Content uploaded by Lisa N Kates
All content in this area was uploaded by Lisa N Kates on Feb 15, 2021
Content may be subject to copyright.
Acute and chronic environmental effects of clandestine
Lisa N. Kates ⁎, Charles W. Knapp, Helen E. Keenan
Department of Civil and Environmental Engineering, University of Strathclyde, 75 Montrose Street, Glasgow G1 1XJ, United Kingdom
•Methamphetamine waste was assessed to estimate its environmental impact.
•Chemical oxygen demand surpassed European Union wastewater discharge regulations.
•Partition coefﬁcients K
were measured for input into a fugacity model.
•Fugacity model indicates waste components will remain in water for 15 to 37 days.
•Sediment compartment is not predicted to contain evidence of methamphetamine waste.
Received 5 March 2014
Received in revised form 16 June 2014
Accepted 16 June 2014
Available online xxxx
Editor: Adrian Covaci
The illicit manufacture of methamphetamine (MAP) produces substantial amounts of hazardous waste that is
dumped illegally. This study presents the ﬁrst environmental evaluation of waste produced from illicit MAP
manufacture. Chemical oxygen demand (COD) was measured to assess immediate oxygen depletion effects. A
mixture of ﬁve waste components (10 mg/L/chemical) was found to have a COD (130 mg/L) higher than the
European Union wastewater discharge regulations (125 mg/L). Two environmental partition coefﬁcients, K
, were measured for several chemicals identiﬁed in MAP waste. Experimental values were input into a
computer fugacity model (EPI Suite™) to estimate environmental fate. Experimental log K
from −0.98 to 4.91, which were in accordance with computer estimated values. Experimental K
from 11 to 72, which were much lower than the default computer values. The experimental fugacity model for
discharge to water estimates that waste components will remain in the water compartment for 15 to 37 days.
Using a combination of laboratory experimentation and computer modelling, the environmental fate of MAP
waste products was estimated. While fugacity models using experimental and computational values were very
similar, default computer models should not take the place of laboratory experimentation.
© 2014 Elsevier B.V. All rights reserved.
The study of pharmaceutical products and illicit drugs in the envi-
ronment rose to prominence in the early 2000s (Heberer, 2002;
Jones-Lepp et al., 2004). Researchers discovered that these com-
pounds are ubiquitous in river water, surface water, and wastewater
(Glassmeyer et al., 2005; Huerta-Fontela et al., 2008; Kasprzyk-
Hordern et al., 2009; Zuccato et al., 2008). As the world population
will be found in the aqueous environment. To date, most research
has focused on detection of these contaminants. Research is lacking
on the short-term and long-term effects of these products in the
environment, such as the adverse physiological and toxic effects on
ecosystems and the persistence of drug residues in both water and
Synthetic drugs that are manufactured illicitly, such as amphetamine
type stimulants (ATS), pose an additional risk to the environment from
clandestine drug laboratories. Methamphetamine (MAP) is the most
commonly produced ATS worldwide, typically manufactured in
clandestine laboratories close to the consumer (UNODC, 2012). The
illicit manufacture of MAP produces a large amount of harmful waste
that is often dumped illegally, creating a potential source of pollution.
One kilogramme of MAP produces 5 to 7 kg of waste that includes
many volatile, ﬂammable, and corrosive chemicals, as well as heavy
metals (White, 2004). Common routes of disposal include poured
down indoor plumbing, direct discharge into surface waters, or the
waste being burned and/or buried (US EPA, 2005). However, illicit
drug manufacturers are often not prosecuted for crimes related to
pollutingthe environment due tothe costs associated with prosecution
and lack of research in this area.
Science of the Total Environment 493 (2014) 781–788
⁎Corresponding author at: Intrinsik Environmental Sciences Inc., 736 8th Avenue SW,
Suite 1060, Calgary, Alberta T2P 1H4, Canada. Tel.: +1 403 237 0275.
E-mail address: firstname.lastname@example.org (L.N. Kates).
0048-9697/© 2014 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
To aid in the detection and prosecution of an illicit dumpsite, an
understanding of the chemical behaviour of the waste components is
essential. The fate of contaminants entering the environment is depen-
dent on their physicochemical properties, such as hydrophobicity,
vapour pressure, and stability (Walker et al., 1996). The use of partition
coefﬁcients in environmental modelling of organic chemicals is useful in
predicting the behaviour of a contaminant in the environment, con-
taminants such as MAP waste. However, an environmental model is
only as reliable as the information input into the scenario. Therefore,
to generate the most reliable model as possible, it is essential to gain
as much information about the dumpsite and the chemicals as feasible.
Information that will aid the reliability of the model includes the organic
carbon content of the dumpsite and partition coefﬁcients of the waste
Two useful partition coefﬁcients that are readily measureable in the
laboratory are the octanol–water partition coefﬁcient (K
) and the
organic carbon partition coefﬁcient (K
is a measure of the
hydrophobicity of a compound, which is inversely related to polarity.
Generally, as the hydrophobicity of a substance increases, so does its
toxicity. The lipophilicity content that permits compounds to enter
cell membranes is also linked to bioconcentration of organic substances
in aquatic organisms. High K
values are associated with high bio-
concentration factors (BCF).
is often referred to as the organic carbon normalized sorption
constant, which is a measurement of the partitioning behaviour
between water and the organic carbon fraction in a water–sediment
system. The K
value factors into account the percentage of organic
carbon present in the sediment, which can greatly inﬂuence the amount
values provide good information regarding the
fate and long-term consequences of an organic pollutant, short-term
changes may cause just as much if not greater harm. In order to assess
the acute environmental impact of MAP waste, the chemical oxygen de-
mand (COD) can be measured to determine possible oxygen depletion
in receiving waters. COD is an indicative value of water and wastewater
quality that measures the amount of oxygen consumed by organic pol-
lutants through oxidation.
This paper aims to assess initial environmental fate and impact of
organic waste products from clandestineMAP laboratories.This was un-
dertaken using computer modelling in conjunction with laboratory
2.1. Target compounds
The target compounds selected for study were based on waste
products identiﬁed in a preliminary proﬁling study (Kates et al.,
2012). In the proﬁling study, organic waste was collected from in-
house MAP synthesis following the Leuckart (Kunalan et al., 2009),
hypophosphorous (Vallely, 1995), or Moscow (Kunalan et al., 2009)
route. Not all of the compounds identiﬁed were available to purchase
as analytical standards, which limited the number of compounds avail-
able for this current study. The MAP waste chemicals tested in thisstudy
are summarized in Table 1. All chemicals in this study were purchased
as analytical grade standards from Sigma-Aldrich, UK. Solvents were
HPLC grade and all analyses were completed using high performance
liquid chromatography with a UV detector.
2.2. Chemical oxygen demand, COD
COD was determined using a commercially prepared reactor diges-
tion test tube kit with a range of 0–1500 mg/L oxygen (Hach-Lange,
UK). Samples were prepared according to the kit instructions and COD
General information on MAP waste compounds used in the current study.
λmax (nm) Molecular
MAP Final product C
N149.24 212 COD, K
BOX L C
NO 121.14 216 COD, K
BA L, M, H C
2,6-DTBP L, M, H C
O206.32 212 K
N-MA L C
NO 73.09 –COD, K
PHE L C
O 94.11 220 COD, K
P2P L precursor
O134.17 208 COD, K
1-P-1,2-P L C
148.16 216 K
L–Leuckart; M –Moscow; H –Hypophosphorous.
782 L.N. Kates et al. / Science of the Total Environment 493 (2014) 781–788
values were measured using a portable colorimeter (DR/850 Hach-
Lange, UK). MAP, P2P, N-MA, PHE, and BOX were tested as individual
compounds at concentrationsfrom 1 mg/Lto 100 mg/L. Those ﬁve com-
pounds were also tested as part of a mixture at concentrations from
0.01 mg/L to 100 mg/L/chemical. Thus the ﬁnal chemical concentrations
in the mixture solution ranged from 0.05 mg/L to 500 mg/L.
Solutions were made up in Nanopure water (Barnstead Nanopure,
ThermoFisher Scientiﬁc, UK). According to the kit manufacturer, the
estimated detection limit of the COD method is 30 mg/L (±16 mg/L).
2.3. Octanol–water partition coefﬁcient, K
(Eq. (1)) values were experimentally determined by reversed
phase high performance liquid chromatography (RPHPLC) following
the Organisation for Economic Co-operation and Development (OECD)
standard method 117 (OECD, 1989). The HPLC used was a Dionex Ulti-
Mate 3000 with a C18 column (Techsphere5ODS, 25 cm × 4.6 mm) and
a variable wavelength detector.
KOW ¼Chemical concentration in octanol½
Chemical concentration in water½
values are associated with high BCFs (Eq. (2)) as most
aquatic organisms will uptake organic pollutants through passive
BCF ¼Chemical concentration in biota½
Chemical concentration in water½
BCF has a linear relationship with K
, as shown in Eq. (3).
logBCF ¼alogKOW þbð3Þ
Three different artiﬁcial soils were prepared for the sorption
experiments to provide a range of organic carbon content and to
reduce matrix interferences from polluted site samples. Garden
humus (B&Q, UK), sand (Portland Builder's sand), and clay (WBB
Minerals, UK) were mixed with silt collected from a stream in
Calderglen Country Park, Glasgow, UK (55°44′57.48″N, 4° 8′34.40″
The moisture and organic carbon content of each soil was deter-
mined following ASTM Standard Method D2974 –07a (ASTM, 2007).
The pH was determined using a 1:1 (w/v) slurry of water and sediment,
which was stirred for 30 min, then left to stand for 1 h before a pH
reading was taken.
The composition and physical properties of each soil are shown in
Table 2. Each soil is classiﬁed as sandy loom according to British
Standard BS 3882:2007(British Standards, 2007).
2.5. Sediment–water partition coefﬁcient, K
values of selected MAP waste impurities were measured
following ASTM standard method E1195 –01 (ASTM, 2008). Sorption
of MAP, N-MA, PHE, BOX, P2P, and 2,6-DTBP was measured by
equilibrating them in a mixture of water and sediment at constant
temperature (20 °C ± 1 °C) in the dark. The amount of chemical
added was determined by taking into account its water solubility,
predicted adsorption coefﬁcient, and limit of detection (LOD) of the
HPLC-UV method used to quantify the amount of chemical left in the
aqueous phase (details below; LODs calculated following Miller and
Miller, 2010). Initial estimate of each chemical's adsorption coefﬁcient
was determined using Eq. (4), which predicts K
to within one order
of magnitude (Eq. (3) in ASTM, 2008).
lnKOC ¼−lnWs−0:01 MP−25ðÞþ15:1621ðÞ=1:7288 ð4Þ
water solubility, mg/mL
MP melting point, °C (for liquids at 25 °C, MP = 25)
Sediment to water ratios were calculated to achieve chemical sorp-
tion between 20 and 80%. With a ﬁxed aqueous volume of 10 mL, the
sediment to water ratios used were 1:2, 1:3, and 1:5. Using 20 mL
glass universal bottles ﬁtted with aluminium foil-lined caps, 1.0 mL of
1.0 mg/mL in water of MAP, N-MA, PHE, BOX, and P2P was added to
the sediment. The volume was brought to 10 mL using Nanopure
water (Barnstead Nanopure, ThermoFisher Scientiﬁc, UK), for a ﬁnal
concentration of 0.1 mg/mL.
The concentration of 2,6-DTBP was less because of its lower water
solubility. The water solubility of 2,6-DTBP is 2.5mg/L at 25 °C, however
one half of that concentration would not completely dissolve in water at
20 °C. As per the standard method, the solution was made up of 10%
acetonitrile (ACN). 2.0 mL of 1.25 mg/L of 2,6-DTBP was added to the
vials, for a ﬁnal concentration of 0.25 mg/L.
The contents of the vials were mixed on a roller shaker for 4 h. Vials
were centrifuged at 4500 rpm for 10 min and the supernatant ﬁltered
using a membrane syringe ﬁlter (0.45 μm; Millex MF-Millipore™). The ﬁl-
trate was added to 2 mL autosampler vials, to which bisphenol A (40 μLof
1.0 mg/mL in methanol) was added as internal standard. While bisphenol
A (BPA) is often found in efﬂuents, it is not related to the compounds of
interest from a clandestine MAP laboratory. By selecting an internal stan-
dard unrelated to the chemicals of interest, cross contamination issues
can be eliminated. The absence of BPA in the artiﬁcial soils was proven
in blank sample runs. The analytical method used to quantify the K
experiment was HPLC with a UV variable wavelength detector. Thus the
internal standard required a chromophore and an elution time that
would not interfere with the compounds of interest, as well as an elution
time that would not signiﬁcantly prolong the run time.
Samples were quantiﬁed using HPLC (Dionex UltiMate 3000) with a
C18 column (25 cm × 4.6 mm, Techsphere5ODS) and a variable wave-
length detector. Wavelengths were set according to the λmax in
Table 1. The mobile phase was a gradient of ACN and water as follows:
20% ACN, 80% H
O for 1 min, increasing to 40% ACN/60% H
5 min, and held for 7 min for a total run time of 13 min.
was calculated using Eq. (5).
total quantity of chemical sorbed to solids, μg
Boven-dry weight of solids, g
concentration of chemical in water, μg/mL
Composition and physical properties artiﬁcial soils.
Soil #1 Soil #2 Soil #3
Sand 41.56 58.17 78.85
Silt 5.55 7.29 9.41
Clay 2.79 3.65 4.70
Humus 50.10 30.89 7.04
pH 5.39 5.50 5.77
Moisture content (%, 105 °C) 13.90 8.75 2.11
TOC (%, 440 °C) 7.46 4.37 1.44
783L.N. Kates et al. / Science of the Total Environment 493 (2014) 781–788
And where G
is determined from HPLC quantiﬁcation as follows:
total quantity of chemical in control sample, μg
T total quantity of chemical left in water, μg
2.6. Organic carbon partition coefﬁcient, K
Using the organic carbon content of the three manufactured soils
(Table 2), Eq. (6) gives K
as a function of K
%OC Percent organic carbon of soil/sediment
2.7. Prediction of environmental fate —fugacity modelling
A user-friendly and freely available fugacity model can be found in
the United States Environmental Protection Agency's (US EPA) com-
puter modelling program Estimation Programs Interface (EPI) Suite™
(US EPA, 2012). EPI Suite™uses a Level III fugacity model, meaning it
assumes that the compartments (air, water, soil, and sediment) are
In the EPI Suite™fugacity model, it is possible to alter the emission
scenario. Emission values for each compartment were changed to create
a model that simulates dumping of chemicals directly into a body of
water. Default emission values for each compartment (air, water, and
soil) are 1000 kg/h. The emission values in this work were changed
to: air: 0 kg/h, water: 1000 kg/h, soil: 0 kg/h.
3. Results and discussion
3.1. Acute impact of MAP waste
COD can be used as an evaluative tool on the immediate impact of
chemical waste in the environment. Results of the COD tests on MAP
waste are shown in Tables 3 and 4. The results are deﬁned as amount
(mg) of oxygen consumed per litre of sample.
COD is an indirect measurement of oxygen consumption by organic
and inorganic chemicals in water (US EPA, 2009). The addition of
oxidisable contaminants into water systems can result in the depletion
of dissolved oxygen concentration (Harrison, 2007), which has the
potential to harm aquatic species.
The European Union legislated value for COD levels of chemicals
discharged into the environment is 125 mg/L (Council of European
Union Communities, 1991). For individual waste components
(Table 3), this threshold is reached at concentrations of 50 mg/L or
100 mg/L. For the mixture of the ﬁve chemicals (Table 4), the legis-
lated threshold is also exceeded at a total chemical concentration
of 50 mg/L. Comparing the results from the individual chemicals to
the results of the mixture, the MAP waste components do not display
additive effects. At a chemical concentration of 100 mg/L, the sum-
mation of COD from the individual components (1056 mg/L COD)
is comparable to the COD values of the mixture (1081 mg/L COD).
The difference is more pronounced at lower chemical concentra-
tions: at 50 mg/L the sum of the individual components (701 mg/L
COD) is over ﬁve times higher than the COD results from the mixture
(130 mg/L COD). This result suggests that the mixture is less harmful
than the individual components. With the exception of phenol, these
chemicals have few legitimate uses and are more likely to be found in
the environment as part of a mixture. The mixture is a better indica-
tion of a real-life dumpsite scenario.
While concentrations of MAP waste in the environment have not
been explored through case study, concentrations of 10 to 100 mg/L
are exceedingly low for environmental dumping. On many occasions,
clandestine MAP manufacturers will stock pile waste before disposing
of it. In such circumstances, several tons of waste may be discharged
in one location over a short period of time. The COD results indicate
that such an event has the potential to cause depletion in the amount
of dissolved oxygen to such an extent that it would become harmful
to aquatic organisms. In one case study in Canada, a clandestine drug
laboratory was seized based on the discovery of dead ﬁsh in a nearby
stream (Hugel, 2010). While it is probable that several factors likely
contributed to the death of the ﬁsh, the COD results from this experi-
ment indicate that oxygen depletion is certainly a potential contributor.
3.2. Chronic effects of MAP waste
Experimentally determined log K
values of MAP waste are com-
pared with EPI Suite™computer estimated log K
values in Table 5.
Once the K
values were determined, it was possible to calculate
BCFs based on the experimental K
and computer estimated K
also shown in Table 5.
Generally, small molecules with low K
values are more likely to be
water soluble, whereas larger molecules with high K
values are more
likely to dissolve in lipids and adsorb to solids. High K
values are also
associated with increased bioconcentration, which is linearly related to
, which is essentially a measurement of
polarity, can help to predict distribution and persistence of a compound
COD of individual MAP waste chemicals (mg/L COD; n = 2).
MAP P2P N-MA PHE BOX COD sum
1 BDL BDL BDL BDL BDL BDL
50 106 141 190 127 137 701
100 201 252 119 235 249 1056
BDL = below commercial kit detection limit of 30 mg/L.
COD of ﬁve MA waste chemicals in a mixture (mg/L COD; n = 2).
0.01 0.05 35 ± 35
0.1 0.5 BDL
1 5 BDL
10 50 130 ± 4
100 500 1081 ± 25
BDL = below commercial kit detection limit of 30 mg/L.
Experimental Log K
values compare to the EPI Suite™computer estimates and BCF
values based on experimental or EPI Suite™Log K
Chemical log K
Experimental EPI Suite™Experimental EPI Suite™
MAP 2.04 2.07 11.13 11.84
BOX 1.45 1.85 2.28 3.68
BA 1.40 1.10 2.27 1.55
2,6-DTBP 4.91 4.92 626.7 639.0
N-MA −0.98 −0.70 0.89 0.90
PHE 1.32 1.46 1.98 2.42
P2P 1.77 1.44 4.99 2.80
1-P-1,2-P 1.71 1.11 5.23 1.95
784 L.N. Kates et al. / Science of the Total Environment 493 (2014) 781–788
in the environment. Hydrophilic compounds tend to be dissolved and
distributed throughout surface water; corollary lipophilic compounds
tend to become associated with particulate matter, mostly sediments
(Walker et al., 1996).
Comparison of the experimental values with the EPI Suite™
shows little variation. Given the range of K
values and their
reporting on a log scale, the experimental and computer estimated
values are remarkably equivalent. The experimental and predicted
values of 2,6-DTBP were nearly identical, whereas the larg-
est difference in values was for 1-P-1,2-P, with a difference of 0.60.
Even though the log K
values are nearly equivalent over the log
scale, the differences become more apparent when the log function
is removed, which may affect environmental compartment distri-
bution, as investigated in Section 3.4.
The chemicals examined from MAP waste display the following
order of lipophilicity, from the lowest to the highest: N-MA b
PHE bBA bBOX b1-P-1,2-P bP2P bMAP b2,6-DTBP. Using the
linear relationship between K
and BCF (Eq. (3)), the same order
can be applied towards the tendency of these chemicals to bio-
accumulate in aquatic organisms.
The BCF values were estimated using the EPI Suite™model
(Table 5). As the estimates are dependent on K
values, BCFs were cal-
culated using both the default K
values and the experimental K
values. BAF values are expressed in L/kg wet weight of ﬁsh, which en-
ables comparison between different species by normalizing for lipid
content.If the percent lipid of the organism is known,this can be accom-
plished by dividing the wet weight (L/kg) by percent lipid, resulting in a
value with units of L/kg lipid weight.
Two chemicals, 1-P-1,2-P and P2P, exhibited notable increases in
BCF when calculated using the experimental log K
sponds to an increase in the potential uptake of these chemicals in
ﬁsh. The BCF of 1-P-1,2-P increased by over threefold, from 1.952 to
6.337, when calculated using the experimental log K
to the computer estimated log K
value. The uptake of P2P also in-
creased when BCF was calculated based on the experimental log K
compared to the computer estimated log K
, displaying a twofold
increase from 2.803 to 5.908 in BCF values.
MAP itself was predicted to be the second most lipophilic compo-
nent of the MAP waste. IfMAP were to bioaccumulate in the lipid layers
of ﬁsh, it would be unlikely to enter into the blood stream of the organ-
ism, meaning MAP would not cross the blood–brain barrier and would
not have the same physiological effects as an organism that ingested
the drug directly. However, other behavioural or toxic effects may
occur as the pollutant is slowly released into the general circulation.
Ghazilou and Ghazilou (2011) observed that male ﬁsh had increased
sexual activity when placed in a ﬁbreglass aquarium with concentra-
tions of MAP ranging from 0.1 to 1.0 mg/L. Changes were observed
after 2 and 5 days of exposure. However after 7 days of exposure
there was a signiﬁcant decrease in sexual activity, suggesting that the
ﬁsh grew acclimatisedto their environment andadapted to the dopami-
nergic effects of the drug.
values can be greatly inﬂuenced by pH, as described in
Bangkedphol et al. (2009) and Wells (2006).
3.2.2. Sorption of MAP waste onto sediment
Sediment properties have a great inﬂuence over the behaviour of
chemicals in the environment. The extent of adsorption of a chem-
ical onto sediment is an important factor in determining the ulti-
mate fate of chemicals in the environment. Adsorption is affected
by a number of soil properties, such as organic matter content,
clay content, and pH. The extent of adsorption is also affected by
the physicochemical properties of the compound, such as water sol-
ubility and K
(Drillia et al., 2005). The physical properties of the
artiﬁcial topsoils used in adsorption experiments are shown in
Of the six target compounds studied in the K
values were able to be determined accurately. For MAP,
matrix interferences prevented the resolution of a peak in the
HPLC chromatogram. The lambda max of MAP was previously de-
termined to be 212 nm, which corresponds to many compounds
present in the soil, such as humic matter. The elution time of MAP
also corresponded to the elution of humic matter despite numerous
program optimization attempts. Using an HPLC equipped with a UV
detector, it was not possible to separate MAP from soil matrix inter-
ferences. The other compound that could not be quantiﬁed on the
HPLC was 2,6-DTBP. Due to its low water solubility and high hydro-
phobicity, ﬁnal water concentrations were below the limits of
detection. Limits of quantiﬁcation for the HPLC method are as fol-
lows (calculated as per Miller and Miller, 2010): N-MA = 34 μg/L,
P2P = 5 μg/L, PHE = 5 μg/L, BOX = 4 μg/L.
were calculated using Eqs. (5) and (6),showninTable 6.
value represents the chemical's propensity to adsorb to
organic carbon. The K
is calculated from K
to be independent of
sediment organic carbon content.
The ASTM method makes the assumption that the main factor affect-
ing adsorption for non-polar organic chemicals is the organic carbon
content of the sediment. In this study, the chemicals under investigation
are fairly polar. In this case, other sediment properties may have greater
effect on theadsorption behaviour. Otherfactors include physical forces,
chemical forces, hydrogen bonding, hydrophobic bonding, electrostatic
bonding, coordination reactions, and ligand exchanges (Tan, 1998).
Adsorption of organic chemicals onto sediment surfaces is also in-
ﬂuenced by several physicochemical properties of the chemical itself.
Examples of those properties include the chemical nature of the ad-
sorbate, water solubility, dissociation capacity, surface charge density,
3.3. Correlation between log K
and log K
It has long been established that there is a linear relationship be-
(US EPA, 1996). In an effort to predict K
for chemicals that were not tested, a plot was constructed of log K
versus log K
(Fig. 1). Three lines are present on the plot: one line
from experimentaldata and two lines fromEPI Suite™estimated values.
EPI Suite™uses two different methods of estimating K
. One method
is based on K
values (Eq. (7)), the other based on molecular con-
nectivity index (MCI, Eq. (8)).
logKOC ¼0:8679 logKOW−0:0004 ð7Þ
logKOC ¼0:5213 MCI þ0:62 ð8Þ
Looking at the correlation coefﬁcients, the R
value from the ex-
perimental data is closer to one (0.978) than both EPI Suite™
values (0.938, 0.840), which indicates a stronger linear correlation.
The poorest correlation between log K
and log K
is from the EPI
Suite™method that calculates log K
from log K
. According to
the EPI Suite™methodology guide, log K
calculated from log
Experimentally determined log K
values for MAP waste components.
N-MA P2P PHE BOX
Soil 1 0.86 1.65 1.44 1.81
Soil 2 1.07 1.69 1.56 1.88
Soil 3 1.20 1.85 1.61 1.90
Average 1.04 1.73 1.54 1.86
10.96 53.70 34.67 72.44
785L.N. Kates et al. / Science of the Total Environment 493 (2014) 781–788
has an R
value of 0.877 (n = 68), and an R
value of 0.967 (n =
69) when calculated using MCI.
Using the equation of the line from the experimental data
(Eq. (9)), it is possible to calculate K
values for MAP, 2,6-DTBP,
1-P-1,2-P, and BA based on experimentally determined K
3.4. Prediction of environmental fate
3.4.1. EPI Suite™Computer Modelling
Equipped with experimental partition coefﬁcients, K
environmental modelling using a fugacity model was conducted. As
each model is only as reliable as the input data, it is important to have
reliable input parameters.
Three different scenarios were run for each chemical –two using the
values, the third using the experimental K
values. The only other parameters altered were the emission
values, as described in Section 2.7.Table 8 displays the compartment
distribution from the fugacity model using K
calculated from the de-
value. Table 9 displays the compartment distribution using
calculated from MCI, and Table 10 is the compartment distribution
from the fugacity model using experimental K
Since compartment distribution is dependent on the physicochemi-
cal properties of the chemicals, it is interesting to compare the three
different fugacity models using three different sources of K
values. For the model scenario of a simulated discharge into water, the
fugacity model in all three instances indicates that seven of eight
chemicals will overwhelmingly remain in the water compartment
value will play a determining factor in the sediment–water
partitioning. Looking at the K
values from all three scenarios, the EPI
Suite™values calculated using K
are within the same order of magni-
tude as the experimental K
values. However, the EPI Suite™K
values from MCI calculations are several orders of magnitude higher
for MAP, BOX, and 2,6-DTBP. This change resulted in a difference in
sediment–water distribution for MAP and BOX of 3%. For 2,6-DTBP,
there are greater disparities between each scenario in the sediment–
water compartments. The distribution in the water compartment
ranged from a high of 98% (Table 10) to a low of 65% (Table 9).
Even with the difference in K
values, all the chemicals from MAP
waste tested in this study will remain predominantly in the water com-
partment. This has implications for environmental sampling, indicating
that water samples should be taken as opposed to sediment samples.
The half-life of each chemical, except 2,6-DTBP, is predicted to be
15 days in the water compartment, with 2,6-DTBP having a half-life of
37.5 days. Given the transient nature of clandestine MAP laboratories,
if a dumpsite is discovered after 15 days, there is a possibility that the
laboratory has already been dismantled.
Under the controlled laboratory conditions, experimental K
values corresponded closely to values calculated using the comput-
er model EPI Suite™. In the absence of laboratory experimentation, the
fugacity model that most closely represents experimental results is the
model that uses K
as calculated from K
. However, under environ-
mental conditions, the partitioning behaviour may be different. Factors
that can inﬂuence the K
values are sediment properties,
salinity, temperature and pH, (Bangkedphol et al., 2009).
3.4.2. Model assumptions
EPI Suite™has several assumptions and limitations that must be
taken into account when interpreting the results. The model is designed
to be a screening tool and should not be used if measured values are
available. EPI Suite™uses a Level III Fugacity model which has several
assumptions of its own. The Level III model assumes steady state condi-
tions, but not equilibrium conditions. This means the model assumes
Fig. 1. Correlation between log K
and log K
values calculated using experimental K
values in Eq. (9).
Chemical Exp. log K
Calc. log K
MAP 2.04 1.77 58.82
BA 1.40 1.59 39.05
2,6-DTBP 4.91 2.57 369.33
1-P-1,2-P 1.71 1.68 47.62
786 L.N. Kates et al. / Science of the Total Environment 493 (2014) 781–788
that chemical concentrations in each compartment will approach zero
over time. The Level III model does not assume that each phase is in
equilibrium, meaning that if a chemical is released into one compart-
ment it can partition into the other compartments. In the Level III
model, a chemical is continuously discharged at a constant rate and
achieves a steady state condition when input and output rates are
equal (CCEMC, 2002; US EPA, 2012).
Chemical losses occur through two methods: reaction and advec-
tion. Reactions include biotic or abiotic degradation of the chemical in
each of the four compartments. Advection is the removal of a chemical
from a compartment through losses other than degradation, such as
bulk media transport via river currents. Advection processes are not
considered for the soil compartment. Additional assumptions of the
Level III model are that there are no direct emissions into the sediment
compartment and it cannot model ionizing or speciating chemicals
(CCEMC, 2002; US EPA, 2012).
There are several parameters that can be changed by the user in EPI
Suite™in order to create a chemical and site speciﬁc model; however,
there are also numerous parameters that cannot be changed. For exam-
ple, a ﬁxed temperature of 25 °C is assumed. That temperature will not
reﬂect many countries mean annual temperatures, nor will it take into
account daytime and seasonal variations.
While a site speciﬁc model can be approximated, the limitations in
setting parameters will prevent a truly site speciﬁc model from being
designed. This once again reinforces the need for laboratory experimen-
tation, particularly in environments that vary considerably from the
model default values. An additional limitation of the model is that mix-
tures cannot be evaluated. After understanding the assumptions and
limitations of the EPI Suite™fugacity model, its advantages are also im-
portant to note. Compared to other environmental models, such as a
mass balance model, the fugacity model is easy to understand and it
does not rely on units, but rather it is based on ratios therefore the
units cancel out. In order to properly use a mass balance model, it is
required to have estimated input concentrations of the chemicals. For
this study, concentrations of MAP waste have never been measured or
studied in a large scale, real-life scenario.
Previous research (Pal et al., 2011), measured the half-life value of
MAP to range from 131 to 502 days, which is in contrast with the EPI
Suite™estimated value in the soil of 30 days. In sediment, EPI Suite™
predicts a half-life of 135 days for MAP. The focus of this current study
was MAP waste products, not necessarily MAP itself, in a sediment–
water system. The US EPA uses EPI Suite™to evaluate new chemicals,
to help estimate harmfulness and persistence in the absence of experi-
mental data. The EPI Suite™model calculates half–lives based on a com-
bination of accumulated published data on 129 organic chemicals and
correlation with 233 test chemicals using the computer model, devel-
oped by Mackay et al. (1999).
The Pal et al. (2011) study was conducted under ideal laboratory
conditions. In a real–life dumpsite scenario, the chemicals will be ex-
posed to environmental conditions, such as temperature ﬂuctuations,
precipitation, and wind. Without conducting a mock–dumpsite experi-
ment, it may not be feasible to determine an absolute half-life value.
Additionally, each dumpsite will behave differently depending on the
biological activity, environmental conditions and organic carbon
Computer models cannot replace experimental values, as shown
by the difference in half-life values in this instance. However, in the
absence of the ability to conduct year-long experiments, a computer
model can serve as a screening tool to ﬂag chemicals that are estimated
to be harmful to the environment.
By using a combination of laboratory experimentation and computer
modelling, the environmental fate of MAP waste products was estimat-
ed. In the immediate term, the waste is likely to be harmful to aquatic
organisms based on the amount of oxygen consumed through the
oxidation reactions of the compounds. A mixture of the individual
waste components was found to consume more oxygen than the indi-
vidual chemicals. For longer-term implications in a discharge-to-water
scenario, the waste is likely to remain in the water compartment and
has a half-life of 15 to 37.5 days. The partitioning indicates that for
suspected water dumpsites of MAP waste, water samples should be
collected within 2 weeks in order to maximize detection. The analysis
of sediment samples is not predicted to contain evidence of clandestine
EPI Suite™fugacity model using K
calculated from default K
BA P2P N-MA PHE MAP 1-P-1,2-P BOX 2,6-DTBP
13 83 2 79 106 12 33 6506
1.10 1.44 −0.70 1.46 2.07 1.11 1.85 4.92
Compartment % t
Air 0 11 0.15 45 0 49 0 10 0.01 3 0 137 0 39 0.02 5
Water 99.7 360 99.2 360 99.8 360 99.5 360 99.4 360 99.7 360 99.6 360 72.7 900
Soil 0.02 720 0.1 720 0.02 720 0.02 720 0.01 720 0.02 720 0.05 720 0.03 1800
Sediment 0.24 3240 0.55 3240 0.2 3240 0.5 3240 0.6 3240 0.23 3240 0.3 3240 27.3 8100
% = Percen t of chemical mass in s peciﬁed compartment; t
EPI Suite™fugacity model using K
calculated from MCI.
BA P2P N-MA PHE MAP 1-P-1,2-P BOX 2,6-DTBP
21 93 3 187 893 10 813 9194
1.10 1.44 -0.70 1.46 2.07 1.11 1.85 4.92
Compartment % t
Air 0 11 0.15 45 0 49 0 10 0.01 3 0 137 0.01 39 0.01 5
Water 99.7 360 99.2 360 99.8 360 99.1 360 97 360 99.8 360 97.1 360 65.3 900
Soil 0.02 720 0.1 720 0.02 720 0.03 720 0.01 720 0.02 720 0.07 720 0.03 1800
Sediment 0.27 3240 0.55 3240 0.2 3240 0.9 3240 3.1 3240 0.2 3240 2.87 3240 34.7 8100
% = Percen t of chemical mass in s peciﬁed compartment; t
787L.N. Kates et al. / Science of the Total Environment 493 (2014) 781–788
From the experimental measurement of K
, a linear corre-
lation was established. K
can be measured very easily, while K
periments are time consuming and labour intensive. By using the
correlation between the two partition coefﬁcients, K
can be estimated
reliably through the measurement of K
. While fugacity models using
experimental and computational values were very similar, default com-
puter models should not take the place of laboratory experimentation.
The authors would like to thank Karl Bresee for a critical review of
the manuscript and for providing valuable feedback regarding the EPI
ASTM (American Society for Testing and Materials). Standard test method for moisture,
ash, and organic matter of peat and other organic soils (Designation: D2974 - 07a);
ASTM (American Society for Testing and Materials). Standard test method for determin-
ing a sorptionconstant (Koc) for an organic chemicalin soil and sediments (Designa-
tion: E1195-01); 2008.
Bangkedphol S, Keenan H, Davidson C, Sakultanimetha A, Songsasen A. The partition be-
havior of tributyltin and prediction of environmental fate, persistence and toxicity in
aquatic environments. Chemosphere 2009;77(10):1326–32.
British Standards. BS 3882:2007 speciﬁcation for topsoil and requirements for use; 2007.
CCEMC (Centre for Environmental Modelling and Chemistry). Level III model. http://
www.trentu.ca/academic/aminss/envmodel/models/VBL3.html, 2002. [March (last
accessed May 14, 2014)].
Council of European Union Communities. Directive 91/271/EEC. Off J Eur Commun 1991;
Drillia P, Stamatelatou K, Lyberatos G. Fate and mobility of pharmaceuticals in solid ma-
trices. Chemosphere 2005;60:1034–44.
Ghazilou A, Ghazilou S. Single and repeated exposure to methamphetamine induces al-
tered sexual behavior in male sailﬁn molly (Poecilia latipinna Lesueur) (Pisces). Afr
J Pharm Pharmacol 2011;5(13):1619–22.
Glassmeyer ST, Furlong ET, Kolpin DW, Cahill JD, Zaugg SD, Werner SL, et al. Transport of
chemical and microbial compounds from known wastewater discharges: potential
for use as indicat ors of human fecal contamination. Environ Sci Technol 2005;
Harrison RM. Principles of Environmental Chemistry. Cambridge, UK: Royal Society of
Heberer T. Occurrence, fate, and removal of pharmaceutical residues in the aquatic envi-
ronment: a review of recent research. Toxicol Lett 2002;131:5–17.
Huerta-Fontela M, Galceran MT, Martin-Alonso J, Ventura F. Occurrence of psychoactive
stimulatory drugs in wastewaters in north-eastern Spain. Sci Total Environ 2008;
Hugel J. Personal communication, Forensic Services Group, New South Wales PoliceForce,
Jones-Lepp TL, Alvarez D, Petty J, Huggins J. Polar Organic Chemical Integrative Sampling
(POCIS) and LC–ES/ITMS for assessing selected prescription and illicit drugs intreated
sewage efﬂuents. Arch Environ Contam Toxicol 2004;47(4):427–39.
Kasprzyk-Hordern B, Dinsdale RM, Guwy AJ. Illicit drugs and pharmaceuticals in the en-
vironment —Forensic applications of environmental data. Part 1: estimation of the
usage of drugs in local communities. Environ Pollut 2009;157(6):1773–7.
Kates LN, Gauchotte-Lindsay C, Nic Daéid N, Kalin RM, Knapp CW, Keenan HE. Prediction
of the environmental fate of methylamphetamine waste. In: Morrison RD, O'Sullivan
G, editors. Environmental forensics: proceedings of the 2011 INEF conference. Cam-
bridge, UK: RSC Publishing; 2012. p. 262–74.
Kunalan V, Nic Daéid N, Kerr WJ, Buchanan HA, McPherson AR. Characterization of route
speciﬁc impurities found in methamphetamine synthesized by the Leuckart and re-
ductive amination methods. Analytical Chemistry 2009;81(17):7342–8.
Mackay D, Shiu WY, Ma KC. Physical–chemical properties and environmental fate hand-
book, CRC netBASE 1999 CD-ROM. Boca Raton, FL US: Chapman and Hall/CRC Press;
Miller JN, Miller JC. Statistics and chemometrics for analytical chemistry. 6th ed. Harlow,
USA: Pearson Education Limited; 2010.
OECD (Organisation for Economic Co-operation and Development). Partition coefﬁcient
(n-octanol/water). high performance liquid chromatography (HPLC) method March
Pal R, Megharaj M, Kirkbride KP, Heinrich T, Naidu R. Biotic and abiotic degradation of il-
licit drugs, their precursor, and by-products in soil. Chemosphere 2011;85(6):
Tan KH. Principles of soil chemistry. 3rd ed. Marcel Dekker, Inc.; 1998
UNODC (United Nations Ofﬁce on Drugs and Crime). World Drug Report 2012, Vienna,
US EPA (United States Environmental Protection Agency). Soil screening guidance: tech-
nical background document. Part 5: Chemical-Speciﬁc Parameters; 1996.
US EPA (United States Environmental Protection Agency). RCRA hazardous waste identi-
ﬁcation of methamphetamine production process by-products. Report to congress;
US EPA (United States Environmental Protection Agency). Drinking water glossary:a dic-
tionary of technical and legal terms related to drinking water. Ofﬁce of Water/Ofﬁce
of Ground Water and Drinking Water; 2009.
US EPA (United States Environmental Protection Agency)Estimation Programs Interface
Suite for Microsoft Windows, v 4.11; 2012.
Vallely P. A single step proce ss for methampheta mine manufacture using
hypophosphorous acid. J Clandestine Lab Invest Chem Assoc 1995;5:14–5.
Walker CH, Hopkin SP, Sibly RM, Peakall DB. Principles of ecotoxicology. Bristol, UK: Tay-
lor & Francis; 1996.
Wells M. Log D
: key to understanding and regulating wastewater-derived contami-
nants. Environ Chem 2006;3:439–49.
White M. FSS report on methylamphetamine: chemistry, seizure statistics, analysis, syn-
thetic routes and history of illicit manufacture in the UK and the USA. UK: Forensic
Science Service; 2004.
Zuccato E, Castiglioni S, Bagnati R, Chiabrando C, Grassi P, Fanelli R. Illicit drugs, a novel
group of environmental contaminants. Water Res 2008;42(4–5):961–8.
EPI Suite™fugacity model using experimental K
BA P2P N-MA PHE MAP 1-P-1,2-P BOX 2,6-DTBP
39.05 53.70 10.96 34.67 58.82 47.62 72.44 369.33
1.40 1.77 −0.98 1.32 2.04 1.71 1.45 4.91
Compartment % t
Air 0 11 0.15 45.5 0 49.4 0 10 0.01 3 0 137 0.01 39 0.02 5.23
Water 99.6 360 99.4 360 99.8 360 99.6 360 99.6 360 99.6 360 99.5 360 97.9 900
Soil 0.02 720 0.08 720 0.02 720 0.02 720 0.01 720 0.03 720 0.06 720 0.04 1800
Sediment 0.34 3240 0.4 3240 0.2 3240 0.33 3240 0.42 3240 0.37 3240 0.47 3240 2.01 8100
% = Percen t of chemical mass in s peciﬁed compartment; t
788 L.N. Kates et al. / Science of the Total Environment 493 (2014) 781–788