FATE AND TRANSPORT OF HYDRAZINE
THROUGH COLUMNS OF SATURATED SANDY SOIL
WAYNE C. DOWNS
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
I would like to thank Dr. Robert Mansell for his patience and persistence
during the data acquisition and document preparation for this manuscript. It
would never have happened without him. I would also like to thank Dr. Michael
Annable for agreeing to step in at a late date to supervise the document and Dr.
Joseph Delfino, Chairman of the Department of Environmental Engineering
Sciences, for devising a way to make it all possible. Thanks, too, go to Dr. Brian
McNeal who recently agreed to become a member of the committee, and to Dr.
Paul Chadik, who patiently waited things out. I would also like to thank Dr.
Stephen Bloom for his ion exchange discussions and modeling assistance, and
Dr. Wayne Huber for his example and guidance during my graduate career at
the University of Florida. He believed in me all along.
Several coworkers deserve special mention for their assistance in the
research described herein: Ana Moliner, for long and fruitful discussions of soil
chemistry, and Denie Augustyn and Robin Roberson for laboratory assistance.
Special thanks go to my wife Jill, for her love and patience during the
years of graduate school and the weeks of managing the family without me
during the preparation of this document. Truly, without her encouragement this
work would never have come to pass.
I would also like to acknowledge the faculty and technical staff of the Soil
and Water Science Department at the University of Florida for their kind
assistance and partial support. The U. S. Environmental Protection Agency's
Robert S. Kerr Environmental Research Laboratory in Ada, Oklahoma, deserves
particular recognition for allowing me the time and providing the equipment to
finish the laboratory experiments. Also, thanks go to EG&G Idaho, Inc. for partial
support in completing the writing of the manuscript. This work was initiated
under a grant to the Soil and Water Science Department of the University of
Florida by the U. S. Air Force Environics Division, Tyndall Air Force Base,
Florida (No. F08635-83-C-0136, CPT Floyd Wiseman, Project Officer).
TABLE OF CONTENTS
ACKNOW LEDGMENTS ............................................................................................ ii
CHAPTER 1. INTRODUCTION ......................................... ............... ..................... 1
CHAPTER 2. LITERATURE REVIEW ...................................... .................................. 7
Literature Review Objectives............................................................ ............... 7
Hydrazine Environmental Chemistry............................................... .............. 7
Hydrazine Fate and Transport Pathways........................................ ............ 10
CHAPTER 3. MATERIALS AND METHODS....................................... ............. 34
Research Objectives ......................................................................................... 34
Soil Characterization ........................................................................................35
Miscible Displacement...................................................................................... 43
CHAPTER 4. RESULTS ........................................................................................ 58
CHAPTER 5. DISCUSSION ..................................................................................86
Introduction ............................................................................................................. 86
Environmental Variables ................................................................................... 87
Process Variables.............................................................................................. 97
CHAPTER 6. SUMMARY AND CONCLUSIONS..........................................................113
Introduction ........................................................................................................ 113
Summary of Experimental Design... .............................................................. 113
Summary of Experimental Results ................................................................ 115
Conclusions .......................................................................................................... 120
LIST O F R EFER EN C ES ....................................................................................... 122
APPEN D IX A ........................................................................................................... 132
A PPEN D IX B........................................................................................................... 138
A PPEN D IX C .......................................................................................................... 149
A PPEN D IX D ..................................................... .................................................... 160
APPEN D IX E...................................................... .................................................... 181
BIO G RA PH IC A L SKETC H ........................................................................................ 190
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
FATE AND TRANSPORT OF HYDRAZINE
THROUGH COLUMNS OF SATURATED SANDY SOIL
Wayne C. Downs
Chairman: M. D. Annable
Cochairman: R. S. Mansell
Major Department: Environmental Engineering Sciences
The effects of environmental and process variables on the fate and
transport of hydrazine were investigated in laboratory columns of three
consecutive horizons of a saturated sandy soil. The investigation of
consecutive horizons containing successively less organic matter showed
hydrazine loss within soil columns to be closely correlated with percentage
organic matter. Percentage clay in each horizon was not well correlated with
The influence of the ion-exchange process was investigated by
observing the ionic composition of column effluent. An effective cation
exchange capacity (CEC) was determined under column-saturated flow, and
found to be an order of magnitude less than CEC values determined from batch
studies. The column-determined value was used in computer simulations to
correctly predict observed hydrazinium breakthrough curves. Results indicated
that ion exchange and ion transport are primary mechanisms that describe the
transport of hydrazine during water flow in these calcium-saturated soils.
A hydrazine mass balance showed losses of 10 to 37 percent in the
columns, depending on the horizon. Losses were found to be correlated with
the duration of column experiments, implying a first-order degradation
Microbial activity also was observed for soil taken from completed
column experiments. Hydrazine concentrations as high as 16 mmol L-1 were
not observed to reduce active microbial populations. Plate counts of
approximately 107 organisms per gram of soil were observed, compared to 108
organisms per gram using acridine-orange counting methods.
Literature values of first order hydrazine degradation rates due to
microbial activity are similar to the rate-controlled losses observed in this study,
though specific experiments to isolate microbial activity were not performed.
Hydrazine (N2H4) and its derivatives are extremely versatile compounds
that have found application for a wide variety of purposes. They are readily
oxidizable and endothermic, and for this reason have been used in fuel cells, as
propellant for gas turbines, as antioxidants, for the deoxygenation of boiler
water, in pharmaceuticals production, and as intermediates for the production of
explosives and propellants. The agricultural industry is a major user of
hydrazine in pesticide production. Hydrazine, along with its derivatives,
monomethyl (MMH) and unsymmetrical dimethylhydrazine (UDMH), also is
used by the defense industry as a liquid propellant in missiles, satellites, and
Hydrazine was first prepared by T. Curtius, a German chemist, in 1887. It
remained little more than a laboratory curiosity with few applications for several
decades (Schiessi, 1980). The first sample of anhydrous hydrazine was
prepared by Lorty DeBruyn in 1893. The method for preparation of hydrazine
hydrate by the Raschig process was discovered in 1907 and cleared the way for
its production in industrial quantities. Hydrazine did not enjoy significant use,
however, until its propellant capabilities were realized by the Germans during
World War II. An energetic propellant was needed for the rocket airplane, the
Me-163B, developed in 1937. A mixture of hydrazine hydrate and methanol
was used as a fuel, with hydrogen peroxide as the oxidizer in a bipropellant
rocket engine (Schmidt, 1987). The first use of hydrazine as a monopropellant
was demonstrated in 1954 at the Jet Propulsion Laboratory in Pasadena,
California The late 1950s saw a great increase in the production of anhydrous
hydrazine in the United States, and hydrazine production increased again in
the 1960s when a blend of UDMH and hydrazine was used to fuel the Titan
series rocket engines.
Development of the Shell 405 catalyst in 1963 allowed almost unlimited
restart capability and opened the way for new hydrazine applications. As a
result, most military, commercial, and scientific satellites in earth orbit use
hydrazine propulsion systems for attitude control and orbit maintenance. Many
unmanned space missions also have used hydrazine propulsion, such as the
Viking landers on Mars; the Pioneer and Voyager space probes to Jupiter,
Saturn, and Uranus; and the Giotto space probe to Haley's comet.
Hydrazine propellant is used extensively for upper-stage rocket
propulsion and for impulse corrections after rocket motor burn is complete. The
space shuttle uses both hydrazine and monomethyl-hydrazine for its second-
stage booster rockets and for orbital maneuvering. Anhydrous hydrazine is
used in the Auxiliary Power Unit on both the space shuttle orbiter and on its two
solid rocket boosters. Another widespread aviation use of hydrazine is in the
Emergency Power Unit on the F-16 fighter (Clewell et al., 1988).
Hydrazine decomposition gasses at high pressure are used to expel
ballast water from submarine ballast tanks in emergency situations. Such
systems are in use on several NATO submarines. Hydrazine systems weigh
only a fraction of comparable compressed-gas systems.
While hydrazine use was dominated by the military and aerospace
industry in the 1960s, by the 1980s other industrial applications were
consuming the major share of all hydrazine produced. The agricultural industry
is a major user of hydrazine in the manufacture of pesticides. Hydrazine is used
in the plastics industry as a chemical intermediate for plastic-foam blowing
agents, and hydrazine is used as an oxygen scavenger in boiler water for
In laboratory animals, exposure to hydrazine may produce either
immediate toxicity or delayed kidney and liver injury in animals that survive the
exposure. Via inhalation, the 4-hr LCso of hydrazine is 7.8 mmol L-1 for mice
and 18.1 mmol L-1 for rats (Clewell etal., 1988). In another study, the 1-hr LC50
in rats was 20 mmol L-1. A six-month inhalation study conducted with dogs,
monkeys, rats, and mice suggested that effects were dose related regardless of
whether the exposures were intermittent or continuous (WHO, 1987).
Hydrazine is a polar molecule, having a high affinity for water.
Consequently, it is extremely irritating to eyes and mucus membranes.
Hydrazine has also been shown to enter the body through the skin. In
anesthetized dogs, topical application of hydrazine in the 100 mg kg-1 range
produced detectable blood concentrations within 30 seconds and a chemical
burn at the site of application (Clewell et al, 1988).
A few instances of hydrazine toxicity in humans have been reported.
Dermal sensitization after exposure to hydrazine has been cited (WHO, 1987).
Accidental ingestion of a concentrated aqueous solution of hydrazine by a
workman caused prolonged unconsciousness and seizures; however, he was
considered reasonably recovered within two weeks (Clewell et al., 1988).
Hydrazine toxicity has been fatal in at least one case where an individual
experienced conjunctivitis, nausea, and tremors each time he handled
hydrazine. After six months of repeated exposure he was admitted to the
hospital and, after three weeks, died despite treatment (Clewell et al., 1988).
Hydrazine is classified as an environmental carcinogen and a suspected
human carcinogen (Stone and Wiseman, 1988). This toxicity has resulted in a
recommendation from the American Conference of Governmental Industrial
Hygienists of a threshold limit value for hydrazine of 0.003 mmol L-1 of air.
Because of uncertainty concerning the relative importance of skin exposure to
hydrazine vapor, it is commonly required that individuals working with
hydrazine wear a self-contained protective suit to provide full-body protection.
The widespread usage of this material provides the opportunity for spills
and subsequent contamination of the environment. In a report on hydrazine use
in the Space Shuttle program, Hudson (1982) indicated that, in addition to the
orbiter and the launch pad where it is fueled, ground facilities involved with
hydrazine activities that service the shuttle include fixed storage tanks, parked
tank trailers, portable service units, piping and vent lines, vent gas scrubbers,
waste tank trailers, contaminated fuel tanks, spill trenches, and ponds. Lewis
(1979) reported that hydrazine fuels are transported between Lake Charles,
Louisiana; Denver, Colorado; Cape Canaveral, Florida; and Vandenberg Air
Force Base, California, as well as Strategic Air Command sites throughout the
country, with an average of 5.2 million pounds of fuels shipped annually over
150,000 miles of rail and highway.
On July 28th, 1991 more than 400 gallons of hydrazine were spilled
when an overheated axle snapped, derailing 12 freight cars that crashed into
an overpass of the Ventura Freeway in Southern California. Three hundred
residents of Seacliff, California, were evacuated and 10 miles of freeway were
closed during the five-day cleanup. Eleven workers were treated for nausea
and respiratory problems during this operation (Reed, 1991).
The potential for leakage of hydrazine from storage tanks and other spills
from transportation accidents makes investigation of the environmental fate of
hydrazine a pressing need.
This study has been proposed to evaluate the fate and transport of
hydrazine in laboratory columns of sandy soil. The transport of hydrazine below
the water table will be simulated using saturated soil columns. Also, saturated
conditions simplify the experimental design as well as the mathematics of
transport. The soils will be characterized with respect to particle size
distribution, organic matter content, elemental composition, pH, and buffering
In addition to soil characteristics, the effect of several environmental
variables known to affect the transport of chemicals in ground water will be
examined, including: solution concentration, water velocity, and time of
hydrazine exposure to soil.
The study approach will be organized in the following manner:
* Through literature review, investigate the environmental chemistry of
hydrazine, and thus elucidate the possible mechanisms governing its fate in
* Identify processes likely to control the fate and transport of hydrazine in soils.
* Identify the soil characteristics and environmental variables likely to impact
the fate and transport processes.
* Isolate the processes thought to be of greatest importance, and quantify
them through column and batch experiments.
* Evaluate the impact of the selected soil characteristics and environmental
variables on the significant fate and transport processes.
* Simulate the selected processes and environmental variables by computer
to confirm their significance and quantify processes difficult to evaluate in
independent laboratory experiments.
Literature Review Objectives
Under saturated soil conditions (as in near-surface aquifers), the fate of
hydrazine is closely linked to its chemical nature and its involvement in a
number of potential degradation and transport pathways. This review will be
devoted to the environmental chemistry of the molecule and to pathways
thought to be available for hydrazine fate and transport.
Hydrazine Environmental Chemistry
Hydrazine is a clear, colorless liquid at ambient temperatures (b.p.
114.50C and m.p. 2.010C). Its vapor pressure is slightly lower than that of water
(10.4 vs 17.5 mm Hg), while its density is quite similar to that of water. It is
extremely soluble in water and, when mixed, an initial difference in refractive
index quickly disappears. Hydrazine vapors are only slightly more dense than
air. Some physical and chemical properties of environmental interest are
tabulated in Table 2-1.
Hydrazine is a symmetrical, yet polar, molecule. It is miscible with polar
solvents such as water, alcohols, ammonia, and amines, and is insoluble in
nonpolar solvents such as hydrocarbons and halogenated hydrocarbons. Its
polarity is due to the attraction of oppositely charged fields for one another; i.e.
the nitrogen molecules rotate internally to minimize electrical repulsion. This
rotation produces a cis-configuration (Audrieth and Ogg, 1951: Figure 2-1).
Due to its polarity, hydrazine most likely will interact with polar groups on the
solid surfaces of soils.
Table 2-1. Some physical and
chemical properties of hydrazine and its hydrate.
Property Anhydrous hydrazine Hydrazine hydrate
(100% N2H4) 1(64% N2H4)
Physical state at STP
Relative vapor density
Solubility in water
Source: World Health Organization (WHO), 1987
The 2s2 shell of each nitrogen atom serves to establish the N-N bond. A
2sp3 hybridization is assumed to occur, in which one of the 2sp3 orbitals from
each nitrogen atom is occupied by a pair of lone electrons with opposite spins
3-9 mg m-3
10.4 mmHg at 200C
1008 g L-1 at 200C
66.7 dyne cm-1
3-9 mg m-3
7.5 mmHg at 200C
1032 g L-1 at 200C
74.2 dyne cm-1
HH H H
\ 110 -
H H 60 H
Figure 2-1. Cis- configuration of the hydrazine molecule.
(Schmidt, 1984). These two electron pairs impart a very strong nucleophilic
character to the molecule. Due to this property, hydrazine can form a large
variety of complexes with metals (Dilworth, 1976). Schmidt (1984) stated that it
is difficult to determine whether a true complex with hydrazine exists or whether
hydrazine is simply included in the crystal lattice of some salts (e.g. calcium
salts) as a solvate analogous to water of crystallization.
The density of hydrazine is higher in the solid state than in the liquid
state. In this respect, it is different from water. The density of the liquid at
ambient temperature is 1.01 g cm-3 (Ahlert et al., 1962).
The use of hydrazine as a fuel is based on its endothermic nature
[(AH)f(l) = +12.1 Kcal mole-1]. However, anhydrous hydrazine is thermally quite
stable (250 C) in the ambient-temperature range (Schiessl, 1980). The
presence of certain metals and oxides can lower the decomposition
Hydrazine is a base slightly weaker than ammonia (pKa=7.95, Condon et
al., 1974) and a strong reducing agent. As such it reduces many metal ions to
lower valence states or to the metal itself, depending on reaction conditions.
The standard redox potentials (Latimer, 1952) of hydrazine:
NH4 + 40H- N2 + 4H,0 + 4e-
and of hydrazinium ion:
Eo = +0.23V [2-2]
indicate that hydrazine is a better reducing agent in alkaline than in acidic
solution. Hydrazine can also act as an oxidizing agent, as indicated by the
following standard redox potentials:
NH4 + 2H20+ 2e- -> 20H- + 2NH3
E = +0.1V [2-3]
N2H,' + 3H+ + 2e- -> 2NH4+
S= +1.27V [2-4]
Hydrazine Fate and Transport Pathways
The fate of hydrazine in soils is determined by chemical, physical, and
biological relationships developed in the soil. Some environmental
relationships have been investigated, and are discussed below:
Oxidation in Solution
Some of the earliest work done with hydrazine was directed towards the
identification of reaction products expected upon reaction with various oxidizing
agents. Browne and Shetterly (1907, 1908, 1909) classified oxidizing agents
on the basis of their reactivity with hydrazine. Other work by Bray and Cuy
N2H5, -- N2 + 5H+ + 4e-
Eo =+1.16V [2-1]
(1924); Cuy et al. (1924); and Kirk and Browne (1928), identified three basic
N2H 4e- + N2 + 5H+ [2-5]
N,H -> e- + N, + NH+ + H- [2-6]
N2H5 -- 2e- + I HN3 + NH + 2}H [2-7]
Higginson et al. (1953) grouped the oxidizing agents according to their
ability to oxidize hydrazine by one-electron or four-electron reactions. They
suggested that the ability of metal ions to adsorb hydrazine in their coordination
spheres determined the path of reaction. Later, Higginson and Sutton (1953),
Cahn and Powell (1954) and Higginson and Wright (1955) used 15N to verify
the oxidation mechanisms that had been proposed.
Cahn and Powell (1954) also did extensive work on the effect of cupric
ion on hydrazine. They noted that the cupric ion does not react appreciably with
hydrazine in acid solution, though its presence greatly increases the proportion
of four-electron oxidation.
Two-electron oxidation, yielding ammonia, was found to be greatest in
acidic solution (pH<3, Higginson, 1957). Subsequent work by Pagsberg (cited
by Sutherland, 1979) and by Adams and Thomas (1963), Atkinson and Bard
(1971), Smith et al. (1971), Haydon and Simic (1972), and Sutherland (1979)
confirmed the validity of the oxidation reactions and identified transient
intermediate reaction steps.
Autoxidation in Solution
Although the two adjacent nitrogen atoms of hydrazine should favor the
formation of nitrogen when the hydrogens are removed by oxidation, this simple
reaction is not usually the only path followed (Schmidt, 1984). In addition to
molecular nitrogen, the following compounds have been identified as reaction
products: ammonia, hydrazoic acid, diazene, and hydrogen peroxide.
Although they are thermodynamically favorable, especially in acidic solution,
there are few examples of such reactions in the literature. Hence, they are likely
very slow in the absence of an appropriate catalyst.
In 1924 Cuy and Bray examined the influence of pH and atmospheric
oxygen on the disappearance of hydrazine from aqueous solution. Hydrazine
solutions in 1.0 M sodium hydroxide were found to be unstable, though acidic
solutions were quite stable. Basic solutions kept under a nitrogen atmosphere
were not found to degrade. They assumed that decomposition was due to
oxidation in air. Gilbert (1929) also examined the effects of pH and oxygen on
hydrazine decomposition. He observed the formation of hydrogen peroxide in
dilute alkaline solutions in the presence of oxygen. Under an oxygen
atmosphere, sodium hydroxide concentrations above 0.03 M were seen to
correspond to a decrease in hydrogen peroxide formation, implying that
hydrazine autoxidation was optimal in dilute alkaline solution. Cuy and Bray
(1924) and Gaunt and Wetton (1966) also did not detect any hydrazine
degradation unless oxygen was present.
Ellis et al. (1960), attempting to determine the kinetics of hydrazine
degradation, found hydrazine to disappear faster than did oxygen. They
modeled the rate of hydrazine disappearance with a second-order empirical
dc ac + bc [2-8]
where a and b are functions of pH and temperature. The rate of reaction
increased by a factor of 1.2 per 10C rise in temperature. Ammonia was
detected when hydrazine was in excess, but no hydrogen peroxide was
detected. Autoxidation by atmospheric oxygen appears to be the most
important factor contributing to the disappearance of hydrazine in the
environment, since oxygen is in great supply.
Other research has shown hydrazine to be degraded in the absence of
oxygen. Gilbert's data (1929) suggested that hydrazine decomposition took
place on surfaces. The possibility of dust particles acting as the active surfaces
was mentioned, and the fritted glass he used to transmit oxygen was also
suspected. Brown (cited by Audrieth and Ogg, 1951) observed that traces of
copper exert a marked catalytic effect on the autoxidation of hydrazine.
Audrieth and Mohr (1951) tested several metallic ions for implication in the
catalysis of hydrazine decomposition. Dissolved copper was by far the most
active catalyst, followed by vanadium (VO-3). They used metal deactivators
which form insoluble salts or stable complexes with copper to inhibit the
catalytic effect, but no inhibitors were found that would totally neutralize the
effect of copper.
More recently, Lim and Fagg (1984) observed that manganese catalyzed
autoxidation of aqueous hydrazine. Schmidt (1984) also mentioned a survey
conducted by Eberstein and Glassman of metals that catalyzed hydrazine
decomposition. They noted that transition metals having incomplete d
subshells act as strong catalysts for hydrazine decomposition, whereas metals
having no dsubshells or completely-filled shells are not catalytic. It was
theorized that electrons from the lone electron pairs in hydrazine interact with
unfilled d orbitals during the early stages of adsorption and chemisorption of
At least two basic equations can be used to describe the heterogeneous
decomposition of hydrazine in solution:
3N2H4 N2 +4NH3 [2-9]
N2H4 -- N2+ 2H, [2-10]
Any possible combination of both reactions has been found to occur,
depending on the catalyst used and the experimental conditions (Maurel and
Menezo, 1978; Oosawa, 1984).
This discussion of oxidation/autoxidation in solution is directly related to
the decomposition of hydrazine in soils, because of the omnipresence of
aqueous solutions even in unsaturated soils. The focus of the experimental
data presented in this dissertation is on hydrazine fate and transport under
saturated conditions, so hydrazine reactivity in solution is of major consequence
Hydrazine Degradation in Natural Waters
Slonim and Gisclard (1976) studied the disappearance of hydrazine in
waters of different origin that varied in hardness, organic matter content, oxygen
content, OH, and temperature. Hydrazine at 5 mg L-1 was added to water
samples and analyses were done each day for five days. Within the first hours
the most polluted water (with the greatest amount of solids in suspension)
caused the greatest breakdown of hydrazine. Water from the same source but
taken under calm weather conditions showed no hydrazine breakdown initially,
though hydrazine was not detectable after four days. A correlation also was
found between the degree of hardness and the rate of hydrazine decay. In city
tap water which was softened and chlorinated, hydrazine concentration
remained approximately the same for 4 days. Slonim and Gisclard (1976)
stated that polluting material rich in organic matter was the leading contributor
to hydrazine degradation. However, they did not mention the possibility of
hydrazine adsorption to organic surfaces. Also, biological activity was not
considered as a possible factor in hydrazine degradation.
MacNaughton et al. (1978) studied sea water and pond water, evaluating
the effects of copper, dissolved organic, and oxygen concentration on
hydrazine behavior. Addition of copper at a concentration of 4x10-6 moles L-1
had a greater effect in sea water than in pond water. They suggested that the
copper was adsorbed to the greater dissolved organic fraction in pond water,
making it less available for subsequent catalysis. Filtered pond water resulted
in no change in oxidation rate, suggesting that suspended material was not
important in catalyzing hydrazine oxidation or complexation of the added
copper. No effect was observed, as well, by varying the oxygen concentration
between 0.5 and 40 mg L-1.
From the literature we can infer that the disappearance of hydrazine from
solution is highly dependent on reaction conditions. At low pH, hydrazine can
be oxidized (mainly to H2 and NH3) by many metals and other oxidizing agents.
At high pH, autoxidation also can take place, the main product being molecular
nitrogen. This reaction is pH dependent with metals, especially copper and
manganese, acting as catalysts.
Some information is available on the effects of hydrazine and its
derivatives, monomethylhydrazine and dimethylhydrazine, on soil
microorganisms under pure and mixed-culture conditions. Since the
hydrazines are nitrogen compounds and may be degraded to NH3, Kane and
Williamson (1983) chose nitrifiers and denitrifiers as test bacteria for the toxicity
of hydrazine and its derivatives. They found that the activities of the autotrophic
nitrifiers Nitrosomonas and Nitrobacter, as well as of denitrifying bacteria, were
inhibited by the three chemicals. Gas production by anaerobic methanogens
was also inhibited by hydrazine. Mantel and London (1980) and London and
Mantel (1983) demonstrated that hydrazines at low concentrations (10 to 50 mg
kg-1) exerted bacteriostatic and bactericidal effects, resulting in prolongation of
the lag phase of growth for the soil bacterium Enterobacter cloaca. At higher
concentrations (100 mg kg-1) the overall growth of the bacteria was inhibited
(London et al., 1983).
Because of the diversity of microorganisms in soils and the buffering
capacity of soil, it has been suggested that toxicity of the compounds to soil
microorganisms would not be as great as to microorganisms maintained in
liquid media (Hollocher et al., 1982). Since hydrazine can be degraded to N2
by autotrophic bacteria such as Nitrosomonas (Kane and Williamson, 1983)
and to NH3 by N2-fixing heterotrophic bacteria (Stiefel et al., 1978), it was
thought likely that hydrazine would be detoxified by these soil bacteria and
possibly by other hydrazine-degrading microorganisms as well.
The microbial degradation of hydrazine in soils and the effect of
hydrazine on soil microbial activity was investigated by Ou (1987, 1988) and Ou
and Street (1987, 1988). They reported that hydrazine applied to Arredondo
soil at concentrations of 0.256 to 1.28 mmol 100 g-1 completely disappeared in
less than 1 and 8 days, respectively. Hydrazine was not observed to be
metabolized to ammonia, which could serve as a nitrogen source for growth. By
comparing degradation rates in sterile and nonsterile soils, they concluded that
biological degradation was responsible for about 20% of the disappearance of
the chemical in Arredondo soil. They also reported that, at 0.256 mmol 100 g-1,
soil respiration and total bacterial and fungal populations were not inhibited by
hydrazine. However, at 1.28 mmol 100 g-1, total bacterial populations in soil
were reduced by the presence of hydrazine (Ou and Street, 1987).
Ou (1987) also reported that three heterotrophic soil bacteria had the
capacity to degrade hydrazine in mixed culture. One of these organisms,
Achromobacter sp., degraded hydrazine at 3.2 mmol L-1 concentration to N2
gas. The organism was not able to grow on hydrazine as a sole source of
nitrogen, however, suggesting that the metabolic process for hydrazine was
Ou and Street (1988) reported that monomethylhydrazine (MMH) was
microbially mineralized to CO2 in Arredondo fine sand. Ou (1988) reported that
two soil bacteria, Achromobacter sp. and Pseudomonas sp., accelerated the
degradation of MMH in culture media and soil samples despite the fact that they
could not utilize MMH as a sole source of carbon.
Soils contain a large variety of active surfaces that have the potential to
interact with hydrazine in a variety of ways. Furthermore, the ionic environment
surrounding soil colloidal particles is quite different from that in the bulk of the
solution. It is anticipated that the reaction taking place near particle surfaces
will be affected by this micro-environment.
Several authors have investigated the effect of added solids on the
autoxidation rate of hydrazine. Ellis and Moreland (cited by MacNaughton et
al., 1978) found that the reaction was dramatically accelerated by the addition of
activated carbon, copper sulfate, brick, or electrolytic carbon. MacNaughton et
al. (1978) also reported that additional surface area in the form of a-quartz,
alumina, or kaolinite did not increase the oxidation rate and, if anything, actually
caused a small reduction in the rate. The presence of small chips of concrete,
however, caused significant oxidation of the hydrazine. This result is in
agreement with findings of a rapid loss of hydrazine spilled on concrete
pavement during spill clean-up studies (Stauffer and Eyl, 1978).
Hayes et al. (1981, 1984, 1987), in an extensive study of the interactions
of hydrazine with soil constituents, found that degradation of hydrazine in the
presence of homoionic Na+-, K+-, Mg+2-, and Ca+2-montmorillonite occurred to
a greater extent than in the corresponding metal-chloride solutions. They
suggested that the higher pH values of the clay suspension contributed to the
enhanced degradation observed in these systems. They examined the effect of
copper in solution (CuC12, 320 mg L-1) and Cu-montmorillonite suspension on
hydrazine degradation, and found greater degradation in the suspension. They
attributed this to an increase in effective Cu+2 concentration at the clay surface
compared to that in the cupric chloride solution. Hydrazine would thus be
brought into closer contact with exchangeable Cu+2 ions, resulting in rapid
degradation in the supernatant solution.
From the data in the literature, several conclusions can be drawn:
* Hydrazine can be oxidized in the presence of clay irrespective of the cation
species on the exchange complex. In the case of ions that are not easily
reduced (K+, Na+, Mg+2, Ca+2), the amount of hydrazine degraded in the
suspension was identical for all cations.
* Hydrazine can also be oxidized by the cation on the exchange complex.
* Hydrazine can be adsorbed directly to clay by exchanging with cations from
the surface. The amount of sorbate held by montmorillonite saturated with
cations not in the transition series was similar for all except K+-
montmorillonite, which did not adsorb any hydrazine at all.
* Hydrazine can also be adsorbed to the clay by completing with the cation
on the exchange complex, with reduced iron in the clay structure, or with
electronegative groups on the surface of the clay.
* The pH of the suspension has a significant effect on the deg adation of
hydrazine in the supernatant solution.
Moliner (1988) examined the effect of Na-montmorillonite partially
saturated with Cu on hydrazine adsorption and degradation. The results then
were compared with the degradation of hydrazine in Cu+2 solutions containing
the same amount of Cu per unit volume. She reported that the clay had a
strong catalyzing effect on hydrazine degradation even when Cu was not
present. The presence of both Cu and clay in the suspension appeared to be
additive, and independent of one another. Degradation due to Cu in the clay
suspension was of the same magnitude as degradation in solution having only
one-tenth the amount of Cu added.
Supporting Moliner's observations, Hayes et al. (1984) reported that
degradation of hydrazine by homoionic clays was independent of the
exchangeable cation as long as the cation was not easily reducible. This
suggests that the clay itself was the active component in the oxidation or
catalytic autoxidation of hydrazine.
Determination of the extent of hydrazine adsorption by soil components
and stability of the complexes thus formed is critical to the study of the
compound's environmental fate. Adsorption depends on the reactivity of the
surface functional groups present on soil colloids (silicate clay minerals such as
montmorillonite and kaolinite, metal oxides and hydroxides such as goethite
and gibbsite, and organic matter such as humic and fulvic acids), and on the
chemical properties of the hydrazine itself.
Adsorption to silicate-clay materials is highly dependent on the physical
configuration of the colloids, and particularly on the charge imbalance in the
clay structure itself. If there is no cation isomorphic substitution, such as is the
case with kaolinite, the surface ditrigonal cavities of the silicate clay act like a
very soft Lewis base (Sposito, 1984), and are likely to complex only neutral
dipolar molecules like water or unprotonated hydrazine. Substitution in the
octahedral layer is able to form weak complexes with cations as well as with
Cation substitution also may occur in the tetrahedral layer, in which the
surface charge is more localized and distributes itself over the three surface
oxygen atoms associated with a single tetrahedron. This makes possible the
establishment of much stronger bonds with dipolar molecules and cations, as
well as the formation of inner-sphere complexes between the siloxane surface
and selected cations (Moliner, 1988).
Hayes et al. (1984) reported that the factors which most strongly
influence the adsorption and/or decomposition of hydrazine and
monomethylhydrazine in homoionically-exchanged montmorillonite clay
suspensions included solution pH, oxygen status, and nature of the
exchangeable cation. They concluded that, in the absence of dissolved
oxygen, the primary interaction of hydrazine with K+-, Ca+2-, Fe+3- ,and AI+3
-montmorillonite and -kaolinite was adsorption rather than decomposition. The
exact mechanisms for hydrazine adsorption to clay minerals were postulated to
be the replacement of protons from water molecules coordinated to adsorbed
cations at low solution pH, and protonation of the more basic hydrazine
molecule by protons from the more acidic coordinated water at high pH.
The second type of surface functional group present in most inorganic
colloids is the hydroxyl group. The broken bonds found at the edges of most
silicate clays create hydroxyl groups which may be coordinated to one or two
cations. This charge imbalance is pH-dependent, and inner- and outer-sphere
complexes can form between the aluminol and silanol groups and available
cations in solution.
Hayes et al. (1984) found that, in studies with hydrous oxides of Fe and
Al, there was evidence of binding and decomposition of hydrazine similar to that
from the silicate clay mineral study. Their results for goethite suggested that the
formation of soluble hydrazine-iron(ll) complexes at pH values less than 8.0
was the primary reaction.
Johnson et al. (1988),using non-invasive Raman spectroscopy and x-ray
diffraction, showed expansion of the kaolinite lattice upon intercalation by
hydrazine. FT-IR spectra indicated that strong hydrogen bonds were formed
between the intercalated hydrazine species and inner-surface hydroxyl groups
on the kaolinite interlamellar surface.
The third soil-colloid group, organic humus, has been observed to
provide a strongly sorptive surface for metals and polar molecules. A variety of
functional groups, including CO, COOH, phenolic OH, enolic OH, lactone,
quinone, hydroxyquinone, ether, alcoholic OH, amino-N, and sulfonic-S, have
been reported for humic substances (Stevenson, 1982). The adsorptive ability
of these groups is a function of the pH of the suspension and depends largely
on the stereochemical configuration of the molecule.
In a study with humic acid preparations at pH 4, Isaacson and Hayes
(1984) found that hydrated hydrazine was more extensively held by H+-
saturated humic acid than by Ca+2- or Al+3- saturated humic substances. This
reflected the greater ability of the hydrazinium ion to exchange with H+, and to
disrupt hydrogen bonding, than to disrupt the divalent and polyvalent cation
bridges between polymer strands. They reported that, in humic acid systems,
exchange by hydrazinium ions, chemisorption through interaction of hydrazine
with humate carbonyl groups, and non-specific sorption involving weakly and
strongly held hydrazinium ions and hydrazine molecules are the major sorptive
processes. They concluded by reporting that decreasing solution pH tends to
increase the importance of ion exchange, but decreases the contribution of
chemisorption in the binding process.
Due to the polarity of the N-H bond, hydrazine can form hydrogen bonds
with electronegative groups on the surfaces of organic matter as well as clays.
Davis et al. (1988), using diffuse-reflectance spectroscopy, found that the
primary surface-hydrazine interaction with silica, silica-alumina, and alumina
surfaces was hydrogen bonding.
Unprotonated hydrazine is a strong nucleophile that can take part in
condensation reactions with carbonyl groups in humic substances to form
hydrazone. This is the basis of the procedure used by Schnitzer and Skinner
(1965) to determine the concentration of carbonyl groups in soil organic matter.
Isaacson and Hayes (1984) showed that this reaction takes place even at pH
4.0. They also pointed out that the maximum rates should occur when the pH of
the media is close to the pka of hydrazine (i.e. near pH 7.95). The condensation
complex is also subject to hydrolysis, and hydrazine can take part as well in
substitution reactions at positions activated by carbonyl groups (Szabo et a.,
1978; Isaacson and Hayes, 1984).
Ion exchange, and particularly cation exchange in soils, is a reversible
process whereby cations held on the surface of soil minerals and even within
the crystal framework of a few mineral species plus those which are held by
certain organic species can be reversibly replaced by those of salt solutions
and acids (Chapman, 1965). This process is often grouped and discussed with
several other processes collectively known as "adsorption." Special notice of its
importance is given here because experimental evidence indicates that it plays
a major role in the retention of hydrazine by soils.
Soils generally possess a negative electrostatic charge of a permanent
or pH-dependent nature. As previously discussed, the permanent charge is the
result of isomorphous substitution within the structures of layer-silicate minerals.
Cations of lower valence are substituted for octahedrally or tetrahedrally
coordinated cations, resulting in a net negative charge. The pH-dependent
charge results from broken bonds at mineral edges and external surfaces,
dissociation of acidic functional groups on organic compounds, and the
preferential adsorption (by chemical reaction) of certain ions on oxide-mineral
surfaces. The magnitude of such charge is dependent upon solution pH,
electrolyte level, valence of the counter-ion, dielectric constant of the medium,
and nature of the anion in the solution phase.
The permanent charge also may be partially neutralized by strongly
adsorbed hydroxy-aluminum polymers that carry a net positive charge. As the
pH rises, these polymers are retained as partially neutralized AI(OH)3,
progressively freeing more negative sites for participation in normal cation
exchange reactions. Negative sites can be similarly neutralized by the
adsorption of positively charged mineral particles, such as hydroxides. The
positive charges of such particles originate from the specific adsorption of
protons on oxide/hydroxide surfaces, with their magnitude depending on the
ionic strength and pH of the solution. Such charge is generally neutralized at
pH >7 (Rhoades, 1982).
The permanent and pH-dependent charges generate a net excess of
negative charge on soils. This excess charge can bring about the formation of a
diffuse layer of positively charged ions about the mineral or humic colloid, with
the density of this layer being greater at the surface and then decreasing
exponentially to the level of the bulk solution. This type of reaction has
important implications in the adsorption of inorganic ions and ionized organic
molecules (Roy et al, 1987).
Published literature on the subject of hydrazine transport in soils is
extremely limited. Only one report (Braun and Zirrolli, 1983) of investigations
into hydrazine movement through soils could be located. That report and
extensive work by Hayes et al. (1984) on interactions of hydrazine with soil
components indicate that hydrazine is highly reactive in soil-water systems.
According to Hayes, hydrazine interactions in soil include at least four
* 1. In acidic soils, hydrazine is hydrolyzed to hydrazinium (N2H5+). This
reaction is at equilibrium at pH 7.96, the pKa of hydrazine. Hydrazinium
undergoes exchange with cations present on the soil surface.
* 2. Under alkaline conditions, hydrazine may be degraded by the process of
catalytic oxidation in the presence of such metals as Fe+3, Cu+2, AI+3, and
Mn+2. Degradation products are likely to include hydrazone, NH4+ ions and
* 3. The formation of hydrazine complexes with adsorbed cations on the
surfaces of clay minerals, oxides, and organic matter may provide an
environment for reversible adsorption mechanisms.
* 4. Condensation reactions may provide an irreversible chemisorption of
hydrazine by humic components of the soil.
Mathematical models for describing the movement of hydrazine through
water-saturated soil must include components for ion exchange, catalytic
oxidation, and reversible and irreversible sorption. The net effect of these
interactions would be to retard the migration of hydrazine-type fuels (Braun and
Zirrolli, 1983) through natural soils. If the soil is not water-saturated,
volatilization of hydrazine would be an additional process for inclusion in the
model. Since investigations of the more complex case of hydrazine transport in
partially water-saturated soil have not been reported in the published literature,
such research would logically follow the work presented here for saturated soils.
Many investigators have given valuable insight into mechanisms of solute
transport in porous media (Rao et. al., 1980; Parker and Jardine, 1986;
Bouchard et al. 1988; Konikow and Mercer, 1988; Selim and Amacher, 1988;
Barnes, 1989; Baveye and Valocchi, 1989; Wierenga and van Genuchten,
1989). Mathematical models for the movement of reactive chemicals in soil
generally assume that transport occurs primarily by mass flow as part of the
mobile soil solution; that displaced and displacing solutions undergo mixing
due to hydrodynamic dispersion; that solute movement may be retarded due to
processes such as ion exchange, reversible adsorption, irreversible
chemisorption, formation of chemical complexes, and chemical precipitation;
and that the solute itself may be altered by microbial degradation, chemical
degradation, etc. (Mansell et al., 1990). Chemical and physical kinetic
processes can be critical to the movement of reactive solutes through
aggregated soil as liquid flow velocity is increased. Mathematical treatment of
many of these mechanisms has been discussed by Nielsen et al. (1986).
The transport of hydrazine under water saturated soil conditions may be
expected to conform to the principles of advective-dispersive transport, with the
chemical and microbial reactions acting as retardation or degradation terms.
The mathematical development of solute transport under steady, saturated flow
begins with the assumption that the solute is partitioned between the solution
and adsorbed phases. The total mass of solute, M, per unit volume of soil is the
sum of the amounts in the solution and adsorbed phases,
M= (9C+pS), [2-11]
8 = volumetric soil-water content (cm3 cm-3 soil),
C = solution-phase concentration (molc cm-3),
p = soil bulk density (g soil cm-3 soil), and
S = mass of solute adsorbed per gram of soil (mole g-1 soil).
Substitution of equation [2-11] into the continuity equation,
M= J [2-12]
t x '
J, = solute flux (g s-1),
provides an appropriate generalized partial differential equation to describe
convective-dispersive transport of a reactive solute in porous media:
( C+ pS) = [2-13]
By expanding the solute-flux term in greater detail:
J, = (-D-+qC), [2-14]
-D- = dispersive flux (with D representing the hydrodynamic
dispersion coefficient (cm2 s-1)), and [2-15]
qC = convective flux (g s-1 cm-2), [2-16]
equation [2-13] becomes the nonlinear partial differential equation
S S SC S
-(C+ pS) =-( )D -(qC). [2-17]
St Sx Sx Sx
For the condition of steady, spatially uniform water flow in porous media, the
constants may be moved outside the partial differentials,
SC SS 52C SC
t + t = OD- q. [2-18]
St St x2 Sx
Equation [2-18] can be simplified by dividing through by 0. Thus,
= rate of change of the solution-phase concentration,
P08S = rate of change of the sorbed-phase concentration,
D-- = dispersive (mixing) transport term, and
q = vo = the average pore-water velocity which results in convective
In order to solve equation [2-18], a functional relationship must be
specified between S and C There are a number of ways to describe this
relationship, depending upon the physical process felt to be of predominant
importance. Observation of the adsorption isotherms and breakthrough curves
performed in conjunction with this investigation suggests that cation exchange
plays a major role in the fate and transport of hydrazine at the pH of these soil
To describe convective-dispersive transport of n cation species in water-
saturated soil during steady liquid flow, a coupled system of n nonlinear partial
differential equations must be solved for C,(x,t)
sc, p 3S, S= c sCi
+ = D -vo [2-19]
8t 0 8t Sx ox
C, = the concentration (mole m-3) of species i in the solution-phase,
Si = the concentration (mole m-3) of species i in the exchange-phase,
D = the hydrodynamic dispersion coefficient (m2 s-1), assumed
to be dependent on pore-water velocity,
v. = the Darcey flux (m s-1)
D(v) = D + Dv, [2-20]
Do = the molecular diffusion coefficient (m2 s-1), and
D, = the dispersivity (m).
The expression for the exchange-phase concentration (8S,/8t) in
equation [2-19] may be related to the solution-phase concentration using an
approach similar to that of Valocchi et al. (1981), as described by Mansell et
-S hi [2-21]
St g,) St g St
h = I + rn JL, [2-22]
jo c, sc ,s,
= 1+S r [2-241
r = the valence of ion species i, and
r = the valence of any additional species.
The Gaines-Thomas binary exchange selectivity coefficient (K,), used in
equations [2-22] and [2-23], expresses the preferential relationship between
solution- and exchange-phases, and is given by
K ..=2 = [2-25]
where S, and S, represent equivalent fractions of ions i and j in the exchange
phase. The total concentration in the exchange phase is assumed constant, and
y7 and y7 are the activity coefficients for ions i and j in the solution phase.
Valocchi et al. (1981) stated that inclusion of solution-phase activity coefficients
is only necessary for the description and prediction of cation exchange if ionic
strengths corresponding to experimental isotherms differ from those used in
A detailed numerical model which combines transient, unsaturated flow
and transport, including variable total solution concentration and binary
exchange selectivity coefficients that vary with total solution concentration and
with ion concentration within the solution phase, is presented by Mansell et al.
Most chemical fate and transport models are based on the assumption of
an instantaneous equilibrium established between solution-phase and sorbed-
phase solute concentrations. Such conditions are not always present.
Nonequilibrium, or rate-limited, sorptive processes have been well
documented, and have been grouped into two general classes: transport-
related and sorption-related (Brusseau and Rao, 1989; Brusseau et al.,1989).
Transport-related nonequilibrium, often referred to as physical nonequilibrium,
results from the existence of a heterogeneous flow domain. The influence of
macroscopic heterogeneities such as aggregates, macropores, and stratified
media on solute transport also has been well documented (Brusseau and Rao,
1989, 1990). Transport-related nonequilibrium affects both sorbing and non-
Sorption-related nonequilibrium may result from chemical
nonequilibrium or from rate-limited diffusive mass transfer. Chemical
nonequilibrium is caused by rate-limited interactions between the sorbate and
sorbent. Specific sorbate-sorbent interactions may be relatively unimportant for
charge-mediated sorption (ion exchange), since such interaction is thought to
be electrostatically driven rather than chemically mediated. Electrostatically
charged sorbates, however, are known to react with organic components of the
sorbent (Isaacson and Hayes, 1984), and rate-limited diffusive mass transfer
within the organic phase may occur.
Three different processes involving diffusive mass transfer can cause
sorption-related nonequilibrium (Brusseau et a.,1991): film diffusion, retarded
intraparticle diffusion, and intrasorbent diffusion. Researchers have shown that
film diffusion is generally insignificant in comparison to other mechanisms
(Brusseau and Rao, 1989), and thus will not be discussed further here.
Retarded intraparticle diffusion involves aqueous-phase diffusion of
solute within pores of granular soil material, and is mediated by instantaneous
sorption to particle walls (Wu and Gschwend, 1986; Ball et al., 1990). Work by
Chantong and Massoth (1983) estimated that the pore diameter required to
produce appreciable diffusive hindrance was approximately 25 nm or less. The
pore-size distribution of a sandy aquifer material as measured by mercury
porosimetry and nitrogen desorption by Ball et al. (1990) revealed that 80% and
greater than 90% of the internal pore volume comprised pores whose diameters
exceeded 25 nm and 10 nm, respectively. Brusseau etal. (1991) concluded
that, if these results are at all representative of other sandy materials, it would
appear that intraparticle diffusion may not be important for many solutes of
interest as well.
Intraorganic diffusion involves the diffusive mass transfer of sorbate
within the organic matrix of the sorbent. Intraorganic diffusion was proposed as
the limiting mechanism for sorption of organic chemicals as early as 1966 by
Hamaker et aL(1966), and has since been embraced by Brusseau and Rao
For the intraorganic diffusion model, the primary assumption is that
sorbent organic matter is a polymeric-type substance within which sorbate can
diffuse. The organic matter associated with natural sorbents has been reported
to be a flexible, cross-linked, branched, amorphous (noncrystalline),
polyelectrolytic substance (Hayes and Swift, 1978; Schnitzer, 1978;
Stevenson, 1982; Choudhry, 1983). Direct confirmation of the 'porous' nature
of organic matter has also been reported (Degens and Mopper, 1976;
Schnitzer, 1978). The conceptualization upon which the intraorganic diffusion
model is based is consistent with the generally accepted view of the process by
which sorbents are adsorbed by native organic matter (Brusseau et a., 1991).
METHODS AND MATERIALS
The experimental design of this study has as its purpose to identify the
transport processes applicable to hydrazine in water-saturated soils, and to
then quantify them through stirred batch suspensions and chemical analysis of
soil-column effluent. Pertinent soil characteristics were determined, and the
effects of solution concentration, water velocity, and time of hydrazine exposure
to soil were evaluated.
Three sequential horizons from a profile of coarse-textured soil were
obtained and characterized as to particle-size distribution, organic carbon
content, elemental composition, and mineralogy. Stirred batch suspensions
were used to determine soil buffering capacity, pH, and cation exchange
capacity. The influence of the organic fraction of the soil was evaluated by
comparing soils with and without appreciable organic carbon content. Rather
than remove the organic carbon by an oxidative process which might damage
microsurfaces, samples of three sequential horizons of the same soil were
obtained, each containing successively less organic carbon.
Flow experiments using saturated soil columns were performed to
determine the dispersion coefficient for saturated flow in each horizon and to
evaluate the effect of changes in pore-water velocity and hydrazine
concentration on hydrazine retention.
Two steady-state solute flux rates were evaluated, 0.5 cm h-1 and 5.0 cm
h-1. These fluxes correspond to Darcy velocities of 1.39x10-4 cm s-1 and
1.39x10-3 cm s-1 and to pore-water velocities of approximately 3.8x10-4 cm s-1
and 3.8x10-3 cm s-1, respectively, assuming a soil porosity of 0.37.
Three influent concentrations of hydrazine were evaluated for each fluid
flux. These were prepared as low (approximately 0.02 mmol L-1), medium
(approximately 6.0 mmol L-1), and high (approximately 20 mmol L-1) as
hydrazine hydrate. Column effluent fractions were analyzed for hydrazine,
cations and, in some cases, pH in order to provide data for analysis.
Data analysis was done by numerically integrating components of each
breakthrough curve for component mass. Experimental parameters and column
influent components were occasionally altered to more easily isolate the
effluent fraction in order to identify and quantify significant processes. The
timing of breakthrough and relative position of breakthrough components also
were examined, to prioritize the influence of various fate and transport
Samples of the Ap, El, and E2 horizons of an Arredondo fine sand were
obtained from a site in NW Alachua County, Florida, 0.4 miles east and 0.6
miles north of the intersection of State Roads 241 and 222, and 0.2 miles south
of a private paved road. Arredondo fine sand is classified as a loamy, siliceous,
hypothermic, Grossarenic Paleudult (Thomas et al., 1985), and is typical of the
well-drained soils of Florida. The Ap horizon at the collection site extended
from the surface to a depth of 20 cm, with the El horizon then extending to 80
cm, and the E2 horizon to 120 cm. The horizons were clearly distinguished
visually from one another in the soil profile. Samples of each horizon were
taken using a clean shovel from sufficiently far from the horizon boundaries to
preclude contamination from above or below. Separately, soil materials were
sieved through a 2-mm mesh screen, spread on a tray, air-dried for three days,
mixed, and stored in 3-gallon plastic buckets prior to use.
The distribution of particle sizes in a soil matrix has a significant effect on
the retention of water and chemicals by the soil. Coarse-textured soils high in
percent sand tend to retain water ineffectively, and are relatively non-reactive
chemically when compared to soils higher in silt and clay content.
Particle-size analysis for mineral components was performed by the
pipette method of Gee and Bauder (1986). Organic matter was removed from
the Ap horizon prior to mechanical analysis by oxidation in 5% sodium
hypochlorite (bleach). Samples from each horizon were suspended in distilled
water and dispersed with sodium hexametaphosphate. The supernatant was
decanted and allowed to settle in a constant-temperature water bath from which
aliquots were removed by pipette at a depth and time corresponding to the
settling velocity determined by Stoke's Law. Samples were dried and weighed
to determine percent clay. Remaining material was washed, dried, and filtered
through 16-, 32-, 60-, 150-, and 325-mesh U.S.A. Standard Testing sieves, and
weighed to determine the various sand fractions. Percent silt was determined
by subtracting the weights of combined sand and clay fractions from the total.
Samples were analyzed in duplicate and averaged for reported values.
X-ray diffraction analysis was performed to determine the principal mineral
species in each soil horizon. Approximately 500 grams of each horizon were
wet-sieved through a 0.0017 mm screen to remove sand particles, and
approximately 100 cm3 chlorox was added to the Ap horizon filtrate to oxidize
soil organic material which would interfere with the x-ray diffraction process.
After two days of oxidation time, 30 cm3 of 0.5 N HCI was added to flocculate
the clays. The suspension was then let stand for one day, and centrifuged at
16,000 rpm for six minutes. The centrifugation process was repeated six times,
each time collecting the supernatant and resuspending the sediment in distilled
water. Approximately 100 cm3 of saturated NaCI was then added to flocculate
all clay materials. An aliquot of clay suspension was placed on porous ceramic
tiles, and 1 N MgCI2, KCI, and/or glycerol were added to the tiles to allow
differentiation of kaolinite from smectite clays.
Organic Carbon Content
The soil organic fraction consists of the cells of microorganisms, plant and
animal residues in various stages of decomposition, stable humus synthesized
from residues, and highly carbonized compounds such as charcoal, graphite,
and coal (Nelson and Sommers, 1982). Determination of the amount of organic
material present in a soil is very important, since many groundwater
contaminants including hydrazine react with organic materials (Isaacson and
The percentage of organic carbon was determined for each soil horizon
by dry combustion in an induction furnace (LECO Model No. 523-300) following
the procedure of Nelson and Sommers (1982). Weighed samples were placed
in a ceramic crucible with iron and copper metal accelerators added. Samples
were heated inside an enclosed combustion tube through which oxygen was
passed. All of the carbon in the samples was oxidized to C02, small particles
were removed in a dust trap, and sulfur was absorbed in a sulfur trap, leaving
only C02 and oxygen. The CO2-oxygen volume was measured in a burette
held at constant temperature and corrected for pressure. The mixture then was
passed through a solution of KOH in another vessel, which absorbed all of the
CO2. The residual oxygen was brought back to the original burette, and the
volume of C02 determined by subtraction from the previous volume. Four
samples from each horizon were analyzed and averaged to give the reported
An analysis of the calcium, aluminum, magnesium, iron, sodium, and
potassium contents of the three soil horizons was made after acid extraction,
using flame atomic adsorption spectroscopy according to the procedure of
Baker and Suhr (1982). Samples of approximately 5 g each were placed in 50
cm3 polysulfone centrifuge tubes into which 20 cm3 of 0.01 M HNO3 was added.
The tubes were mechanically shaken for 4 hours at low speed, then centrifuged
for 10 minutes at 10C at 10,000 rpm with a 2,000 rpm per minute acceleration
rate. Following centrifugation, the supernatant in each tube was decanted into
an acid-washed glass scintillation vial and analyzed on an atomic adsorption
spectrometer (Perkin-Elmer Model No.460). All samples were run in duplicate.
Elemental standards were prepared from stock solutions and diluted until linear
in response over the sample range tested.
Elemental concentrations in oxide form were calculated by converting the
elemental concentration into molar form, and then stoichiometrically adding the
proper molar amount of oxygen. Oxide weight percentages were determined as
milligrams of elemental oxide per milligram of soil times 100.
In an acidic environment hydrazine (N2H4) is hydrolyzed to hydrazinium
(N2H5+). This reaction is at equilibrium at pH 7.96, the pKa of hydrazine. That
is, at pH 7.96 there are equal proportions of hydrazine and hydrazinium
present. The protonated hydrazinium molecule at lower pH may undergo ion
exchange reactions on soil particle surfaces, having a potentially significant
impact on the transport process.
The pH of each soil horizon was determined in a 2:1 (v:w) suspension of
0.01 N CaCI2, according to the method of McLean (1982). The pH
determination was made using a glass-calomel electrode (Ross combination
pH electrode No. 8103) on an Orion meter (No. 601A), in triplicate.
The ability of the three horizons of Arredondo fine sand to resist changes
in pH was measured by preparing titration curves for each, using Ca(OH)2. Five
grams of soil and 25 cm3 of approximately 0.01 N Ca(OH)2 were added to a
beaker and allowed to stand for 3 minutes (with stirring) before the pH was
read. Ca(OH)2 was then added in 0.1 ml increments, allowed to equilibrate with
stirring, and the pH again was noted. The true normality of the Ca(OH)2 was
determined to be 0.0084 N by titration with 0.01 N potassium phthalate.
Cation Exchange Capacity (CEC)
The Soil Characterization Laboratory at the University of Florida
performed an extractable cation analysis on Arredondo fine sand which allowed
an estimate of its CEC (Thomas et al., 1985). Extractable bases (Ca, Mg, Na,
and K) and extractable acidity were summed to give a total of 6.27 cmolc Kg-1
for Ap horizon soil, 3.42 cmolc Kg-1 for El horizon soil, and 2.29 cmolc Kg-1 for
the E2 horizon.
This analysis was confirmed for our samples from the plateaus of the
adsorption isotherm obtained by plotting the amount of K+ adsorbed against the
amount added in exchange with Ca+2. Two grams of soil were placed in a
polysulfone centrifuge tube along with 10 ml of 0.01 N CaCI2 and shaken gently
for four hours, then centrifuged for 10 minutes at 10,000 rpm, and decanted.
Dilutions to 0.1 N were made from a stock solution of 0.1 N KCI, and the pH of
each solution was adjusted to the pH of the horizon with which it would be used.
Ten milliliters of each dilution were placed in a centrifuge tube containing the
two grams of drained soil, mixed on a vortex stirrer, and shaken gently for four
hours. The tubes were again centrifuged for 10 minutes at 10,000 rpm, and the
supernatant was analyzed for potassium. The decrease in potassium in the
supernatant was considered to be due to that adsorbed onto the soil surface,
and the plateau of the plot of adsorbed potassium versus potassium added was
considered to reflect the exchange capacity. The CEC determination was done
A second approach to the determination of CEC was performed using a
colorimetric measurement of methylene blue adsorption, as described by Soon
(1988). Two grams of soil were weighed into a 250 cm-3 Erlenmeyer flask, 50
cm-3 of 0.5 mM methylene blue solution (buffered at pH 6.8 in 50 mM sodium
acetate) was added, and the flask and contents were allowed to settle for two
hours. A 0.25 cm-3 aliquot of the supernatant solution was then pipetted into a
test tube containing 12.25 cm-3 of distilled water, and mixed. Standards were
prepared containing 0 to 0.5 cm-3 of 0.5 mM methylene blue solution in a final
volume of 12.5 cm-3. Transmittance was measured at 550 nm in an optically
clear test tube using a Coleman 54B spectrophotometer. A straight line was
fitted through the standard curve (R2=0.997), and the equation of the line was
used to convert measured transmittance to concentration.
The adsorption of hydrazine onto Arredondo soil was evaluated by
exposing samples of each horizon to incremental concentrations of hydrazine,
and then analyzing the solution for hydrazine loss. An assumption inherent in
determining adsorption isotherms is that loss of the sorbate from solution is a
valid measure of adsorption. However, this assumption may not be valid with
hydrazine, given its reactive nature. Studies by Moliner and Street (1989a)
indicated that, in aqueous systems with 02 present, hydrazine may undergo
autoxidation. This was especially true when a catalyst such as Cu+2 was
present. Bott and Rassoul (1970) suggested that there is no decomposition of
hydrazine in the absence of oxygen in contact with polyethylene,
polypropylene, or Pyrex glass. On the other hand, polyvinyl cholride (PVC)
interacted with hydrazine and was considered an unsuitable material for
containing hydrazine solutions. Hydrazine may also react with metals such as
Fe+3 and Mn+3, which are widely present in soils, reducing them to lower
valence states (Griffeth et al., 1980). Additionally, there may be hydrazine
losses due to volatilization and degradation.
Other than these losses, the remaining possibilities for hydrazine
disappearance from the supernatant include ion exchange and sorptive
reactions, both reversible and irreversible. It is the combined effect of all these
reactions which contributes to the proper interpretation of an adsorption
All adsorption isotherms described here were obtained in an anaerobic
glove box to eliminate hazards associated with the potential autoxidation of
Adsorption isotherms were performed on each of the three horizons of
Arredondo fine sand, with two sets of isotherms being measured. In the first set,
measured at pH 4.8 and 8.0, five-gram samples of each Arredondo soil horizon
were placed in glass serum vials and washed five times with 0.1 N CaCI2
(maintained at the soil pH) to saturate the exchange complex with Ca+2.
Samples were shaken, equilibrated overnight, centrifuged, and the supernatant
Ten cm3 of constant ionic strength solution with increasing hydrazine
concentration were added to each of the Ca+2-saturated soils The pH of the
hydrazine solution was adjusted with HCI or Ca(OH)2. Because N2H5+ Cl-
contributes to the solution salt content, its concentration was also taken into
account when preparing the solutions of constant ionic strength. After
incubation for 48 hours in an anaerobic glove box, samples were centrifuged
and hydrazine was measured in the supernatant.
The second set of isotherms was conducted at pH 4.0 and 8.0. Twenty
grams of soil were washed five times with 0.1 N NaCI at the desired pH to
saturate the exchange complex with a single cation. The procedure followed
was then identical to that described above. Afterwards. the soils were extracted
with 0.1 N KCI, and then with 0.1 N HCI.
Information from the isotherms was used to determine selectivity
coefficients and ratios of exchangeable hydrazinium to calcium in soil columns
under equilibrium conditions. The method used to obtain those coefficients is
described as follows:
At equilibrium, the relative proportions of hydrazinium and calcium on
exchange sites are determined by ionic concentrations, valence, and solution
normality. The Gaines-Thomas binary exchange selectivity coefficient (K,) may
be used to express this preferential relationship (Valocchi et al. ,1981):
K. =- S [3-1]
S( ) r c,) c
where S" and Sj represent the equivalent fractions of ions i and j on the
exchange phase, and C, and C, are the solution-phase concentrations. The
total quantity of sorbed phase is assumed constant, and y7 and are the
activity coefficients for ions i and j in the solution phase. Valocchi et al. (1981)
stated that inclusion of solution-phase activity coefficients is only necessary for
the description and prediction of cation exchange if ionic strengths
corresponding to experimental isotherms differ from those used in actual
transport experiments. Since the background ionic strengths for the exchange
isotherms and the transport experiments in this work were the same, y, and
y,were set to unity. However, had the ionic strengths not been equal, the ratio
of y, to y, could have been incorporated into the value of K. The relationship
between the solution-phase and sorbed-phase concentrations may be clarified
by rearranging equation [3-1]:
The equivalent fractions of ions i and j in the solution-phase, C* and C, are
readily determined since the total solution concentration, CT (the solution
normality) is known, and the ionic solution-phase concentrations of interest, Ci
and Cj are readily measured:
The denominator of equation [3-2] then becomes
(CTc)" = CT (F
K,, can thus be written as
S= CT- (CI)'(C)( [3-8]
The Rothmund-Kornfeld binary exchange equation (Bond and Phillips,
= k V" [3-9]
is an empirical expression which provides a valuable mathematical means for
incorporating the characteristic shape of measured binary exchange isotherms
into a functional description of selectivity coefficients across a range of solution
concentration values (Mansell et al, 1993).
To determine the coefficients k and n the logs of both sides of the
equation [3-9] can be taken:
log T = k + n log i [3-10]
and the result regressed as log r against log p. Values for r and V are known
from information in the sorption isotherm (C, and S,). The intercept is k and
the slope n at the normality at which the isotherm was acquired.
When equation [3-9] is substituted into equation [3-6], sorbed-phase
concentrations drop out and K, can be expressed n terms of Cr and CT :
This expression shows that nonunity values for the Rothman-Komfeld
parameter n allow the selectivity coefficient Ki, to vary with local solution
concentration (C,) in the soil and with normality (C,), if ion valences are not
equal. When n =1, K, becomes a constant (k) for a given solution normality.
Preliminary Column Studies
Glass columns packed with soils from each of the Ap, El, and E2 horizons
of Arredondo fine sand were used in the laboratory to examine the fate and
transport of hydrazine under saturated soil conditions. Soil characteristics were
determined through slurry studies and soil extraction and analysis. A number of
these studies (elemental composition, pH, particle-size distribution, organic
carbon analysis, etc.) have been described previously. Transport
characteristics were studied by packing the soil into a column in such a way as
to imitate its natural configuration, with fluid designed to simulate the aqueous
soil solution then being pumped through the column at natural flow rates.
The preliminary studies reported here were designed to determine how to
best bring the soil columns to a steady-state operation, mimicking natural
conditions prior to the addition of hydrazine. The packing, wetting-up process,
saturation, deoxygenation, and measurement of hydrodynamic dispersive
characteristics of the wetted columns are each described in turn.
Glass chromatography columns 26.8 or 27.8 cm long by 5.08 cm i.d.
(Kontes No. 420800-3020) were hand-packed by sequentially adding
approximately 80 g of soil to the column and tamping with a plastic rod to a
maximum resistance (100 tamps, determined by prior experience to yield a bulk
density approximately of 1.6 g cm-3) before adding another 80 g. This
procedure was carefully followed when packing all columns for each horizon.
Column bulk densities were calculated by dividing the actual weight of soil in
each column by the column's volume.
Column porosities (r7) were estimated from the bulk density:
S= 100-p) [3-12]
p, = soil bulk density, and
a soil particle density of 2.65 g cm-3 was assumed.
The saturated column water content was calculated by dividing the weight
of liquid in a column by the volume of the column, assuming a liquid density of 1
The percent saturation of each column was determined by dividing the
water content by the porosity.
Ground water containing hydrazine is likely to be anoxic due to reduction
of 02 by the hydrazine, so packed soil columns were deaerated by introducing
a flowing stream of helium (or, in later experiments, nitrogen) into the bottom of
each capped column for two hours prior to saturation. Thus, oxygen originally
present was displaced by nonreactive gas to prevent oxidation in the column
and to be more representative of environmental conditions..
The soil columns were saturated from the bottom using deaerated CaCI2..
A CaCl2 solution was used to approximate the ionic solution of natural ground
water, which is dominated by the calcium cation in Florida. The CaCI2 solution
was prepared at 0.01 N, and deaerated by bubbling helium (or, in later
experiments, nitrogen) from compressed gas tanks through a 3-cm sparger into
continuously stirred flasks. The dissolved oxygen content of the CaCI2 solution
was monitored by the Winkler technique (Clesceri et al., 1989), and a standard
procedure was established to deaerate a new carboy of CaCI2 for at least four
hours prior to use (Figure 3-1).
0 50 100 150 200 250 300 350
Figure 3-1. Deoxygenation of CaCI2 influent solution.
Influent CaCl2 solutions were acidified to the appropriate horizon pH by
adjustment with HCI. Columns were filled from the bottom and allowed to
continually flow for at least 24 hours. Soil water content was determined by
weighing each column before and after saturation.
Liquid material was transferred from unpressurized flasks to the soil
columns using a Gilson peristaltic pump through small-diameter Tygon tubing.
Constant pumping rates (Q) of 101.2 and 10.12 cm3 h-1 were used. These rates
correspond to Darcy velocities (q = Q/A) of 5.0 and 0.5 cm h-1 through a
completely saturated soil column, and represent a range typical of flow rates
expected in-situ for fine sands of North Florida. Flow-rate adjustment was made
initially by adjusting pump speed while collecting column effluent in a
Short-term transport of water and soluble chemicals through saturated soil
is dependent on both physical and chemical processes. The chemical
processes act to retard and/or transform the chemical in solution as it is moved
down-gradient under the influence of the physical processes. Hydrodynamic
dispersion is an important physical mixing process which occurs due to
diffusion gradients and velocity distributions among soil pores. Dispersion
coefficients at Darcy velocities of 0.5 cm h-1 and 5.0 cm h-1 were determined
using the derivation of Kirkham and Powers (1972) applied to data obtained by
passing a pulse of tritiated water (3H20) through soil columns of each horizon.
Approximately two pore volumes of tritiated water diluted to approximately
10,000 counts per minute in 0.01 N CaCI2 were pumped through separate soil
columns at the low and high flow rate, respectively. Columns had been
saturated with deaerated 0.01 N CaCI2 prior to each introduction of tritiated
water. Effluent fractions were collected at nine-minute intervals (13 tubes per
pore volume, 30 cm3 tubes) for the high flow rate and at two-hour intervals (10
tubes per pore volume) for the low flow rate. Small aliquots from each fraction
were suspended in scintillation fluid (Scintiverse II), shaken, and counted in an
automatic liquid scintillation counter. Background counts were determined by
counting aliquots of CaCI2 which had not passed through the soil column, and
were subtracted from the effluent fraction count before plotting. This procedure
was duplicated at each flow rate for each horizon.
Hydrazine Column Investigations
Miscible displacement studies were performed using soil columns,
pumping various concentrations of influent hydrazine solution for either a pulse
or continuous duration through hand-packed columns of soil. Effluent fractions
were collected and analyzed for hydrazine, calcium, pH, and other components
of interest. Graphic displays of results were examined by plotting the relative
concentration ratios of effluent (C) and influent (Co) concentrations against the
number of pore volumes. A diagram of the equipment configuration utilized for
the miscible displacement experiments is shown in Figure 3-2.
Influent Hvdrazinium Solutions
Influent hydrazinium solutions were prepared at low (approximately 0.2
mmol L-1), medium (approximately 6 mmol L-1), and high (approximately 20
mmol L-1) concentrations as hydrazine hydrate (m.w. 50.06) in 0.01 N CaCI2 for
use in the miscible displacement studies. Preliminary column work had shown
that 20 mmol L-1 pulses of hydrazine were concentrated enough to overwhelm
most soil sorption sites as well as any sorption/degradation processes in the
topsoil, while 6 mmol L-1 appeared to give definable results, and 0.2 mmol L-1
solutions were severely retarded/degraded by the topsoil. The pH of the
hydrazine solution was adjusted to the pH of the soil horizon in use by the
addition of HCI, since each of the three horizons was acidic. Acidification
protonated the hydrazine molecule to hydrazinium (N2H5+) cations, with the
relative proportions of hydrazine and hydrazinium then being determined at any
given pH by knowing the pKa of hydrazine. Thus, the hydrazine solution is
found to exist as approximately 99.9 % hydrazinium in the pH range (4.46 to
5.13) for the soils used.
Hydrazinium solutions at the two flow rates were pumped through the
columns as either an approximately two pore-volume pulse or a continuous
step-function input. Pulse inputs followed by the acidified 0.01 N CaCl2 solution
allowed an observation of both the ascending and descending limbs of the
breakthrough curve, which gives insight about sorption and desorption
processes. Breakthrough curves following continuous input provide information
in turn about the irreversible processes of chemisorption and degradation.
Column Effluent Collection
A fraction collector (TRIS Retriever II) was positioned to collect solute
emerging from small-diameter Tygon tubing connected to the top of the soil
column. During the high flow rate (5.0 cm hrl) studies, glass test tubes 10-mm
in diameter (20 cm3 capacity) were rotated under the emerging effluent at nine-
minute intervals, collecting approximately 15 cm3 of effluent per tube, or
approximately 13 tubes per pore volume of solute. One cm3 of 1 N HCI was
added to alternate test tubes to ensure acidic conditions, thus stabilizing the
hydrazinium molecule against further oxidation. Analyses for hydrazinium were
performed on the acidified effluent fractions, while non-acidified aliquots were
examined for pH and calcium. The pH determination was made using a glass-
calomel electrode (Ross combination pH electrode No. 8103 on an Orion meter,
No. 601A) as effluent fractions were collected. In later column studies an in-line
flow-through cell and pH probe (Cole-Parmer No. L-05662-49) was connected
to the column effluent line near the top of the column, interpreted on a calibrated
Orion Model EA940 pH meter, and recorded on a Varian strip-chart for analysis.
For the low flow rate (0.5 cm h-1) studies, 12-mm diameter test tubes (25
cm3 capacity) were used to collect fractions at 2-hour intervals. Each tube
contained approximately 20 cm3 of effluent, or about 10 tubes per pore volume.
Again, one cm3 of 1 N HCI was added to alternate test tubes, and non-acidified
tubes were examined for pH and calcium.
Hydrazine analysis was performed using a modification of the method of
Watt and Crisp (1952). Small aliquots of collected fractions were placed into
25-cm3 volumetric flasks along with 15 cm3 of 4-dimethylaminobenzaldehyde
(PDBA) solution. Hydrazine reacts with PDBA to form an intense orange color
which is proportional to the concentration of hydrazine present. The solution
was diluted and stabilized by the addition of 1 N HCI to bring the volume up to
25 cm3. Color intensity was read on a spectrophotometer (Coleman 54B) as
percent transmission, which was then converted to absorbence. Incremental
dilutions of the hydrazine stock solution were read at the time of the column
effluent fractions and used as the standard curve from which to interpolate
hydrazine concentrations. The data were entered on a computer spreadsheet,
and a linear regression was used to fit a straight line through the standard
curve. Only fits with R2 > 0.98 were accepted for interpretation. The equation of
the fitted line was used to calculate the hydrazine concentrations in the various
Calcium analysis was performed by atomic adsorption spectrometry on a
Perkin Elmer flame spectrophotometer Model No. 460. Effluent samples were
diluted 1 to 200, and absorbence was determined using a nitrous oxide flame.
Interpolation of calcium concentration was done from a standard curve made
from dilutions of a standard stock solution analyzed at the same time as the
samples. The standard solution was maintained under refrigeration in a Teflon
vial, and regularly compared to pure hydrazine solution for consistency.
Analysis of other elements in the column effluent (aluminum, potassium,
and sodium) was also performed on the atomic adsorption spectrophotometer,
using acetylene or nitrous oxide flame as appropriate. Interpolations of
elemental concentrations were again made from a standard curve prepared
from dilutions of a standard stock solution analyzed along with the effluent
Analysis of the data obtained from each column effluent fraction was
accomplished using a computer spreadsheet. The relative concentration of
each fraction was determined by dividing its measured concentration (C) by the
initial input concentration (Co) to establish a relative scale from 0 to 1. Analysis
of initial concentrations was done on solute saved from flasks containing input
These relative effluent-fraction concentrations were then plotted against
number of pore volumes. Pore volume is the volume of liquid contained in a
saturated column (determined by weighing the dry and saturated column,
assuming a liquid density of 1.0 g cm-3). The cumulative volume of effluent was
divided by the pore volume to establish the abscissa of the graph. This plot is
known as a breakthrough curve (BTC), and reveals important information about
the dynamics of physical and chemical interactions within a column.
The mass represented under portions of the breakthrough curves was
evaluated by numerical analysis of the spreadsheet data, with the trapezoidal
rule being used to integrate the area under portions of the breakthrough curve.
The computer spreadsheet lends itself well to this kind of analysis, since the
data may be entered by row and column. A trapezoidal rule expression can be
written on the spreadsheet to evaluate selected portions of the data.
Mass balances were computed for hydrazinium added to and detected in
the effluent from each soil column. Mass input was determined by the weight
difference of the flask containing hydrazinium input solution before and after the
column experiment, multiplied by the hydrazinium concentration (assuming a
solution density of 1.0 g cm-3). Hydrazinium mass out was determined by
multiplying the concentration associated with each trapezoid under the output
curve by the effluent volume of the fraction collected, and summing over all
trapezoids. The difference between hydrazinium input and output was
assumed to be either adsorbed to the soil or degraded. Soil from a completed
column experiment was also exposed to the PDBA colorimetric detector
solution, and was observed to turn the orange color characteristic of PDBA
reaction with hydrazinium. However, a quantitative analysis of residual
hydrazinium was not possible.
At the initiation of this work, no information was available on microbial
degradation of hydrazine in soils. Throughout the duration of this study, several
checks were made to observe any influence of microbial degradation in the soil
The column studies reported here were performed with either pulse or
continuous-duration input of aqueous hydrazine solutions. One purpose of the
continuous input was to observe a rate-controlled degradation process which
might be operative after all sorptive requirements were met.
Additionally, plate counts and direct acridine orange (A-O) counts were
made of microbial biomass within completed soil-column experiments. Plate
counts are an estimation of the number of viable cells able to reproduce on the
plate-culture media. Approximately 2.5 grams of soil obtained from a location
near the center of the soil column were diluted with 100 cm-3 of distilled water
and plated onto tryptone broth agar media following the procedure of Wollum
(1982). Tryptone agar is a general-purpose growth medium on which most
microorganisms will develop. Agar plates were incubated overnight or until
observable colony growth was noted at 24C.
A-O counts give a total microbial estimation, living or dead, of
microorganisms which will adsorb the acridine orange stain. The technique is a
microscopic direct-counting method in which 2.5 grams of soil are obtained,
diluted in 0.1% sodium pyrophosphate, fixed in 5% Noble agar solution, placed
in a 1 cm-2 depressed circle on a microscope slide, stained with 0.01% acridine
orange, and counted under a phase-contrast microscope (Trolldenier, 1973).
Mineral components of the three upper horizons of Arredondo fine sand
were found to consist of a predominant sand fraction and relatively small
percentages of silt and clay (Table 4-1). From the standpoint of particle size, the
El and E2 horizons are more similar to one another than they are to the Ap
horizon. The 2.6 percent clay and 7.3 percent silt fraction in the Ap horizon set it
apart as somewhat different from the two lower horizons.
Table 4-1. Particle-size distribution.
Horizon % Sand % Silt % Clay
VC C M F VF Total Total Total
(2-1 mm) (1-.5) (.5-.25) (.25-.1) (.1-.05) (2-.05) (.05-.002) (<.002)
Ap 0.0 3.0 21.6 57.3 8.2 90.1 7.3 2.6
El 0.0 1.5 21.6 54.2 16.1 93.4 4.9 1.7
E2 0.0 3.3 29.6 50.1 11.5 94.5 3.7 1.8
X-ray analysis of clay films on the ceramic tiles revealed peaks at angles
corresponding to the d-spacing of kaolinite. No smectite clays or significant
amounts of oxide minerals were found.
The finding of kaolinite as the dominant clay mineral in this soil has
important implications for the adsorption of charged ions such as hydrazinium.
Most of the surface functional groups of kaolinite consist of inorganic OH
groups, which may be coordinated to one or two cations. The charge is found
predominantly at the edges, arising from broken bonds, and is pH-dependent.
At these broken bonds, hydrazine could replace other cations from exchange
sites under acidic conditions. Under alkaline conditions, the siloxane ditrigonal
cavity on the outer planer surfaces of the kaolinite particles also would be
available for hydrogen bonding, thus allowing the adsorption of hydrazine at
sites not previously occupied by a cation (Moliner, 1988).
Organic Carbon Content
The analysis of Arredondo fine sand revealed that successively deeper
soil horizons contained less organic carbon. The topsoil, or upper horizon, was
found to contain 1.84 percent organic carbon, compared to 0.34 and 0.14
percent for the El and E2 horizons, respectively (Table 4-2). While all of these
percentages are low, it is significant to note that the Ap horizon contains
approximately five and a half times as much organic carbon as the underlying
The dry combustion method described here determines total carbon
present in the soil, with total carbon being the sum of both organic and
inorganic carbon. Inorganic carbon is found in carbonate materials such as
calcite, dolomite, and soluble carbonate salts, and is not generally found in
well-leached soils of low pH (Nelson and Sommers, 1982) such as north
Florida Arredondo fine sand. In such acid soils, total carbon content can
generally be considered equivalent to organic carbon content.
Table 4-2. Organic carbon percentages.
Ap El E2
1 1.02 0.41 0.07
2 1.38 0.30 0.18
3 2.44 0.36 0.09
4 2.52 0.29 0.22
Average 1.840.75 0.340.06 0.140.07
Organic carbon has been observed to contribute significantly to the
adsorptive capacity of soils. A variety of functional groups, including carboxyl,
phenolic OH, enolic OH, lactone, quinone, hydroxyquinone, ether, alcoholic OH,
amino, and sulfonic, have been reported on humic substances (Stevenson,
1982). The ability of these groups to complex metals and polar molecules like
hydrazine is typically a function of the pH of the suspension and depends in part
on the stereochemical configuration, of the molecule.
Schnitzer and Skinner (1965) utilized the reactivity of hydrazine with the
carbonyl groups of humic substances as a procedure for determining the
concentration of carbonyl groups in soil organic matter. Isaacson and Hayes
(1984) showed that this reaction took place even at pH 4.0, and that hydrazine
can also take part in substitution reactions at positions occupied by carbonyl
Soil samples from each of the Ap, El, and E2 horizons were analyzed for
Ca, Mg, Na, K, Fe, and Al. While the soils unquestionably contain other
elements as well, these are the prominent ones expected in soils such as
Arredondo fine sand (Thomas et al, 1985). Results of the analysis are shown in
Table 4-3. Elemental analysis.
(mg kg-1 soil)
Ap El E2
Calcium 560.96 28.57 16.27
Magnesium 24.37 4.33 4.26
Sodium 23.94 7.48 8.36
Potassium 18.86 7.36 5.42
Iron 77.58 76.55 45.89
Aluminum 691.10 358.73 140.33
Total 1396.81 483.02 220.53
Total wt. % 0.14 0.048 0.022
Although the relative percentages of calcium and aluminum are high, the
weight percentages indicate that the total metal composition within each
horizon is low. For example, in the Ap horizon the elements analyzed account
for only 0.14 percent of the total weight, leaving the other 99.86 percent as sand
(SiO2), silt, clay, and organic matter. As previously mentioned, the organic
matter fraction in the Ap horizon accounts for 1.84 percent by weight, leaving
the remaining 98.02 percent as sand, silt, and clay.
The particle-size distribution for the Ap horizon indicated 90.1 percent
sand composition, 7.3 percent silt, and 2.6 percent clay. Thus, the elemental
analysis corroborates the particle-size analysis by identifying the relatively
small weight percentage of elemental components, and gives additional
information about the relative predominance of various elements.
The relatively high calcium concentration may be indicative of the calcitic
origin of this soil's parent material, and the relatively high percentage of
aluminum is not surprising, given the kaolin structure observed in the x-ray
diffraction analysis. Silicates weather to kaolin, which eventually breaks down
to oxide materials (especially aluminum-rich oxy-hydroxides and other oxide
materials). Aluminum bauxite ores are commonly found associated with kaolin
The three horizons of Arredondo fine sand are each acidic (Table 4-4),
thus confirming the predominance of hydrazinium as the prevalent form of
hydrazine in these soil conditions.
Table 4-4. Soil pH
Ap El E2
1 4.46 5.05 5.09
2 4.43 5.06 5.10
3a 4.49 4.98 5.13
Average 4.460.03 5.030.04 5.130.05
The buffering-capacity titration curves (Figure 4-1) are plotted as
centimoles of charge (as Ca(OH)2) per kilogram of soil versus pH. None of the
three curves show the characteristic sigmoid shape indicative of a truly buffered
plateau with less-buffered regions to either side. The titration curve of the Ap
horizon is seen to have a lower slope than that of the El horizon, indicating
greater resistance to pH change by increasing amounts of Ca(OH)2. The best-
fit lines for the titration curves, fitted by least squares, show the Ap curve to have
a slope of 2.32 (R2=0.96), the El curve to have a slope of 3.99 (R2=0.91), and
the E2 curve to have a slope of 6.96 (R2=0.95). Buffering capacities as
indicated by the slopes of the curves were also calculated for 0.1, 0.3, and 0.5
cmolc Kg-1 soil (Table 4-5). The order of buffering capacity for the three
horizons was Ap > El > E2, with none of the horizons possessing a strong
0.2 0.4 0.6 0.8
meq Ca(OH)2 per 100 g soil
Figure 4-1. Titration curves for three horizons of Arredondo fine sand.
Table 4-5. Curve slopes for various increments of Ca(OH)2 addition.
Horizon Curve Slope
(meq 100 g-1 soil)
.01 .03 .05
Ap 4.58 2.44 1.43
El 7.54 3.56 2.85
E2 11.30 5.60 2.85
Cation Exchange Capacity (CEC)
Cation exchange capacity, usually expressed in centimoles of charge per
kilogram of soil (formerly milliequivalents of charge per 100 g of soil), is a
measure of the quantity of readily exchangeable cations neutralizing negative
charge of the soil. While CEC is considered a soil property, its value also is
dependent upon the conditions under which it is measured. In these CEC
determinations, the soil material in each polysulfone tube was initially saturated
with 0.1 N CaCI2, and potassium (a monovalent cation like hydrazinium) was
used to exchange the calcium. The dilutions of 0.1 N KCI used for exchange
were adjusted to the pH of the corresponding soil.
Experimental results are plotted as potassium in the original solution
versus the difference between solution values before and after exchange
(assumed to reflect adsorption). The CEC was inferred from the plateau of this
adsorption isotherm. The procedure was performed in duplicate, with the data
displayed in Figures 4-2 and 4-3.
The colorimetric measurement of methylene blue adsorption also
produced estimations of CEC. These were similar to those obtained from the
exchange isotherm method. Data from the two approaches are shown in Table
Table 4-6. Results of the exchange isotherm and methylene blue approaches
to CEC determination.
Horizon Cation Exchange Capacity
Exchange1 Exchange2 Met Blue1 Met Blue2
Ap 8.0 8.05 8.57 8.0
E1 6.7 7.3 4.03 6.27
E2 5.2 5.2 3.57 4.35
0 20 40 60 80 100
K+ in Solution (mmol(+) L-1)
0 20 40 60 80 100
K+ in Solution (mmol(+) L-1)
Figure 4-2 (top) and 4-3 (bottom). Duplicate exchange isotherms for the Ap, El,
and E2 horizons of Arredondo fine sand.
Adsorption isotherms were plotted from the results of batch experiments
for Ca+2 saturated soils exchanged with hydrazinium (Figures 4-4 through 4-9).
Experiments were performed in an anaerobic glove box to minimize potential
oxidation, and also were performed at pH 4.8 and 8.0 to examine the effect of
pH on adsorption.
Examination of the isotherms shows that adsorption was higher at pH 8.0
than at 4.8 for all three horizons (Table 4-7), suggesting that both the neutral
hydrazine and the charged hydrazinium were adsorbed. (At pH 8,
approximately half of the hydrazine is protonated and half is in neutral form.)
The convex nature of the isotherms indicates that hydrazine is a relatively
strong competitor against calcium.
Table 4-7. Maximum hydrazinium adsorbed and percentage organic
carbon for three soil horizons.
Horizon Hydrazinium Sorbed Percentage
(plmol g-1) Organic
pH 4.8 pH 8.0
Ap 24 42 1.84
E1 14 20 0.34
E2 11 15 0.14
Note that hydrazinium adsorption appears to correlate well with the
percentage of organic carbon in each horizon A correlation analysis was
performed in which percentage organic matter (independent variable) was
linearly regressed by least squares against hydrazinium sorption (dependent
variable). The regression statistics indicated a coefficient of determination (R2)
of 0.987 for the sorption at pH 4.8 versus percentage organic carbon
regression, and 0.996 for the sorption at pH 8.0 versus percentage organic
A best-fit line determined from the regression statistics (y = 7.28x+ 10.70)
was plotted for the hydrazinium sorption versus percentage organic carbon
data at pH 4.8 (Figure 4-10). Only three soils were investigated so it is best not
to overly conclude information from the scant data. However, other researchers
have also noted the correlation between organic carbon and adsorption
capacity (Isaacson and Hayes, 1984; Brusseau et al, 1991).
S40 r ,40
30 30 -
3 20 20
10 10 -
0 30 60 90 120 0 30 60 90 120
Supernatant Hz (mmol L-1) Supernatant Hz (mmol L-1)
Figures 4-4 (left) and 4-5 (right). Adsorption isotherms for the Ap horizon at pH
4.8 and 8.0, respectively. Soil initially saturated with Ca+2.
E 12 E 12
N 8 N 8
,6 4 4
< 0' < 0
0 30 60 90 120 0 30 60 90 120
Supernatant Hz (mmol L-1) Supernatant Hz (mmol L-1)
Figures 4-6 (left) and 4-7 (right). Adsorption isotherms for the El horizon at pH
4.8 and 8.0, respectively. Soil initially saturated with Ca+2.
0 30 60 90 120
Supernatant Hz (mmol L-1)
0 30 60 90 120
Supernatant Hz (mmol L-1)
Figures 4-8 (left) and 4-9 (right). Adsorption isotherms for the E2 horizon at pH
4.8 and 8.0, respectively. Soil initially saturated with Ca+2.
0 0.2 0.4 0.6 0.8
1 1.2 1.4 1.6 1.8 2
PERCENTAGE ORGANIC MATTER
Figure 4-10. Regressed fit between sorbed hydrazinium
organic carbon. pH 4.8.
These isotherms were converted to dimensionless isotherms in order to
derive dimensionless values for the relative sorbed and solution fractions of
hydrazinium and calcium for use in predicting the influence of ion exchange on
the hydrazinium transport process. Data from dimensionless isotherms were
used to determine the selectivity coefficients and ratios of exchangeable
hydrazine to calcium at equilibrium. Dimensionless isotherms are developed
by dividing measured solution concentrations by the total normality on the
abscissa, and inferred sorbed concentrations by the maximum sorbed value on
the ordinate (Figures 4-11 through 4-13). Exchange parameters developed
from the dimensionless isotherms are listed in tabular form in Table 4-8. The
equilibrium ratio is the equivalent fraction of hydrazine and equivalent fraction
of calcium on the soil exchange sites at 0.01 N. Two-thirds of the normality is
represented by hydrazine.
Table 4-8. Ion exchange coefficients from adsorption isotherms (Hz-Ca).
Horizon K, n k Equilibrium
Ap 73.53 0.91 55.30 0.79 to 0.21
El 61.42 0.88 42.22 0.76 to 0.24
E2 35.78 0.83 20.53 0.68 to 0.32
An additional set of isotherms was performed to evaluate the efficiency of
replacement with hydrazine for a monovalent cation. Soil from the Ap and E2
horizons was saturated with Na+, and exchanged with hydrazine at pH 4.0 and
8.0 (Figures 4-14 through 4-17). The supernatant was analyzed in each case
for both Na+ and hydrazine.
Analyzing the supernatant for displaced Na+ in the second set of
isotherms showed a difference between the amount of hydrazine adsorbed and
0.20 0.40 0.60 0.80
RELATIVE CONCENTRATION (C/CT)
0.2 0.4 0.6 0.8
RELATIVE CONCENTRATION (C/CT)
Figures 4-11 (top) and 4-12 (bottom). Dimensionless adsorption isotherms for
the Ap and El horizon, respectively, at pH 4.8. Soil initially saturated with Ca+2.
I 0.6 -
0.0 I i
0.0 0.2 0.4 0.6 0.8 1.0
RELATIVE CONCENTRATION (C/CT)
Figure 4-13. Dimensionless adsorption isotherm for the E2 horizon at pH 4.8.
Soil initially saturated with Ca+2.
the amount of hydrazine retained on exchange sites. At low pH on the E2
horizon material (Figure 4-16), which had only a small percentage of clay and
organic carbon, the amount of hydrazine adsorbed was equivalent to the
amount of Na+ released at low hydrazine concentrations. This indicates that, at
pH 4.0 where the protonated form of hydrazine is dominant (99.9%), the primary
mechanism of adsorption is cation exchange.
Under alkaline conditions (pH 8) hydrazine was adsorbed even more
readily than at pH 4 (Figures 4-15 and 4-17), and there was an even greater
amount of adsorption at high concentrations. In addition, the amount of Na+
released from the soil was less than at pH 4.0, and there was little apparent
exchange between hydrazine and Na+.
2.00 4.00 6.00 8.00
SUPERNATANT HYDRAZINE (mmol L-1)
0.00 IF I I
0.00 2.00 4.00 6.00 8.0(
SUPERNATANT HYDRAZINE (mmol L-1)
Figures 4-14 (top) and 4-15 (bottom). Adsorption isotherms for the Ap horizon
at pH 4 and 8, respectively. Soil initially saturated with Na+.
| I | I
4.00 6.00 8.00
SUPERNATANT HYDRAZINE (mmol L-1)
Figures 4-16 (top) and 4-17 (bottom). Adsorption isotherms for the E2 horizon
at pH 4 and 8, respectively. Soil initially saturated with Na+.
--"- Adsorbed Hz
2.00 4.00 6.00 8.00
SUPERNATANT HYDRAZINE (mmol L-1)
" Adsorbed Hz
-- Released Na
Preliminary Column Studies
In situ soil bulk density measurements recorded by Thomas et al. (1985)
for Arredondo fine sand were in the range of 1.7 g cm-3, so the columns
prepared for these studies were packed in this range. Tables A-1 and A-2 in
Appendix A report calculations of column bulk densities as well as porosity,
volumetric water content, and percent water saturation.
Percent water saturation was calculated for the columns after a full set of
experiments had been performed at the two flow rates and the three
concentrations for each of the three horizons. Results showed that the degree
of saturation varied between 70 and 89 percent, with an average of 76 percent
(Appendix A, Table A-1). Incomplete saturation of these columns was due to air
or flushing-gas entrapment in soil pores.
New experiments designed to improve the percentage water saturation
were performed by purging the columns with carbon dioxide gas (CO2) rather
than helium, and deaerating the CaCl2 influent solutions with nitrogen gas (N2).
The solubilities of carbon dioxide, oxygen, nitrogen, and helium in water are as
follows (Table 4-9):
Using CO2 as a purge gas and N2 to deaerate the influent solution,
percent saturation increased to an average of 94 percent in the 40 additional
column experiments (Appendix A, Table A-2). This appears due to the much
greater solubility of CO2 in water than for any of the other gases used.
A series of column experiments was performed to determine the length of
time necessary to saturate the columns at each horizon and flow rate with
Table 4-9. Solubility of four gasses in water.
(cm3 gas per 100 cm3 H20
at 200C and 760 mmHg)
Source: Budavari, 1989.
CaCI2 prior to the introduction of hydrazine. This was necessary to ensure that
the column influent and effluent concentrations of CaCl2 were equal. That is,
that sorption/ion exchange processes within the columns were at steady state
with respect to Ca+2 Columns were packed and purged as previously
described before deaerated CaCI2 was introduced. Calcium was analyzed in
the effluent fractions and plotted as relative concentration (effluent
concentration (C) divided by influent concentration (Co)) versus pore volume.
Figures 4-18 through 4-22 display these results. Columns were saturated in 24-
hour increments so that experiments could begin early in the working day, and
equipment performance could be observed for an extended period of time.
Table 4-10 displays the saturation times for each horizon and flow rate.
Table 4-10. Selected saturation times prior to hydrazine addition.
Horizon Saturation Time (hrs.)
0.5 cm hr1 5.0 cm hr-1
Ap 72 48
E1 48 24
E2 48 24
0 9 18 27 36 0 50 100 150 200
TIME (hours) TIME (hours)
Figures 4-18 (left) and 4-19 (right). Calcium breakthrough curves for the Ap
horizon. High and low flow rates, respectively.
U -- U ------------
0 5 10 15 20 0 25 50 75 100
TIME (hours) TIME (hours)
Figures 4-20 (left) and 4-21 (right). Calcium breakthrough curves for the El
horizon. High and low flow rates, respectively.
0 25 50 75 100
Figure 4-22. Calcium breakthrough curve for the E2 horizon. Low flow rate.
I I I I
Data from the three breakthrough curves of each soil horizon at each flow
rate are displayed in Figures 4-23 through 4-34. From the scintillation counter,
tritium counts as counts per minute of each fraction were normalized by dividing
by the count of the input solution. These normalized counts were plotted
against their corresponding numbers of effluent pore volumes to produce a
breakthrough curve for each flow rate for each horizon.
Kirkham and Powers (1972) differentiated the solution to the convective-
dispersive transport equation for conditions of steady water flow and a step-
function input of non-reactive solute, to obtain the slope of the breakthrough
curve at the normalized concentration (C/Co) equal to 0.5 and approximately a
pore volume (p) of 1.0. The slope (Sp,)thus becomes a function of the
dispersion coefficient (D):
C = F 1-p
C= T erfc1--=p [4-1]
Co =/ 2DvL
p (C/C = 2 and [4-2]
D = [4-3]
erfc = complementary error function,
p= effluent volume, expressed as column pore volumes
D = dispersion coefficient (cm2 s-1),
v = pore water velocity, (cm s-1)
L = column length (cm), and
S = slope of the breakthrough curve at p=1 and C/Co=0.5
0 1 2 3 4
Figures 4-23 (left) and 4-24 (right).
Ap horizon at low flow rate.
0 1 2 3
Figures 4-25 (left) and 4-26 (right).
Ap horizon at high flow rate.
5 0.0 0.5 1.0 1.5 2.0 2.5
Replicate tritium breakthrough curves for the
4 0 1 2 3 4
Replicate tritium breakthrough curves for the
z \ -
U --,j --
0 2 4 6 8 10 0.0 0.5 1.0 1.5 2.0 2.5
PORE VOLUMES PORE VOLUMES
Figures 4-27 (left) and 4-28 (right). Replicate tritium breakthrough curves for the
El horizon at low flow rate.
Figures 4-29 (left) and 4-30 (right).
El horizon at high flow rate.
0 1 2 3 4
Figures 4-31 (left) and 4-32 (right).
E2 horizon at highlow flow rate.
0 1 2 3
Figures 4-31 (left) and 4-32 (right).
E2 horizon at low flow rate.
o 0.2 -
O 0.2 -
5 0 1 2 3 4 5
Replicate tritium breakthrough curves for the
4 0 1 2 3
Replicate tritium breakthrough curves for the
0 1 2 3 4 5 6 0 1 2 3 4
PORE VOLUMES PORE VOLUMES
Figures 4-33 (left) and 4-34 (right). Replicate tritium breakthrough curves for the
E2 horizon at high flow rate.
The slopes of all breakthrough curves were calculated from the data and
substituted into the analytical solution of Kirkham and Powers to determine a
dispersion coefficient for each horizon at each flow rate (Table 4-11):
Table 4-11. Experimental dispersion coefficients (cm2 h-1) for 2 water
flow velocities and 3 soil horizons.
Horizon Darcey Flow Rate
0.5 cm h-1 5.0 cm h-1
Ap 2.5x10-4 8.0x10-4
E1 3.0x10-4 4.5x10-4
E2 5.5x10-4 9.0x10-4
Hydrazine Column Studies
After saturation, hydrazine solutions were miscibly displaced for two flow
velocities through columns of soil materials from Ap, El, and E2 horizons.
Hydrazinium influent solutions were applied as finite pulses and as step
function inflows. As previously discussed, one full set of experiments was
performed with helium purging as the deaeration procedure, and another using
the carbon dioxide-nitrogen gas procedure. The carbon dioxide-nitrogen
procedure enabled column percent saturation values averaging approximately
18 percent higher.
Graphics depicting the experimental design and identifying the
breakthrough curve associated with each variable are shown in Figures 4-35
and 4-36. A number of replicate experiments were performed, and several
columns were intentionally omitted because experience at other flow rates and
for other horizons suggested that no useful additional information would be
gained from the omitted data.
Breakthrough curves were numbered 1 through 95 in consecutive order.
Early experiments to establish experimental protocol and refine analytical
technique have not been included in this discussion, and thus the first
breakthrough curve mentioned is number 17. Experimental results were
entered in a computer spreadsheet (Microsoft Excel), and are included on
diskette, rather than in hardcopy table form. The numerical data were used to
perform all analyses. The results of each column experiment are displayed in
graphical form in Appendices B through E. Two leading tables in each
appendix summarize the pertinent data used for analysis. Table 1 in each
appendix displays summary information focused on mass balance. Hydrazine
losses and rates of loss are presented for each column experiment. Table 2
summarizes information necessary to draw conclusions about the effects of ion
exchange on hydrazine loss. Having four appendices allowed the column
experiments to be grouped by method of preparation and duration:
Appendix B. Helium preparation, pulse duration (Tables B-1 and B-2,
Figures B-1 through B-16).
* Appendix C. Helium preparation, continuous duration Tables C-1 and C-2,
Figures C-1 through C-16).
* Appendix D. Carbon dioxide-Nitrogen preparation, pulse duration. (Tables
D-1 and D-2, Figures D-1 through D-32).
* Appendix E. Carbon dioxide-Nitrogen preparation, continuous duration
(Tables E-1 and E-2, Figures E-1 through E-11).
0.5 cm hr-1
5.0 cm hr-1
0.5 cm hr"1
5.0 cm hr-1
0.5 cm hr-1 5.0 cm hr-1
0.5 cm hr-1 5.0 cm hr-1
0.5 cm hr-1 5.0 cm hr1
0.5 cm hr-1 5.0 cm hr-1
Figure 4-35. Breakthrough curves associated with the variables of duration of
hydrazinium input, flow, concentration, and horizon. Helium preparation.
Numbers shown in the blocks are designations for specific column experiments.
0.5 cm hr-1
5.0 cm hr"1
7 Ap 70 56
E El 69 55
E2 54 66,60
0.5 cm hr-1 5.0 cm hr"1
SAp 74,76 57
E El 72,81 58,67
SE2 73,80 59
0.5 cm hr-1 5.0 cm hr-1
- Ap 75,79 61,63
c4 El 77 62,65
E2 78 64,68
0.5 cm hr"1 5.0 cm hr-1
0.5 cm hr1 5.0 cm hr-1
0.5 cm hr'1 5.0 cm hr-1
Figure 4-36. Breakthrough curves associated with the variables of duration of
hydrazinium input, flow, concentration, and horizon. Carbon dioxide and
nitrogen preparation. Numbers shown in the blocks are designations for
specific column experiments.
Microbial Activity in Soil Columns
The presence of microbial populations in the soil of the column
experiments was examined by performing acridine orange (A-O) direct counts
on soil from a number of the completed soil column experiments. A
considerable microbial population was found in each soil examined (Table 4-
14). Since the acridine orange dye stains both active and inactive cells, a
number of plate counts for active microbial populations were also performed.
An average of the A-O counts by horizon from completed column
experiments showed soil from the Ap horizon to contain 7.6x108 organisms per
gram, with the El horizon containing 1.8x108, and the E2 horizon containing
1.2x108 organisms per gram of wet column soil. Plate counts averaged 1.9x107
organisms per gram, or approximately an order of magnitude less than the A-O
count performed on soil from the same column. Background counts of column
soil from experiments not containing hydrazine contained the following counts
per gram: Ap, 8.0x108; El, 6.1x106; and E2, 7.3x106.
One purpose of performing continuous duration column experiments was
to observe the plateau of the hydrazine output curve. It was anticipated that the
plateau observed in the effluent measurements of long-duration pulses might
be lowered by a process such as microbial degradation once the sorptive
demand of the soils was met, and that this effect might be observed by
normalizing the effluent hydrazine concentrations by dividing by the influent
concentration. Thus, a relative effluent concentration less than the value of 1
would be observed. However, this lowering of the effluent plateau was not
observed, even though a comparison of influent and effluent masses indicated
the loss of hydrazine. Apparently, the rate of hydrazine loss was sufficiently
small (subsequently measured as 0.05 to 0.1 mmol Hz kg-1 soil hr-), so as to be
masked by experimental scatter or diluted by the normalization process.
A review of published literature investigating the environmental fate of
hydrazine in soil and water suggests that the processes of greatest effect are
oxidation, autoxidation, microbial and chemical degradation, and sorption (both
reversible and irreversible, and ion exchange). The potential degradation
pathways of oxidation and autoxidation are minimized in this work by
performing experiments in the pH range in which hydrazine is stable. The
hydrazine influent for soil-column experiments was adjusted to the pH of the
soils, the acidic pH range of 4.46 to 5.13, in which hydrazine (N2H4), with a pKa
of 7.96, occurs as 99.99% hydrazinium (N2H5+). Hydrazinium in water has
been shown to be stable with respect to oxygen at acidic pH. Additionally, this
protonated condition suggests that ion exchange should be investigated as a
potentially significant fate and transport process.
Several environmental variables were studied in conjunction with this
examination, including variable hydrazine solution concentration, water
velocity, percentage of soil organic matter, and soil pH. Three successive
horizons of coarse-textured sand were used to investigate the effects of organic
matter without the necessity of subjecting a single soil to the oxidative treatment
necessary to remove humic material (Wolf et al., 1989).
Some general but significant qualitative conclusions can be made about
the effect of such environmental variables as percentages of organic matter and
clay, solution concentration, and pore water velocity by examining the column
mass input-output calculations shown in Tables B-1, C-1, D-1, and E-1 of their
Percentage of Organic Matter and Clay
The influence of organic matter and clay within a soil horizon on the
disappearance, or loss, of solute from column effluent has been noted in the
literature review. "Hydrazine loss" is defined here as that portion of the influent
hydrazine which did not appear in the effluent. Since the soil columns were not
analyzed for residual hydrazine at the termination of miscible displacement,
hydrazine could have been "lost" by irreversible sorption, slow desorption by
soil solids, or by microbial degradation.
Hydrazine losses from the 27 column pulse experiments described in
Appendix D, Table 1 were determined by subtracting the mass of hydrazine
recovered in the column effluent from that introduced into the columns, and then
normalizing by the mass of soil in each column. Duplicate soil-column
experiments within the data-set were averaged to give one value of hydrazine
loss for each experimental configuration of flow rate and concentration. The
losses are summarized by horizon in Table 5-1, together with the percentage of
organic matter and clay determined from batch experiments (reported earlier in
Table 4-1 and 4-2). Percentage hydrazine loss was determined by summing
total hydrazine mass loss for the three experimental trials within each horizon,
and dividing by the total hydrazine mass input for those trials, times 100.
Twenty-seven column experiments are represented in the table, nine for each
horizon. Influent mass loadings for all columns within each horizon were within
20 percent of one another.
Table 5-1. Hydrazine losses, percentage organic matter, and percentage
clay by horizon for a total of 27 column experiments.
Horizon Hz Loss % Hz %OM % Clay
(mmol Hz Loss
Ap 12.76 37.5 1.84 2.6
E1 4.12 14.4 0.34 1.7
E2 2.77 10.4 0.14 1.8
The percentage hydrazine losses for each horizon were regressed
against the percentage organic matter and percentage clay content of each
horizon in an attempt to establish a correlation, which is depicted graphically in
Figure 5-1. The best-fit line for the hydrazine loss versus percentage organic
matter has a correlation coefficient (R2) of 0.999, while the correlation of the
hydrazine loss versus percentage clay content has an R2 of 0.944.
The fit of the predicted hydrazine loss due to the percentage of organic
matter in each horizon [%Hz Loss = 15.74*(%OM) + 8.59] intersects the Y-axis
at a positive value, suggesting that there is hydrazine loss when there is no
organic matter present, or that the presence of organic matter does not account
for all hydrazine losses (a reasonable conclusion). While the fitted line is only
regressed through three points, the points each represent averages of many
values, lending confidence to the regression and to the conclusion that there is
a correlation between percentage organic matter and percentage hydrazine
4 0 -Pred.Loss (OM)
------Pred. Loss (Clay) I
) 30 -
N- % Organic Matter
W 15 '
O El ,El
1 E2 E2
0 0.5 1 1.5 2 2.5 3
PERCENTAGE ORGANIC MATTER AND CLAY
Figure 5-1. Linear regression of both percentage organic matter and
percentage clay against percentage hydrazine loss.
The best-fit line through the clay data crosses the Y-axis (representing
percentage hydrazine loss) at minus 37.82 percent [%Hz Loss = 28.81*(%Clay)
- 37.82]. Care should be taken here in assuming that a good correlation
coefficient (R2=0.944) implies a truthful relationship. The best-fit line implies the
negative loss, or unrealistic manufacture, of hydrazine when passing through a
soil column. Taken by itself, the line also indicates that the E2 horizon should
be more sorptive that the El horizon, a conclusion not supported by the data.
A multiple regression including both percentage organic matter and
percentage clay versus percentage hydrazine loss was observed to fit the data
with a correlation coefficient of 1.0, though the fitted line passed through none of
the data points. The equation [Hz loss = 17.91*(%OM) 4.18*(%Clay) + 15.42]
shows a negative correlation with percentage clay, as well as unreasonable
coefficients. It is simply a mathematical fit through too few data points to allow
Thus, there appears to be a good correlation between the percentage of
organic matter and the percentage loss of hydrazine within soil columns. The
equation of the best-fit line also appears reasonable when interpreted with the
data. The correlation of percentage clay and percentage hydrazine loss, while
having a high correlation coefficient, does not appear reasonable.
An examination of the effect of variable hydrazine concentration on
hydrazine transport was included in the initial experimental protocol. Through
the course of the experimental studies, replicate experiments were performed at
each of the three targeted concentration ranges (0.20 mmol L-1, 5.0 mmol L-1,
and 20 mmol L-1). In the analysis of the data, an attempt was made to correlate
hydrazine solution concentration with percentage hydrazine loss from columns
receiving a pulse of hydrazine influent. To minimize the number of variables
which might affect the correlation, a separate linear regression was performed
on data from each horizon at each velocity, for a total of six regressions (three
horizons, two velocities in each). The data, equations of predicted linear
hydrazine loss best-fit lines, and correlation coefficients are shown on Table 5-
2. Correlation coefficients are observed to range from 0.63 to 0.99. With the
exception of the slow velocity data (R2=0.99) of the Ap horizon, the data appear
only sufficiently correlated to define a general trend. The single high correlation
coefficient appears fortuitous, given the relatively low correlation of the other
The data and predicted best-fit line from the slow velocity E2 horizon is
felt to be typical of the data-set, and is shown graphically in Figure 5-2. Note the
negative slope of the best-fit line, indicating that greater percentage hydrazine
losses are associated with lower concentrations. This trend would be expected
if there existed a sink, or loss mechanism, of finite extent which exerted its
demand relatively early in the transport process. A low concentration (and thus
Table 5-2. Hydrazine loss as a function of soil-column influent concentration.
Horizon Flow Conc. % Hz Hz Predicted Loss R2
Rate (mmol Loss equation
Ap Fast 22.15 21.4
0.18 100 -3.10*(Conc.)+84.9 0.78
Slow 16.38 41.0
0.20 100 -3.57*(Conc.)+98.5 0.99
El Fast 15.00 12.6
0.19 100 -5.24*(Conc.)+79.1 0.63
Slow 18.14 9.0
0.23 100 -4.17*(Conc.)+77.4 0.66
E2 Fast 18.39 1.5
0.21 36.7 -1.63*(Conc.)+27.8 0.64
Slow 12.28 14.9
0.19 56.3 -3.28*(Conc.)+50.4 0.80
low total mass) input would be lost in the sink, whereas a higher concentration
input would fill the demand of a finite sink term, leaving excess solute to be
affected by other more slowly acting, perhaps rate-limited, loss/degradation
The concept of a finite-extent sink is well-known, and frequently modeled
as a sorptive loss. The data also support this concept. Note that, in slow-
velocity experiments, the solute input to the Ap and El horizons does not
emerge in the effluent, whereas hydrazine solute is observed in the effluent
fractions from the E2 horizon, implying that the sink term for the E2 horizon may
be smaller than for the other horizons. The same trend is also noted at similar
velocities in each horizon .
0 I I I I
0 5 10 15 20
CONCENTRATION (mmol Hz L-1)
Figure 5-2. Hydrazine concentration versus percentage hydrazine lost in the El
horizon at slow flow-rate.
The relationship between the rate of advective movement of a solute and
the rate of its interaction with the soil environment is a critical factor in assessing
its environmental fate and transport. The assumption is often made that
reactions occur instantaneously, or at least quickly with respect to solute
transport, and that process equilibrium has been established. Although this
assumption simplifies the conceptual understanding of the processes involved
as well as the mathematics utilized in fate and transport predictions, kinetic
processes commonly are associated with the fate and transport of chemicals in
Two factors were incorporated into the design of the hydrazine
experiments to evaluate transport equilibrium. The Darcy water velocity was
varied an order of magnitude as either 5.0 cm hr1 or 0.5 cm hr1, and the
hydrazine exposure time in the columns was varied using either pulse or
continuous solute input. The assumption was made that even slowly reacting
processes would eventually come to equilibrium over a long exposure.
Column experiments for each soil horizon were duplicated for each
velocity. Summary results are shown in Table 1 in appendices B and D, and
are condensed here in Table 5-3:
Table 5-3. Percentage hydrazine losses by flow rate for otherwise
replicate experiments (from Table 1 in Appendices B and D).
Horizon Percentage Hydrazine
5.0 cm hr1 0.5 cm hr1
Ap 36 63
E1 17 20
E2 5 17
Total 20 33
Percentage hydrazine losses shown in this table were determined by
taking the difference between input and output hydrazine masses from each
column trial, dividing them into groups by velocity, and then summing each
group and determining a percentage loss for each horizon. Thirty-six column
experiments are represented in the table, 18 for the nitrogen-CO2 prepared