Phelan, J.M. and S.W. Webb, 1998a. Chemical Detection of Buried Landmines.

Proceeding of the 3rd International Symposium on Technology and the Mine

Problem, April 6-9, 1998. Mine Warfare Association.

Chemical Detection of Buried Landmines

James M. Phelan1 and Stephen W. Webb2

Of all the buried landmine identification technologies currently II. CHEMICALS IN THE SOIL ENVIRONMENT
available, sensing the chemical signature from the explosive
components found in landmines is the only technique that can
Soils are porous media with a number of physico-chemical
classify non-explosive objects from the real threat. In the last two properties that affect the transport of explosive chemicals.
decades, advances in chemical detection methods have brought Soil bulk density is a measure of the compaction of the soil
chemical sensing technology to the foreground as an emerging and is defined as
technological solution. In addition, advances have been made in M
the understanding of the fundamental transport processes that ρb = s [1]
Vs
allow the chemical signature to migrate from the buried source to
the ground surface. A systematic evaluation of the transport of the where ∆b is the soil bulk density (g/cm3), Ms is the mass of
chemical signature from inside the mine into the soil environment, soil particles (g), and Vs is the volume of soil (cm3). Soils
and through the soil to the ground surface is being explored to under natural conditions have bulk densities ranging from
determine the constraints on the use of chemical sensing 1.0 to 1.8 g/cm3. However, soils that have been excavated
technology. This effort reports on the results of simulation and replaced, such as during the emplacement of a
modeling using a one-dimensional screening model to evaluate the landmine, may have bulk densities much less than 1. The
impacts on the transport of the chemical signature by variation of soil bulk density is inversely proportional to the soil
some of the principal soil transport parameters.
porosity as follows
landmines, chemical sensors, soil transport φ = 1 − ρb / ρs [2]
where ∆s is the soil particle density (ranges from 2.6 to 2.8
I. INTRODUCTION g/cm3 for most soils). The soil porosity, or void volume, is
defined as
The organic chemicals of the explosives in the buried Vw + Va
landmine environment can exist in or on four phases: solid φ= [3]
Vs
phase of the neat explosive material, vapor phase in the soil
air, aqueous phase in the soil water solution, and sorbed where Ν is the soil porosity (cm3/cm3), Vw is the volume of
onto soil solid phases. The chemical signature begins as a soil water (cm3) and Va is the volume of soil air (cm3). Soil
surface coating from production or depot storage and porosity values range from 0.3 for sands to 0.6 for clay rich
through continuous emission by permeation through the soils. The volumetric moisture content describes how much
mine case or through leaks in seals and seams. Once the water is present in the soil and changes greatly during
chemicals enter the soil environment, they experience phase precipitation/drainage events and evaporation conditions.
transitions, partitioning into the soil air, soil water and Volumetric water content is defined as
sorbing onto soil particles. The impact of temperature and Vw
chemical gradients, and precipitation/evaporation will cause
θ= [4]
Vs
movement of the chemical signature. Part of this transport
where 2 is the volumetric water content (cm3/cm3). Soil
is upward to the soil surface where chemical detection
moisture contents have values from near zero up to the soil
technology is envisioned to be used. Simulation modeling
porosity value. When the soils are not fully saturated, the
is a technique that can evaluate the impacts of many of the
balance of the soil pore space not filled with water is termed
environmental variables that can dampen or accentuate the
the air filled porosity, and is defined as
surface expression of the chemical signature. Model results
will be shown that describe the magnitude of the changes Va
a= [5]
that accompany variations due to chemical properties of the Vs
explosive and properties of the soil environment.
where Va is the volumetric air content (cm3/cm3).

It is often more convenient to use soil saturation (Sl)

because it is a measure of the relative saturation of a
1
Environmental Restoration Technologies Department, particular soil pore space with water.
2
Mission Analysis and Simulation Department,
at Sandia National Laboratories, Albuquerque, NM.
 θ
Sl = [6] Table 1. Vapor Density, Water Solubility and Henry’s Law
φ Constant of Explosive Compounds at 20°C
Since the explosive chemicals can exist as solutes in the soil TNT DNT RDX
water and the movement of soil water can be a significant 9DSRU'HQVLW\ JP3) 43.5 122 0.024
transport mechanism, water solubility is an important Water Solubility (mg/l) 130 270 50
parameter. Water solubility is defined as Henry’s Law Constant 3.35E-7 4.51E-7 4.73E-10
M chem
CL = [7]
Vw Pennington and Patrick (1990) measured the soil water
where CL is the concentration in aqueous phase partition coefficient (Kd) of TNT in fourteen soils from
(g/cm3 soil water) and Mchem is the mass of chemical (e.g. locations across North America. The mean value was 3.8
TNT) (g). Water solubility, however, is not constant and is cm3/g with a standard deviation of 1.34. The highest value
typically an increasing function with temperature. was 6.8 cm3/g and the lowest value was 2.3 cm3/g. Xue et
al. (1995) evaluated two soils and showed mean values for
Henry’s Law constant is a relative measure of the amount of TNT of 2.66 cm3/g and 3.64 cm3/g. DNT and RDX have
the chemical that exists in the gas phase to that in the very little data. Xue et al. (1995) showed values for RDX
aqueous phase at equilibrium, and is defined as of 1.59 cm3/g and 1.57 cm3/g. McGrath (1995) showed a
Koc value of 251 for DNT. For the fourteen soils evaluated
CG
KH = [8] by Pennington and Patrick (1990), the mean value for the
CL fraction of organic carbon was 0.0173 with a standard
where KH is the Henry’s Law constant (unitless) and CG is deviation of 0.011. Using these values, the Kd for DNT has
the concentration in gas phase (g/cm3 soil gas). Henry’s a mean value of 4.4∀2.7 cm3/g (one std. dev.). In summary,
Law constant is also a function of temperature because both the soil water partitioning coefficients for TNT, DNT and
CG and CL are functions of temperature. RDX all fall into an approximate range between 1.5 and 7.0
cm3/g. This is a rather narrow range as common chemicals
The soil partition coefficient is a relative measure of how can have values one to two orders of magnitude lesser and
much of the chemical is temporarily bound to the soil to greater than these.
that in the soil aqueous phase
The biochemical half-life of explosives in near surface soils
C
Kd = S [9] has not been studied well outside of the biotreatment
CL technology area for contaminated soils. However, long-
where Kd is the linear soil-water partition coefficient term surface soil degradation tests at Los Alamos National
(cm3/g) for water saturated soils and CS is the concentration Laboratory (LANL) followed the degradation of soils doped
sorbed on the soil solid phase (g/g of soil). with 1000 mg/kg of various explosives over 20 years
(Dubois and Baytos, 1991). Table 2 shows the half-lives
The soil water partition coefficient is often correlated with estimated from these long-term experiments.
the fraction of organic carbon found in the soils. In this
way, the variability between soils can be reduced. The Table 2. Estimates of Half-Lives of Explosives from LANL
organic carbon distribution coefficient is defined as Long-Term Surface Soil Tests
Kd Explosive Half-Life (years)
Koc = [10] TNT 1
f oc RDX 36
where Koc is the organic carbon distribution coefficient and HMX 39
foc is the fraction of organic carbon. PETN 92

IV. SCREENING MODEL

III. PHYSICAL CHEMICAL PROPERTIES OF EXPLOSIVES
The environmental fate and transport of organic chemicals
The principal explosive chemicals found in landmines are including volatilization and leaching losses has been used to
TNT and RDX (NGIC, 1995). DNT, as a production by- explore the distribution of agricultural pesticides in soils
product of TNT, is also considered to be a significant (Mayer et al. 1974, Farmer et al. 1980, and Jury et al.
signature chemical for buried landmines. As a group, these 1980). These models were primarily intended to simulate
chemicals have very low vapor densities and moderately specific circumstances. However, Jury et al. (1983, 1984a,
low water solubilities. Table 1 shows these properties and 1984b, 1984c) developed and validated a general screening
the Henry’s Law constant at 20°C (Phelan and Webb, model (Behavior Assessment Model, BAM) that included
1997). volatilization, leaching, and degradation to explore the
major loss pathways of agricultural pesticides as a function
of specific environmental conditions. The model
simulations can be used to assess the behavior of different 6. The adsorbed and dissolved phases undergo
chemicals under particular environmental conditions, but is reversible, linear adsorption.
not intended to predict a definitive concentration 7. The dissolved and gaseous phase concentrations are
distribution in the field. As such, the predictions from the related through Henry's law.
screening model are only an indication of expected 8. The soil properties are constant in space and time.
conditions. 9. Water flux is constant in space and time (relaxed in
the present application).
This model is valuable in that it can express the total 10. Volatilization of the chemical to the atmosphere is by
concentration of a chemical in the gas, aqueous and sorbed vapor diffusion through an air boundary layer of
phases. The total concentration is expressed as constant thickness.

CT = ρb CS + θC L + aCG [11] In the present implementation of Jury’s model, a constant

source term has been added to reflect the chemical source
where CS is the concentration sorbed to the soil, CL is the from the landmine at a specific location.
solute concentration in the aqueous phase, and CG is the gas
phase concentration. In addition, Jury (1983) shows how Under these assumptions (including the source term) the
equation [11] can be rewritten in terms of one of the model formulation becomes
variables alone ∂C T ∂ 2 CT ∂CT
+ µC T = D E − VE +σ [16]
∂t ∂z 2
∂z
CT = RS CS = R L C L = RG CG [12] where CT is the total chemical concentration, : is the bio-
chemical decay constant, and Φ is the source term. The
where effective velocity (VE) is defined as
θ KH Jw
RS = ρb + +a [13] VE = [17]
Kd Kd ρb K d + θ + aK H
where Jw is the precipitation/evaporation flux. The effective
R L = ρb K d + θ + aK H , and [14] diffusion coefficient (DE) of the chemical is defined as
a 10 / 3 K H Dga + θ 10 / 3 Dlw
θ DE =
φ 2 ( ρb Kd + θ + aK H )
Kd [18]
RG = ρb + +a [15]
KH KH
where Dga is the diffusivity of the gas phase of the chemical
are the solid, liquid and gas phase partition coefficients, w
in air and Dl is the diffusivity of the chemical in aqueous
respectively.
phase . The boundary conditions for the problem are
diffusion through a boundary layer at the upper surface, and
An adaptation of the BAM was developed to be applicable
a zero chemical concentration at infinity at the lower
to the conditions of contaminated soil buried under a known
boundary. These boundary conditions can be expressed as
depth of clean soil - Buried Chemical Model, BCM (Jury et
∂CT
al., 1990). Simulations based on a modification of Jury’s − DE + VE CT = − HE CT [19]
BCM are used in this report to simulate the behavior of the ∂z
chemical signature from buried landmines. The Buried where
Chemical Model of Jury et al. (1990) is based on the hK H
following assumptions. A detailed discussion of these HE = [20]
assumptions is given in Jury et al. (1990). ρb K d + θ + aK H
and
1. The chemical may adsorb on the solid phase, be Dga
dissolved in the aqueous phase, or exist in the vapor h= [21]
phase. d
2. The chemical flux is the sum of the vapor flux and the and
dissolved solute flux using Fick's law.
3. The porous medium factors for gas and liquid phase CT ( ∞, t ) = 0.
diffusion are given by the Millington and Quirk
(1961) model as extended for liquid diffusion by Jury
The initial conditions are an initial concentration, C0, over
et al. (1983).
an interval from L to W, or
4. The chemical will undergo first-order degradation due
to biological and chemical effects.
5. Chemical movement is one dimensional. CT ( z ,0) = 0 L 〉 z 〉W
 CT ( z ,0) = C0 L ≤ z ≤W Table 2 shows the input parameters used in the simulations.

The above model results in a closed form solution as a Table 2. Simulation Parameters
function of space and time; the results are rather lengthy and parameter units base case variant
will not be presented here but are given by Jury et al. (1983, cases
1990). In the present simulation, the assumption of constant θ cm3/ cm3 0.25 0.375
water flux in time will be relaxed. Therefore, sequences of
water fluxes representing desired conditions (rainfall φ cm3/ cm3 0.5 *
followed by evaporation) can be simulated to determine the ρb g/cm3 1.5 *
effect of water flux variations on the location of TNT in the Kd cm3/g 1.6 3.8
soil and the surface TNT vapor flux. A numerical solution 6.0
was developed and verified by comparison to the results KH -- 5.9E-7 4.73E-10
given by Spencer et al. (1988) and Jury et al. (1990) air boundary cm 0.5 *
(Phelan and Webb, 1997). layer
t1/2 days 365 180
Using this solution, simulations were performed using a 60
landmine that has contributed an initial soil concentration Co µg/cm3 4.6E-3 0
(Co) based on the surface contamination of the landmine. It Jc µg/cm2-day 8.6E-6 0
has been assumed that the entire surface contamination was Dlw cm2/day 0.432 *
completely and uniformly transferred to the soil just prior to
Dga cm2/day 4320 *
the beginning of the simulation runs. Surface contamination
data (Hogan et al., 1992) showed a median surface burial depth, cm 10 *
contamination of 15 ng/cm2 from 42 domestic and foreign top
landmines. Using the dimensions of an anti-tank (AT) mine burial depth, cm 20 *
of 30 cm diameter by 10 cm high, the surface contamination bottom
would provide 3.5x10-5 g of TNT for initial distribution in cycles/yr -- 45 *
the soil. Using the volume of the AT mine that this mass of precipitation days 1 *
TNT is distributed into, the initial concentration (Co) would evaporation days 7 *
be ~5x10-3 µg/cm3. precipitation cm/day 0.44 *
rate
The constant source term emanation rate was derived from evaporation cm/day - 0.063 *
vapor collection chamber experiments on two mines rate
(Spangler, 1975). Values ranged from 10-16 to 10-18 g/cm2- total cm/year 20 *
s. The higher rate of 10-16 g/cm2-s (8.6x10-6 µg/cm2-day) precipitation/
was used in these simulations. If the top of the AT mine evaporation
was buried at a depth of 10 cm, the burial zone of the initial * - same as the base case
contamination is from 10 to 20 cm, and the constant source
term is placed at a depth of 15 cm. V. DISCUSSION
a For each of the figures shown, there is a distinct oscillation
The diffusivity of gas in air ( Dg ) and diffusivity of liquid of the surface vapor flux. This feature is a result of the
w
in water ( Dl ) were selected from Jury et al. (1983). The cycling of precipitation and evaporation. To evaluate the
effect of the Henry’s Law constant, two simulations were
biochemical half-life value of 365 days was selected from a performed where all parameters were kept constant with one
long term field experiment (Dubois and Bayton, 1991). case using a KH equal to that of TNT and one for RDX
(both at 20°C). Figure 1 shows the results and indicate that
The precipitation/evaporation rates and periods followed in
the TNT surface vapor flux would be expected to reach a
all the simulations here were the low desert scenario from
greater steady state value than RDX, approximately
Phelan and Webb (1997). This scenario was derived from
proportional to the ratio of Henry’s Law constants. For
data found in HELP (Hydrological Evaluation of Landfill
TNT, a temperature increase from 0°C to 40°C will increase
Performance) model (Schroeder et al., 1994a and 1994b). .
the KH value by a factor of about 100. It appears that
The HELP model showed that the low desert had 1 day of seasonal and diurnal soil temperature changes could make a
precipitation followed by 7 days of evaporation. For significant effect on the subsurface transport and surface
simplicity, total precipitation and total evaporation for each flux of explosive signatures.
cycle are assumed to be equal and for these simulations the
cycles were continued for approximately four to ten years.
 1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
1.00E-05
1.00E-06
1.00E-07

Surface Flux (ug/cm2-day)

1.00E-08
Kh
1.00E-09
1.00E-10 TNT
1.00E-11
1.00E-12
RDX
1.00E-13
1.00E-14
1.00E-15
1.00E-16
1.00E-17
1.00E-18
1.00E-19
1.00E-20
1.00E-21
1 2
1.00E-22
1.00E-23
0 200 400 600 800 1000 1200 1400 1600
Time (days)

Figure 1. Effect of Henry’s Law Constant on Surface Vapor Flux

Depth concentration profiles at the end of the simulation Next, simulations were performed to evaluate small changes
period show that for both cases the concentration of TNT in the soil water partitioning coefficient. The Kd values for
and RDX are essentially equal. This implies that most of TNT, DNT and RDX have values from about 1.5 to 7.
the transport upward to the ground surface is within the Figure 2 shows the surface flux over time for Kd values of
aqueous phase and that the release of the chemical into the 1.6, 3.8 and 6.0 cm3/g.
vapor phase above the ground surface is directly
proportional to the Henry’s Law constant.
1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
1.00E-05
1.00E-06
1.00E-07 Kd
Surface Vapor Flux (ug/cm2-day)

1.00E-08
1.6
1.00E-09
1.00E-10
1.00E-11 3.8

1.00E-12
6.0
1.00E-13
1.00E-14
1.00E-15
1.00E-16
1.00E-17
1.00E-18
1.00E-19
1.00E-20
1 2 3
1.00E-21
1.00E-22
1.00E-23
0 200 400 600 800 1000 1200 1400 1600
Time (days)

Figure 2. Effect of Soil Water Partitioning Coefficient on Surface Vapor Flux

These simulations show that even though the soil water evaporation cycles becomes more pronounced. This is
partitioning coefficients appear to vary only slightly among consistent with the lower Kd value, since more of the mass
many soils, there is a significant impact to the transport of of the chemical is found in the aqueous phase and is
the chemical to the ground surface. As the Kd value affected by the upward and downward flux of water. Figure
increases, the lag period becomes much longer and the 3 shows the depth profiles from the simulations varying the
steady state concentrations stabilize at much lower levels. soil water partitioning coefficients. These curves show that
Also with the lower Kd values, the effect of precipitation/ the simulations with lower Kd values have more significant
transport of the chemicals to soils both above and below the source zone than the higher Kd values.
2.50E-03

Kd = 6

2.00E-03

Kd = 3.8

Total Concentration (ug/cm3) 1.50E-03

Kd = 1.6

1.00E-03

5.00E-04

0.00E+00
0 5 10 15 20 25 30
Depth (cm)

Figure 3. Depth Profile After 4 Years With Variant Kd Values

The next simulations were an evaluation of the source term and the amount from the surface emission flux (0.02
parameters. Very little data is available on the initial ug/day), it would take about 4.8 years for the total mass of
concentration of explosives coating the outer surface of a explosive from the surface emission flux to equal the
mine and even less so on the surface emission flux. Figure amount from the initial surface contamination. Curve two
4, curve one shows the effect of reducing the surface in Figure 4 shows the effect where the initial concentration
emission flux to essentially zero (a value of 1E-20 ug/cm2- on the mine is essentially zero (a value of 1E-20 ug/cm3
day was used). This curve is essentially the same as curve was used). This shows a significantly longer lag time and
one in Figure 1 where the surface emission flux (Jc) was about four orders of magnitude lower steady state surface
equal to 8.6E-6 ug/cm2-day. This implies that the surface vapor flux at the end of the simulation period. Figure 4
emission flux makes very little contribution to the overall indicates that the initial concentration is a much more
mass transport in the soil. If one considers the total mass of important parameter than the surface flux for the mass
explosive contributed by the initial concentration (35 ug) transport of chemicals to the ground surface.

1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
1.00E-05
1.00E-06
1.00E-07
Surface Vapor Flux (ug/cm2-day)

Co = 4.6E-3 ug/cm3
1.00E-08
Jc = 1E-20 ug/cm2-day (approximates zero)
1.00E-09
1.00E-10
1
1.00E-11
1.00E-12
1.00E-13
1.00E-14
1.00E-15
2
1.00E-16
1.00E-17
1.00E-18
Co = 1E-20 ug/cm3 (approximates zero)
1.00E-19
Jc = 8.6E-6 ug/cm2-day
1.00E-20
1.00E-21
1.00E-22
1.00E-23
0 200 400 600 800 1000 1200 1400 1600
Time (days)

Figure 4. Effect of Source Term Variation on Surface Vapor Flux

Figure 5 shows the cases where the initial concentration is (curve 1) over the base case (curve 3). These simulations
increased by a factor of 10 (curve 2) and a factor of 10E+5 show that the lag period remains about the same; however,
the increase in steady state flux is proportional to the increase in initial concentration (C0).

1.00E+00

1.00E-01

1.00E-02

1.00E-03

1.00E-04
Co = 4.6E+2 ug/cm3
1.00E-05
1
Surface Vapor Flux (ug/cm2-day)
1.00E-06

1.00E-07

1.00E-08
Co = 4.6E-2 ug/cm3
1.00E-09
2
1.00E-10
3
1.00E-11

1.00E-12
Co = 4.6E-3 ug/cm3
1.00E-13

1.00E-14

1.00E-15

1.00E-16

1.00E-17

1.00E-18

1.00E-19
0 200 400 600 800 1000 1200 1400 1600
Time (days)

Figure 5. Effect of Source Term Variation on Surface Vapor Flux

Another parameter where there is very little data is the (curve 2). Figure 6 shows that over the long-term, the
biochemical half life (t1/2). There are many influences on shorter half-life will significantly decrease the steady state
the magnitude of this parameter and the variability is surface flux. Biochemical decay constants that are very
expected to be large. Simulations over ten (10) years were large (e.g. RDX values over 30 years) appears to have
completed to assess the impact of decreases in the minimal impact to the short term soil transport phenomena
biochemical half-life from 365 days (curve 1) to 180 days evaluated in these simulations.

1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
1.00E-05
1.00E-06
1.00E-07
Surface Vapor Flux (ug/cm2-day)

T1/2 = 365 days

1.00E-08
1.00E-09
1.00E-10
1
1.00E-11
1.00E-12
1.00E-13
T1/2 = 180 days 2
1.00E-14
1.00E-15
1.00E-16
1.00E-17
1.00E-18 T1/2 = 60 days

1.00E-19
3
1.00E-20
1.00E-21
1.00E-22
1.00E-23
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Time (days)

Figure 6. Effect of Half-Life on Surface Flux

Initial simulations using this approach showed that for a greater steady state concentrations. Figure 7 shows how
particular scenario, the constant moisture content value had increasing the soil saturation from 0.5 to 0.75, decreases the
a significant impact on the lag period (when the vapor flux lag period substantially and increases the steady state
reached the ground surface) and the steady state surface flux by two orders of magnitude. The constant
concentrations (Phelan and Webb, 1997). Higher moisture moisture content assumption in this screening model allows
contents showed significantly shorter lag periods and for a simplified mathematical solution to complex transport
phenomena. One must recognize this assumption and not transport phenomena of chemical signatures from buried
over generalize the information gleaned from these landmines in more detail and with fewer simplifying
simulations. Future efforts will include the development of assumptions.
a numerical simulation capability that can explore the

1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
1.00E-05
1.00E-06
1.00E-07 Sl=0.75
1
Surface Flux (ug/cm2-day)

1.00E-08
1.00E-09
1.00E-10
2
1.00E-11
1.00E-12
Sl=0.5
1.00E-13
1.00E-14
1.00E-15
1.00E-16
1.00E-17
1.00E-18
1.00E-19
1.00E-20
1.00E-21
1.00E-22
1.00E-23
0 200 400 600 800 1000 1200 1400 1600
Time (days)

Figure 7. Effect of Soil Moisture Content on Surface Flux

In some parts of the world, minefields are located in areas followed by -0.5 cm/day of evaporation for 60 days. This
that experience very distinct wet and dry climatic periods. shows the immediate drop in the steady-state surface flux
Figure 8 shows the result of a simulation to assess the effect after precipitation begins. Once the evaporation period
of a short-term continuos precipitation period followed by a begins, there is a short lag period where the surface vapor
short-term evaporation period. The baseline simulation flux stays nearly constant before rising to just above the flux
(Figure 1, curve 1) was run for 1440 days followed by 1 before the precipitation began.
cm/day of precipitation for 30 days which was then

1.00E+00

1.00E-02

1.00E-04

Precipitation begins
1.00E-06

1.00E-08
Surface Flux (ug/cm2-day)

1.00E-10

1.00E-12

1.00E-14

1.00E-16

1.00E-18

1.00E-20

Evaporation begins
1.00E-22

1.00E-24
1200 1250 1300 1350 1400 1450 1500 1550
Time (days)

Figure 8. Effect of Continuous Precipitation Followed by Continuous Evaporation

 [1] DuBois, F.W. and J. F. Baytos. 1991. Weathering of Explosives
for Twenty Years. Los Alamos National Laboratory, Report LA-
VI. SUMMARY
11931.
The detection of buried landmines has become a significant [2] Farmer, W.J., M.S. Yang, J. Letey, and W.F. Spencer. 1980.
challenge to the technical community. The challenges of Hexachlorobenzene: its vapor pressure and vapor phase diffusion in
soil. Soil Sci. Soc. Am. Proc. 44:676-680.
locating a buried object in near surface soils have driven the [3] Hogan, A., D. Leggett, T. Jenkins, P. Miyares. 1992. Results of
need for technology from traditional geophysical sensor Preliminary Analysis, Surface Contamination of Depot-Stored Land
systems to those looking for the chemical signatures of the mines, Prepared as a briefing guide for planning DARPA mine
explosives derived from within the landmine. The transport detector trials. USACRREL. June 23, 1992. Includes: Test Plan
Summary: Sampling of Mine Surfaces to Determine the Extent of
of the chemicals found in the explosive charge through soils Explosive Contamination.
is a complex process involving phase changes, interactions [4] Mayer, R., J. Letey, and W.J. Farmer. 1974. Models for
with the soils, and biochemical reactions. predicting volatilization of soil-incorporated pesticides. Soil Sci.
Soc. Am. Proc. 38:563-568.
[5] Jury, W.A., R. Grover, W.F. Spencer, and W. J. Farmer. 1980.
An investigation with computational simulation was Modeling Vapor Lossess of Soil-Incorporated with Triallate. Soil
initiated to explore the impacts of several of the various Sci. Soc. of Am. J., 44, 445-50 (May-June 1980).
input parameters with a pesticide screening model adapted [6] Jury, W.A., W.F. Spencer, and W.J. Farmer. 1983. Behavior
to the landmine chemical sensing problem. It was found Assessment Model for Trace Organics in Soil: I. Model
Description. J. Environ. Qual., Vol 12, no. 4, 558-564.
that the variations in the Henry’s Law constant are directly [7] Jury, W.A., W.J. Farmer, and W.F. Spencer. 1984a. Behavior
proportional to the changes seen in the steady-state surface Assessment Model for Trace Organics in Soil: II. Chemical
vapor flux. Small changes in the soil water partitioning Classification and Parameter Sensitivity. J. Environ. Qual., Vol. 13,
coefficient made large changes in the lag period and steady- no. 4, 567-572.
[8] Jury, W.A., W.F. Spencer, and W.J. Farmer. 1984b. Behavior
state surface vapor flux. The initial source term and
Assessment Model for Trace Organics in Soil: III. Application of
continuous source flux of chemicals from the landmines is Screening Model. J. Environ. Qual., Vol 13, no. 4, 573-579.
poorly understood and expected to be highly variable. [9] Jury, W.A., W.F. Spencer, and W.J. Farmer. 1984c. Behavior
These simulations found that with the baseline case, if the Assessment Model for Trace Organics in Soil: IV. Review of
continuous source flux was absent, there was no significant Experimental Evidence. J. Environ. Qual., Vol 13, no. 4, 580-586.
[10] Jury, W.A., D. Russo, G. Streile, and H. Abd. 1990. Evaluation
difference in the surface vapor flux or subsurface of Volatilization by Organic Chemicals Residing Below the Surface.
distribution at the end of the simulation period. This Water Resources Research vol 26, no 1, p 13-20, January 1990.
implies that the continuous source flux may be much less [11] McGrath, C. 1995. Review of Formulations for Processes
important than the initial surface contamination. It appears Affecting the Subsurface Transport of Explosives. US Army Corps
of Engineers, Waterways Experiment Station, Technical Report
that the magnitude of the surface vapor flux is directly IRRP-95-2, August 1995.
proportional to the amount of the initial surface [12] Millington, R.J., and J.M. Quirk. 1961. Permeability of porous
contamination. The biochemical half-life is another solids. Trans. Faraday Soc., 57:1200-1207.
parameter that is likely to have very different values [13] NGIC. 1995. Landmines and Demining: A Global Problem CD-
ROM. National Ground Intelligence Center, Charlottesville, VA.
depending on the location and climatic conditions. [14] Pennington, J.C. and W.H. Patrick. 1990. Adsorption and
Simulations over ten years showed that the steady state desorption of 2,4,6-trinitrotoluene by soils. J. Environ. Qual.
surface vapor flux declines steadily when the biochemical 19:559-567.
half-life becomes smaller than one year. Finally, the [15] Phelan, J.M. and S. W. Webb. 1997. Environmental Fate and
Transport of Chemical Signatures from Buried Landmines -
importance of heavy precipitation (such as a monsoon Screening Model Formulation and Initial Simulations. Sandia
season) followed by a dry season was explored. The impact National Laboratories, SAND97-1426. June 1997.
of the heavy precipitation was to lower the surface flux [16] Schroeder, P.R., C.M. Lloyd, and P.A. Zappi. 1994a. The
seven orders of magnitude; however, the evaporation Hydrological Evaluation of Landfill Performance (HELP) Model
User’s Guide for Version 3. EPA/600/R-94/168a. U.S.
period that followed returned the surface flux to Environmental Protection Agency, Risk Reduction Engineering
approximately the pre-monsoon surface flux. Laboratory, Cincinnati, Ohio.
[17] Schroeder, P.R., T.S. Dozier, P.A. Zappi, B.M. McEnroe, J.W.
VII. ACKNOWLEDGMENT Sjostrom, and R.L. Peyton. 1994b. The Hydrological Evaluation of
Landfill Performance (HELP) Model: Engineering Documentation
This work was performed with internal laboratory-directed for Version 3. EPA/600-R-94/168b. U.S. Environmental Protection
research and development funds to explore applications for Agency, Risk Reduction Engineering Laboratory, Cincinnati, Ohio.
miniaturized chemical sensors. Follow-on work to develop [18] Spangler, G.E.. 1975. Measurements on the Availability of
TNT Vapor from Antitank Mines. Report No: MERADCOM-R-
a numerical model is just beginning with funds from the 2159
Strategic Environmental Research and Development [19] Spencer, W.F., M.M. Cliath, W.A. Jury, and L.-Z. Zhang. 1988.
Program (SERDP) for UXO and from the Defense Volatilization of Organic Chemicals from Soil as Related to Their
Advanced Research Projects Agency (DARPA) for Henry's Law Constants. J. Environ. Qual. 17:504-509.
[20] Xue, S.K., I.K. Iskandar, and H.M. Selim. 1995. Adsorption-
landmines. Sandia is a multiprogram laboratory operated by Desorption of 2,4,6-Trinitrotoluene and Hexahydro-1,3,5-Trinitro-
Sandia Corporation, a Lockheed Martin Company, for the 1,3,5-Triazine in Soils. Soil Science, Vol. 160, No. 5. November
United States Department of Energy under Contract DE- 1995.
AC04-94AL85000.