Humin Fractions Isolated from It a

Compound identified



Soil



HA



FA



Humin



Carbohydrates

Phenols

Lingin monomers

Lingin dimers

n-Fatty acids

n-Alkyl monoesters

n-Alkyl diesters

n-Alkyl benzenes

Methylnaphthalenes

Methylphenanthrenes

N compounds

n-Alkanes



ϩϩb

ϩϩ

ϩϩ

ϩϩ

ϩ

ϩ

ϩϩ

ϩ

ϩ

ϩ

ϩ

ϩ



ϩϩ

ϩϩ

ϩϩ

ϩϩ

ϩϩϩ

ϩ

ϩ



ϩϩ

ϩϩ

ϩϩ

ϩϩ

ϩ

ϩ

ϩ

ϩ



ϩϩ

ϩϩ

ϩϩ

ϩϩ

ϩϩ

ϩ

ϩϩ

ϩϩ

ϩ

ϩ

ϩ



a From



ϩ



Schnitzer and Schulten (1995).

weak (relative intensity <20%); ϩϩ, intense (relative intensity 20–60%); ϩϩϩ, very intense

(relative intensity >60%).

bϩ,



A LIFETIME PERSPECTIVE



27



viously (Schulten and Schnitzer, 1997; Schulten et al., 1998). While in Py-FIMS,

the sample was heated at a rate of 10 K minϪ1 from 323 to 973 K; the final pyrolysis temperatures of 573, 773, and 973 K were attained with the Curie-point pyrolyzer between 3 and 9.9 sec. This fast transfer of thermal energy to the sample

makes this method a valuable tool for structural studies on humic materials. The

resulting thermal shock produces small, stable organic pyrolysis products.

While Py-FIMS shows that carbohydrates, phenols, lignin monomers, lignin

dimers, lipids (alkanes, alkenes, fatty acids, and n-alkyl esters), alkylaromatics,

and N-containing compounds are major HA components, Curie-point pyrolysis

GC/MS of HAs indicates the presence of relatively large amounts of alkyl-substituted aromatic hydrocarbons (Schulten et al., 1991). Of special significance is the

identification of a series of C1 to C22 n-alkyl benzenes. In addition, ethylmethyl

benzene, methylpropyl benzene, methylheptyl benzene, methyloctyl benzene, and

methylundecyl benzene were also detected. Other compounds identified are

trimethyl- and tetramethylbenzenes, alkylnaphthalenes, and alkylphenanthrenes.

The alkyl substitution of naphthalene ranges from 1 to 5 methyls, whereas that of

phenanthrene ranges from 1 to 4 methyls.



E. A TWO-DIMENSIONAL STRUCTURE OF HA

On the basis of both Py-FIMS and Curie-point pyrolysis GC/MS data, Schulten et al. (1991) proposed that HA consists of isolated aromatic rings linked covalently by aliphatic chains. In the hand-drawn HA structure in Fig. 15 (Schulten

and Schnitzer, 1993), n-alkyl aromatics play a significant role. Oxygen is present

in the form of carboxyls, phenolic and alcoholic hydroxyls, esters, ethers, and ketones, whereas nitrogen occurs in nitriles and heterocyclic structures. The resulting carbon skeleton shows high microporosity with voids of various dimensions,

which can trap and bind other organic and inorganic soil constituents as well as

water. The elemental composition of the HA is C308O90N5, its molecular mass is

5540 Da, and its elemental analysis is 66.8% C, 6.0% H, 26.0% O, and 1.3% N.

The HA structure in Fig. 15 is supported by chemical (Schnitzer and Khan, 1978;

Schnitzer, 1978), oxidative, and reductive degradative (Schnitzer and Khan,

1978; Schnitzer (1978), colloid-chemical (Ghosh and Schnitzer, 1980) electron microscopic (Stevenson and Schnitzer, 1982), and 13C NMR and X-ray

(Schnitzer, 1994) data obtained on HAs over many years and by exhaustive consultations of the voluminous literature on humic substances. As far as the model

HA structure is concerned, Schnitzer and Schulten (1995) assumed that carbohydrates and proteinaceous materials are adsorbed on external HA surfaces and in

internal voids and that hydrogen bonds play an important role in their immobilization. Aside from carbohydrates and proteinaceous materials, the voids can

also trap and bind lipids and biocides as well as inorganics such as clay minerals

and hydrous oxides.



28



M. SCHNITZER



Figure 15 Two-dimensional HA model structure. From Schulten and Schnitzer (1993), with permission of the publisher.



F. A THREE-DIMENSIONAL STRUCTURE OF HA, SOM,

AND WHOLE SOIL

The two-dimensional (2D) HA structure (Fig. 15) was converted to a three-dimensional (3D) structure model with the aid of Hyper Chem software. Details of

the different steps involved in this conversion are published elsewhere (Schulten

and Schnitzer, 1997; Schulten et al., 1998) so only the most significant findings

will be discussed here. The first 3D HA model structure was published by Schulten and Schnitzer in 1993. Its elementary composition is C308 H335O90N5, with a

molecular mass of 5547.9 g molϪ1 and an elemental analysis of 66.78% C, 5.79%

H, 25.99% O, and 1.26% N.

There are different views in the literature on SOM as to whether carbohydrates

and proteinaceous materials are adsorbed on or loosely retained by HA or whether

they are bonded covalently to HA. Regardless of which mechanism is considered,

carbohydrates and proteinaceous materials are HA components for analytical

purposes because their presence affects the elemental analysis, functional group

content, and molecular weight of HA. According to Lowe (1978), carbohydrates

constitute about 10% of the HA weight; a similar value has been suggested for proteinaceous materials in HA (Khan and Sowden, 1971). Thus, Schulten and



A LIFETIME PERSPECTIVE



29



Schnitzer (1993) assumed that one molecular weight of HA interacts with 10% carbohydrates and 10% proteinaceous materials. The resulting HA has an elemental

composition of C342H388O124N12, with a molecular weight of 6650.8 g molϪ1 and

an elemental analysis of 61.8% C, 5.9% H, 29.8% O, and 2.5% N. When more carbohydrates and proteins are added to the HA, the C content decreases, but the O

content increases. For the development of the HA structure, Schulten and Schnitzer

(1993) assumed that carbohydrates and proteinaceous materials were not integral

HA components but were adsorbed in internal voids and on external surfaces. In

1997, Schulten and Schnitzer modified the model because it was too small to accommodate all oxygen-containing functional groups. Also, on average, (CH2)n

chains were too long because the proposed preliminary C–C skeleton (Schulten et

al., 1991) was based mainly on data obtained by Py-GC/MS and Py-FIMS, which

quantitatively showed methylene units ranging from n ϭ 1 to n ϭ 20. The improved HA model, which includes a trapped trisaccharide and a polypeptide in its

voids, has the following elemental composition (Schulten and Schnitzer, 1997):

C305H299N16O134S1. Its elemental analysis is 57.56% C, 4.73% H, 3.52% N,

33.68% O, and 0.5% S. Its molecular mass is 6365 Da. The sizes of the voids are

large enough to occlude polysaccharides, peptides, water, biocides, etc. The improved model contains 5 aliphatic and 21 aromatic carboxyl groups, 17 phenolic

hydroxyls, 17 alcoholic hydroxyls, 7 quinonoid and ketonic carbonyls, 3 methoxyls, and 1 sulfur function. The aliphatic links between aromatic units have been

shortened to between 1 and 10 CH2 units, with an average of n ϭ 5.

Schulten and Schnitzer (1997) considered the relationship between HA and

SOM. In agricultural soils, the bulk of SOM consists of humic substances (HA,

FA, and humin). Several workers (Schnitzer and Khan, 1978; Schnitzer, 1978;

Preston et al.,1989) have shown that the chemical structures of HA and humin are

similar. According to these data, humin is HA so strongly complexed by clays and

hydrous oxides that it can no longer be extracted by dilute base or acid. As far as

FA is concerned, 13C NMR spectra of HA and FA are also similar. The major differences are that FA has a lower molecular weight and is richer in CO2H groups,

in O, and in carbohydrates than HA, but structurally the two fractions are similar

(Schnitzer, 1994). The same type of information also comes from oxidative degradation studies (Schnitzer, 1978) and Py-FI mass spectrometry (Sorge et al., 1994).

Also, in many agricultural soils, except Spodosols, FA constitutes less than 10%

of the SOM, so that is a minor humic fraction.

Thus, for agricultural soils we can define SOM in the following manner (Schulten and Schnitzer, 1997): SOM ϭ HA ϩ carbohydrates ϩ proteins (1). For example, for a soil containing 3.0% SOM we can write: SOM ϭ 2.50% HA ϩ 0.25%

carbohydrates ϩ 0.25% proteins (2). The improved 3D SOM model is based on

this definition.

The only major SOM component that has not been considered so far is water.

The water content of air-dry HA, FA, and humin is of the order of 3.0% (M.

Schnitzer, unpublished data).



30



M. SCHNITZER



As a next step, Schulten and Schnitzer (1997) further developed the improved

SOM model structure to include 3% H2O. The elemental composition of this structure (Fig. 16, see color insert) is C349H401N26O173S1. Its elemental analysis is

54.0% C, 5.2% H, 4.7% N, 35.7% O, and 0.4% S, with a molecular mass of 7760

Da. Note that the elemental analysis of this three-dimensional model HA is close

to naturally occurring HAs (Schnitzer, 1978).

In an attempt to develop a 3D chemical model of a whole soil, Schulten and

Schnitzer (1997) proposed the following definition for an average agricultural soil:

Soil ϭ 3% SOM ϩ 3% H2O ϩ 94% inorganics (3).

Detailed structural features of a soil particle are shown in Fig. 17 (see color insert). Voids in the model SOM structure are capable of occluding organics, inorganics, and water, and the functional groups are involved in reactions with metals and

minerals and provide nutrients to plant roots and microbes. Note that in Fig. 17, SOM

is bound to silicates via Fe3+ and Al3+ ions. The SOM in the simulated soil particle

is surrounded by a model matrix of silica sheets. Of interest to soil chemists is that

the modeled soil particle displays 23 hydrogen bonds, which again emphasizes the

importance of this type of linkage. Schulten and Schnitzer (1997) calculated that 13

of the hydrogen bonds are intramolecular, 9 are in the mineral matrix, and only 1 is

between SOM and a silica sheet. The spaces in Fig. 17 between the mineral matrix

and SOM are several magnitudes larger than the voids in SOM so that the mineral

surfaces are not completely covered or shielded by SOM. This allows access to the

mineral surfaces of metal ions, small organic molecules, and water.

Other SOM characteristics that can be determined by computational chemistry

are surface area, volume, refractivity, polarizability, hydrophobicity, and hydration

energy (Schulten and Schnitzer, 1997). These characteristics can also be helpful in

the development of model SOM structures.

We can expect that applications of computational chemistry to the development

of model SOM structures will increase and lead to a better understanding of the

spatial arrangements of the molecular constituents of SOM, organic mineral complexes, and soil aggregates. A more comprehensive knowledge of chemistry and

reactions of SOM will certainly be beneficial to a sustainable agriculture and help

us protect the environment.



V. NITROGEN-, PHOSPHORUS- AND SULFURCONTAINING COMPONENTS OF SOM

A. ORIGINS AND FUNCTIONS OF SOIL NITROGEN

SOM acts as a storehouse and supplier of N to plant roots and microorganisms;

almost 95% of the total soil N is closely associated with SOM (Schnitzer, 1978).



A LIFETIME PERSPECTIVE



31



Nitrogen is essential for crop production as it is an important constituent of proteins, nucleic acids, porphyrins, and alkaloids. Nitrogen is the only essential plant

nutrient that is not released by the weathering of minerals in the soil. The main

source of soil N is the atmosphere, where dinitrogen (N2) is the predominant gas.

Only a few microorganisms have the ability to use molecular N2; all remaining living organisms require combined N for carrying out their life activities. Increases

in the level of soil N occur through the fixation of N2 by some microorganisms and

from the return of ammonia and nitrate in rain water; losses are due to the harvesting of crops, leaching, and volatilization. Atmospheric ammonia originates

from the volatilization from soil surfaces, lightening, fossil fuel combustion, and

natural fires.

While a considerable amount of research has been done over the years on soil

N, most of this work has been limited to the quantitative and qualitative determinations of proteinaceous materials, amino acids, amino sugars, ammonia, and

nitrates. Reviews on soil N summarize the known organic N forms in soils

(Stevenson, 1994) as well as their mineralization and importance in plant nutrition (Mengel, 1996). Because about one-half of total soil N remains unidentified

and poorly understood, there is a need for more research and information in this

area.



B. NITROGEN DISTRIBUTION IN SOILS

AND HUMIC SUBSTANCES

Sowden et al. (1977) determined the distribution of major N compounds in samples taken from soils formed under widely different climatic and geologic conditions on the earth’s surface. While the total N contents of the samples analyzed

ranged from 0.01 to 1.61%, the proportions of total N that could be hydrolyzed by

hot 6 M HCl were quite similar, ranging from 84.2 to 88.9%. Amino acid N varied from 33.1 to 41.7%, amino sugar N from 4.5 to 7.4%, and ammonia N from

18.0 to 32.0%. Proportions of unidentified hydrolyzable N ranged from 16.5 to

17.8%, whereas those of nonhydrolyzable N ranged from 11.1 to 15.8%. Estimates

of non-protein N ranged from 55% in tropical soils to 64% in arctic soils, averaging 61% of the total N in all soils (Sowden et al., 1977). From these data it appears

that about 60% of the total soil N is non-protein or, conversely, that 40% of the total soil N is protein N. To establish whether hydrolysis with hot 6 M HCl hydrolyzed all proteinaceous materials in soils and humic substances, Griffith et

al.(1976) hydrolyzed a number of soils and humic materials first with hot 6 M HCl

and then hydrolyzed separate samples of the acid-treated residues with either 0.2

M Ba(OH)2 or 2.5 M NaOH under reflux. The results obtained showed that hot 6

M HCl released almost all of the amino acids in the soil and humic substances in

24 hr.