Environmental Science and Pollution Research

https://doi.org/10.1007/s11356-019-04607-2

RESEARCH ARTICLE

Evaluating the dual porosity of landfilled municipal solid waste

Wenjie Zhang 1 & Mingfeng Lin 1

Received: 13 October 2018 / Accepted: 18 February 2019

# Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract
The objective of this paper was to study the pore-size distribution of municipal solid waste (MSW) and provide a basis for
understanding the mechanism of preferential flow in MSW. Two methods were used to investigate the pore-size distribution in
MSW. The first was an indirect method based on the soil-water characteristic curve (SWCC) and the second was a direct method
using nuclear magnetic resonance (NMR). Samples taken from different depths of a landfill were used. In the SWCC method,
SWCCs of the matrix region were obtained by a pressure plate extractor and SWCCs of the macropore region were determined by
water breakthrough tests. A SWCC equation is proposed by modifying the Van Genuchten equation to consider the dual-porosity
feature of MSW and the pore-size distribution was obtained based on the Young–Laplace equation. In the NMR method, the pore-
size distribution of MSW was obtained by analyzing the T2 curves. The results of the two methods were qualitatively similar and
both showed a dual-porosity characteristic of MSW. The average macropore radii of shallow, middle, and deep MSWs obtained
from the SWCC method are 0.193, 0.184, and 0.173 mm, and those obtained from the NMR method are 0.213, 0.138, and 0.145
mm, respectively. The proportion of macropores decreases with depth. The average micropore radii given by the SWCC method
are 0.022, 0.011, and 0.008 mm, and those given by the NMR method are 0.013, 0.011, and 0.008 mm, respectively. As the depth
and fill age increase, the average micropore size becomes smaller and the proportion of micropores increases. The volume ratios
of macropores obtained by the two methods are quantitatively quite different. The discrepancy is mainly caused by the different
test principles adopted by the two methods.

Keywords Municipal solid waste . Dual porosity . Soil-water characteristic curve . Nuclear magnetic resonance

Introduction studied preferential flow in MSW by model interpretation of

Municipal solid waste (MSW) is characterized by a complex
composition. There are macropores as well as matrix regions
in MSW, such that there exists fast flow through Bpreferential
pathways^ that bypass other parts of the MSW because of the
macropore characteristics. As a result, the common assump-
tion of a homogenous flow regime in MSW has been
questioned and preferential flow may be the reason why
existing models are not in agreement with actual field obser-
vations. Preferential flow has been identified as an important
flow pattern in MSW (Fellner et al. 2009; Manzur et al. 2016;
Aharoni et al. 2017; Feng et al. 2018). Rosqvist et al. (2005)
Responsible editor: Philippe Garrigues
* Wenjie Zhang
wjzhang2008@gmail.com
1
Department of Civil Engineering, Shanghai University,
Shanghai 200444, China
the water breakthrough curves and found that between 19 and
41% of the total water content participated in the transport of
solute through preferential flow paths. A two-domain ap-
proach was used by Fellner and Brunner (2010) to model
leachate generation from MSW landfills in which the prefer-
ential pathway is characterized by high permeability (Ks = 300
m/d) and zero retention capacity, while the bulk of the landfill
(the matrix domain) is characterized by low permeability (Ks
= 0.1 m/d) and high retention capacity.
The reason for existence of preferential flow in MSW lies
in the pore-size distribution. Pore-size distribution is one of
the important intrinsic properties of soil, which is associated
with the porosity and void ratio and controls the physical,
mechanical, and hydraulic behaviors of soil. However, our
knowledge of the pore-size distribution in MSW is limited.
Mercury intrusion porosimetry, which is based on the princi-
ple that mercury does not enter soil/rock pores with certain
radii unless a corresponding pressure is applied, was exten-
sively used to investigate the pore structures of porous media
(Zhang et al. 2018; Ma et al. 2018). However, at high

pressures, breaking of the particles and opening of closed
pores may significantly affect the accuracy of the measured
pore-size (Romero and Simms 2008). The principle of mea-
suring the soil-water characteristic curve (SWCC) with a pres-
sure plate extractor is similar. Air cannot enter certain wet soil
pores until a corresponding air pressure is applied.
Associating the air pressure with pore radius, the pore-size
distribution can be determined from the SWCC. However,
water will issue from macropores before any air pressure is
applied. As there are a lot of macropores in MSW, the pore-
size distribution based on such a SWCC may yield a signifi-
cant error (Solone et al. 2012).
The existence of macropores and micropores in MSW re-
sults in a bimodal SWCC. Bimodal SWCCs have been devel-
oped for gap-graded soils (Burger and Shackelford 2001;
Zhang and Chen 2005; Liu et al. 2013). However, bimodal
SWCCs of MSW have not previously been reported. Nuclear
magnetic resonance (NMR) has been used to determine the
pore-size distribution of soils and rocks (Jaeger et al. 2010;
Meyer et al. 2018). Compared with other traditional experi-
mental methods, NMR is more straightforward and more rap-
id, while providing convincing results (Tian et al. 2014). In
order to get a quantitative pore-size distribution, the surface
relaxivity, ρ2, needs to be determined. The ρ2 value of sand-
stone ranged from13.10 to 29.68 μm/s (Pape et al. 2009),
while the ρ2 of clastic rock was in the range of 1–10 μm/s
(Saidian and Prasad 2015). However, no NMR studies on
MSW have been reported yet. Compared with soils and rocks,
MSW is different in composition and pore structure, so the ρ2
value of MSW remains unknown.
The objective of this study was to determine the dual-
porosity character of MSW. A bimodal SWCC of MSW was
obtained by the combination of the pressure plate extractor
test and water breakthrough test. The SWCC method and the
NMR method were used to quantify the pore-size distribution.
The quantitative results will be beneficial in terms of the use of
a dual-porosity model to simulate water migration in MSW.
Theoretical background
Determination of pore-size distribution from SWCC
In order to calculate the pore-size distribution from SWCCs,
pores in the MSW are regarded as a collection of pipes with
different radii. The general form of the SWCC can be
expressed as:
∞ The pore-size distribution function can be written as fol-
θðψÞ ¼ θs ∫ψ gðψÞdψ ð1Þ
where ψ is matric suction (kPa), θ(ψ) represents the volumet-
ric water content when all pores with a matric suction more
than or equal to ψ are filled with water, θs is the saturated
Environ Sci Pollut Res
volumetric water content, and g(ψ) is the probability density
function of the pore radius represented by matric suction. For
dual-porosity media, the probability density function can be
regarded as a linear superposition of the macropore probabil-
ity density function, g1(ψ), and the micropore probability den-
sity function, g2(ψ). Then, Eq. (1) can be expressed as:
∞ ∞
θðψÞ ¼ θs1 ∫ψ g1 ðψÞdψ þ θs2 ∫ψ g2 ðψÞdψ
∞ ∞
¼ w1 θs ∫ψ g1 ðψÞdψ þ w2 θs ∫ψ g2 ðψÞdψ ð2Þ
where θs1 and θs2 are the saturated volumetric water contents
corresponding to the macropores and micropores, and w1
and w2 are weighting coefficients describing the propor-
tion of macropores and micropores. The sum of θs1 and
θs2 equals to θs, and the sum of w1 and w2 equals to 1.
In the test, the maximum matric suction is approximated
as the matric suction at residual water content, ψr, and
then Eq. (2) is rewritten as:
ψ ψ
θðψÞ ¼ w1 ðθs −θr2 Þ∫ψr1 g 1 ðψÞdψ þ θr1 þ w2 ðθs −θr2 Þ∫ψr2 g 2 ðψÞdψ þ θr2
¼ w1 ðθs −θr Þθ1 ðψÞ þ w2 ðθs −θr Þθ2 ðψÞ þ θr ð3Þ
¼ θs1 θ1 ðψÞ þ ðθfc −θr Þθ2 ðψÞ þ θr
where θ r1 and θ r2 are residual water content for the
macropores and micropores, ψr1 and ψr2 are the matric
suction corresponding to θr1 and θr2, and θfc is the field
capacity. Note that θ r 1 is approximately zero as
macropores can hardly hold water at high matric suction,
and then θr approximately equals to θr2.
By combining Eq. (3) and the Van Genuchten equation
(Van Genuchten 1980), a bimodal SWCC model can be ob-
tained as:

mi
2 1
θe ¼ ∑ wi ð4Þ
i¼1 1 þ ðαi ψÞni
where θe is effective volumetric water content defined by θe =
(θ − θr) / (θs − θr), i = 1 denotes macropores, and i = 2 denotes
micropores, and αi, ni, and mi are fitting parameters.
The relation between matric suction, ψ, and pore radius, r,
can be described by the Young–Laplace equation:
2T cosα
ψ¼ ð5Þ
r
where T is surface tension (N/m), and α is contact angle (°).
Under the assumption of a complete wetting interface and a
temperature of 20 °C, Eq. (5) gives ψ = 0.1455/r.
lows:

mi
θ−θr 2 1
F ðrÞ ¼ ¼ ∑ wi ð6Þ
θs −θr i¼1 1 þ ðαi ⋅0:1455=rÞni
Environ Sci Pollut Res

The pore-size probability density function is obtained by Materials and methods

taking the derivative of Eq. (6) with respect to r:

mi þ1 Materials
2 ðαi ⋅0:1455=rÞni 1
f ðrÞ ¼ ∑ wi mi ni ⋅ ni ð7Þ
i¼1 r 1 þ ðαi ⋅0:1455=rÞ
Samples were collected from the Qizishan MSW landfill,
By fitting the measured SWCC with Eq. (4), parameters αi, which is located in a valley about 13 km to the south of
ni, and mi can be obtained, then the pore-size distribution can Suzhou city, China. This facility commenced operations in
be achieved. July 1993. The first stage of the project had disposed around
5.6 MT of MSW until 2008 when it was closed. The second
stage of the project was constructed in 2006 and started oper-
Determination of pore-size distribution from NMR ation in 2008. At present, almost 50% of the MSW in this city
was deposited in the landfill at a loading rate of 1200 t/day (the
The NMR signal is an exponential decay, characterized by rest 50% was mainly dealt with by incineration). According to
initial signal amplitude and distribution of the transverse re- the record of filling, the deposit age of the samples collected
laxation times, T2. Since the NMR T2 of targeted molecules from the depth of 0–10 m was 0–3 years, the deposit age of the
reflects the structural arrangements of surrounding water mol- samples collected from the depth of 10–20 m was 3–6 years,
ecules, it is expected that the measurement of the T2 distribu- and the deposit age of the samples collected from the depth of
tion can shed insights into the microscopic mechanisms of 20–30 m was 6–9 years. The compositions and properties of
pore water distribution in a porous media. For a water- these specimens are distinct and, due to compaction and deg-
saturated porous material, T2 is dominated by the surface re- radation, these parameters vary with depth and age. In order to
laxation time (Coates et al. 1999) and can be expressed as avoid the discreteness caused by heterogeneity, and thus
achieve a more general trend, the specimens were divided into
1 1 1 1
¼ þ þ ð8Þ three types by age and depth (Table 1). Composition and dry
T 2 T 2Surface T 2Bulk T 2Diffusion density of every sample used in the SWCC tests and NMR
where T2 is transverse relaxation time; T2Bulk is the transverse tests were controlled artificially according to the statistical
bulk fluid relaxation time; T2Surface is the transverse surface parameters in Table 1.
relaxation; T2Diffusion is the diffusion relaxation time that ac- About 200 g of each MSW sample was used to determine
counts for transverse relaxation in an inhomogeneous magnet- the particle-size distribution (ASTM D422). Particle-size dis-
ic field. According to Kleinberg (2006), at low-field NMR, tributions for shallow, middle, and deep MSW are shown in
both T2Bulk and T2Diffusion are negligible compared with Fig. 1. As the depth and age increase, the content of large
T2Surface. And the T2 relaxation rate, 1/T2, is proportional to particles decreases while the content of fine-grained particles
the surface to volume (S/V) ratio of the pore (Behroozmand increases. The average particle size, d50, decreases from 19.89
et al. 2015), so Eq. (8) can be expressed as: to 0.84 mm. The uniformity coefficients Cu of shallow, mid-
  dle, and deep MSW samples are 173, 121, and 21, and the
1 1 S curvature coefficients Cc are 0.54, 0.17, and 0.73, respectively.
≈ ¼ ρ2 ð9Þ
T 2 T 2Surface V These values illustrate a wide variety of particle sizes and a
lack of medium-sized particles.
where ρ2 is the surface relaxivity coefficient, which is a char-
acteristic of magnetic interactions at the fluid-solid interface. Table 1 Parameters of landfilled wastes at different depths
In NMR, water in smaller pores experiences a greater surface
relaxation, and thus relaxes faster than that in larger pores. As MSW type Shallow Middle Deep
described by Eq. (9), the smaller pore has higher S/V value, Depth (m) 0–10 10–20 20–30
corresponding to a higher T2 relaxation rate. Age (years) 0–3 3–6 6–10
In order to relate T2 with pore-size distribution, pore geom- Bulk density (g/cm3) 0.79 0.97 1.25
etry needs to be assumed. To be consistent with the previous Dry density (g/cm3) 0.41 0.60 0.81
SWCC method, pores in MSW are assumed to be circular Specific gravity 1.51 1.91 2.10
tubes with the radius r, so Eq. (9) can be expressed as:
Porosity (%) 71.43 66.67 61.54
1 2ρ2 Content of plastic (%) 23.5 11.83 4.91
≈ ð10Þ Content of metal and glass (%) 3.28 1.18 1.10
T2 r
Content of paper, wood, fiber etc. (%) 20.46 13.49 3.75
Equation (10) indicates that T2 is proportional to the pore Content of humus (%) 17.42 16.83 14.92
radius. Therefore, the pore-size distribution can be determined Content of cinder and dust (%) 35.34 56.67 75.32
from T2 provided that the surface relaxivity ρ2 is known.

Fig. 1 Particle-size distributions of municipal solid waste (MSW)
Pressure plate extractor test
After cutting large-sized objects (including plastic bag
and sheeting) into less than 2 cm, MSW sample was filled
into a PVC container (diameter 7.0 cm and height 5.2 cm)
to form a column. The MSW column was first saturated
with tap water and then put into a volumetric pressure
plate extractor (Soil Moisture Ltd., 2 bar) to get the
SWCC. Outflow of water from macropores is inevitable
as soon as the MSW sample is removed from the satura-
tor. This part of water will discharge from the extractor as
the first-level air pressure (3 kPa in this study) is applied,
such that there is always a straight line between 0 and
3 kPa on the SWCC (Fig. 2). The lines are steep and
Fig. 2 Soil-water characteristic curves of municipal solid waste (MSW)
measured by volumetric pressure plate extractor
Environ Sci Pollut Res
cover a large range of water content values. In order to
determine the pore-size distribution in this suction range,
more data are needed. As only the 3-kPa point was ob-
tained in this test, this part of the SWCC cannot be used
to determine the pore-size distribution in this suction
range. In addition, the drainage curves and the wetting
curves are quite different. This can be attributed to the
different flow paths of wetting and drainage conditions,
and a dual-porosity structure (Stoltz et al. 2012).
Water breakthrough test
Macropores were defined as those pores that have
drained when the soil reaches field capacity. The re-
maining smaller pores were defined as micropores and
the corresponding region was termed the matrix region.
In this test, a MSW column was first saturated with tap
water and then allowed to drain freely until the field
capacity was reached. During this process, the amount
of drained w ater equals to the volum e of the
macropores. The field capacity of the shallow, middle,
and deep MSW samples is 38.45, 42.15, and 45.80%,
respectively. As the total porosity of 71.52, 66.70, and
60.54% is concerned, the macroscopic porosity accounts
for 46.23, 36.81, and 24.35% of the total porosity,
respectively.
Water was added to the fully drained column by main-
taining a constant water depth of 2 cm. The effluent
volume was recorded at intervals of 5 s until a steady-
state outflow was reached (Fig. 3). During this process,
effluent from the largest macropores appeared first. As
effluent from the smallest macropores appeared, all the
macropores transmitted water and the flow reached a
steady state. Macropore radii were calculated from the
Fig. 3 Water breakthrough curves of landfilled waste
Environ Sci Pollut Res

water breakthrough curves based on the basic flow equa- Table 2 Proportion of macropore volume to total volume of MSW
tion in circular pipes (Radulovich et al. 1989): Pore radius (mm) Shallow MSW Middle MSW Deep MSW

Q ¼ Av ¼ πr τL=t 2
ð11Þ 1.45–1.3 2.15% – –
1.2–1.3 2.54% – –
combined with the Poiseulle equation:
1.1–1.2 3.49% 1.63% 1.23%
Q ¼ πr4 ΔP=ð8μτLÞ ð12Þ 1.0–1.1 5.64% 2.10% 0.77%
0.9–1.0 3.76% 3.19% 0.81%
then solving for the pore radius: 0.8–0.9 2.78% 4.37% 1.11%
0.7–0.8 3.22% 4.00% 1.87%
r ¼ τL½8μ=ðtΔPÞ1=2 ð13Þ
0.6–0.7 – 0.92% 1.79%

where Q is flow (m3·s−1), A is pore area (m2), v is velocity (m3·

s−1), τ is tortuosity, r is pore radius (m), L is sample length (m),
t is time since water was added (s), ΔP is pressure (Pa), and μ Based on the Young–Laplace equation (Eq. (5)), the
is the viscosity of water (Pa·s). The assumption of circular matric suction and the corresponding volumetric water
pores should be addressed here. The aim of studying MSW content of the macropore region were calculated from
pores was to investigate the water flow pattern or flow capa- Table 2. They were coupled to the SWCC of the matrix
bility and so pores were equalized to circular tubes and the region to yield a complete SWCC (Fig. 4). In the 0–3-
radii were termed equivalent radii. kPa suction range, the water content first decreases signif-
In order to determine tortuosity, τ, the boundary between icantly and then becomes constant. The steep part on the
the micropore and macropore should be specified. No speci- SWCC indicates many macropores in this range (0–1 kPa)
fied boundary value has been reported. In practice, a matric whereas the less steep part (1–3 kPa) indicates a lack of
suction range of 0.3 to 10 kPa has been used as the limit pores corresponding to this suction range.
(Steinberg et al. 2005; Kuncoro et al. 2014), which corre-
sponds to the air pressure needed to completely discharge Nuclear magnetic resonance test
water from large pores. A matric suction value of 1 kPa is
used in this paper, corresponding to an equivalent radius of A customized container with an inner diameter of 4.6 cm
0.145 mm. The theoretical travel time can be calculated from and a height of 4.5 cm was used to hold the MSW sam-
Eq. (13) by taking r = 0.145 mm. Then, τ is determined from ples in the NMR test. The container is made of
the ratio of the actual measured time to the theoretical travel polytetrafluoroethylene, which does not influence the
time. The calculated τ is 7.7, 8.2, and 9.0 for shallow, middle, main magnetic field and produces no signal during the
and deep MSW samples, respectively. The Reynolds number test. The same MSW samples as in the SWCC test were
corresponding to the limit radius was calculated to be 1.11, used, except that metal pieces were removed from the
1.00, and 0.94 for shallow, middle, and deep MSW samples.
The values coincide with the critical Reynolds number for
Darcy’s Law which was reported to be a certain value
between 1 and 10 (Bear 1972). As preferential flow is
characterized by rapid flow in macropores that cannot be
described by Darcy’s Law, the 1-kPa limit here reflects the
essence of preferential flow.
The total effluent time was partitioned at certain time inter-
vals, such that each radius calculated from Eq. (13) represents
the average of a range of pores corresponding to the time
interval. Flow rate through the pore of midpoint radius for
interval i, Qmi, was calculated from Eq. (12). The measured
effluent volume during interval i, ΔQi, is divided by Qmi to
give the number of macropores in this interval. Considering
the average radius and column height, the approximate pore
area and volume can be calculated. The pore volume of each
interval divided by the total MSW volume gives the propor-
tion of pores in each interval (Table 2). It can be seen that the
amount and size of macropores both decrease with depth. Fig. 4 Complete SWCC of landfilled waste and the regression

samples. The NMR tests were performed at a frequency
of 21 MHz at 20 ± 1 °C. The echo time was set to 0.2 ms
and 18,000 echoes were acquired. Unlike rocks and soils,
the organic matter and plastics in MSW contain hydrogen
protons, so the dry MSW also produces signals.
Therefore, the free induction decay curves of dry MSW
samples were collected first and were taken as the back-
ground. Then, the saturated samples were tested and the
free induction decay curves were determined by compar-
ison with the background values. Relaxometry T2 was
obtained by applying the Fourier transform to the free
induction decay curves, as shown in Fig. 5. The T2 curve
shows a two-peaks profile, indicating a dual-porosity
characteristic of MSW.
Surface relaxivity coefficient
The surface relaxivity, ρ2, is an important parameter in the
NMR measurements (Silletta et al. 2018). In order to
quantitatively determine the pore-size distribution from
T2, the surface relaxivity must be known. However, the
ρ2 value of MSW has not been reported so far. In this
study, a well-accepted NMR-permeability equation,
known as Schlumberger-Doll Research (SDR) equation
developed by Kenyon et al. (1988), was used to determine
the ρ2 value of MSW which is
K ¼ Cn4 T 22LM ð14Þ
where K is the intrinsic permeability (m2), n is the porosity
(listed in Table 1), and T2LM is the geometric mean value of
the T2 distribution (Daigle and Dugan 2009). From Fig. 5,
T2LM corresponding to shallow, middle, and deep MSW
was calculated to be 121.95, 150.69, and 71.9 ms. The
Fig. 5 T2 distribution curves of landfilled wastes
Environ Sci Pollut Res
constant C is expected to be proportional to the square of
ρ2 (Kleinberg et al. 2003). Therefore, Eq. (14) is written as:
K ¼ ρ22 n4 T 22LM ð15Þ
Then ρ2 can be expressed as
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
K
ρ2 ¼ ð16Þ
n4 T 22LM
The intrinsic permeability, K, can be calculated from the
saturated hydraulic conductivity, ks, as
ksμ
K¼ ð17Þ
γw
where μ is the viscosity of water (Pa·s) and γw is the bulk
density of water (N·m−3). The previously described water
breakthrough test can also be approximated to a constant head
permeability test when a steady-state outflow is reached.
Based on the test results, the saturated hydraulic conductivity,
ks, of the shallow, middle, and deep MSW was calculated to be
2.10 × 10−4, 1.12 × 10−4, and 4.47 × 10−5 m/s, respectively.
From Eq. (17), the intrinsic permeability, K, was calculated to
be 2.15 × 10−11, 1.15 × 10−11, and 4.59 × 10−12 m2 for shallow,
middle, and deep MSW, respectively.
From Eq. (16), the ρ2 value of shallow, middle, and deep
MSW is calculated to be 74.58, 50.56, and 78.42 μm/s, re-
spectively. As ρ2 is known, the pore-size distribution can be
achieved from Eq. (10).
Results and discussion
Pore-size distribution obtained by the SWCC method
The complete SWCCs are fitted by the bimodal SWCC model
(Fig. 4). The model has high fitting accuracy (R2 > 0.998) and
fitting parameters are listed in Table 3. There is an approxi-
mately horizontal section on the fitted SWCC (Fig. 4) between
the low-suction part and the high-suction part, which reflects
that the drainage from macropores has ended whereas the
drainage from the matrix region has not yet started. The hor-
izontal sections on the three curves correspond to volumetric
water contents of 38.5, 42.2, and 45.8%, which are in good
agreement with the previously measured field capacity. The
residual water contents obtained from the curve fits were 17.0,
25.3, and 28.6% for the shallow, middle, and deep MSW
samples. These water contents all correspond to a matric suc-
tion of about 100 kPa on the fitted SWCC curves. The residual
water contents were much higher than common soils such as
clay and sand, which indicates a unique pore distribution of
MSW. There is a large number of macropores in MSW, which
separates the matrix region into many individual parts. As the
Environ Sci Pollut Res

Table 3 Parameters of the bimodal water retention curves

Samples w1 θs1 α1 n1 θs2 α2 n2 θr θs R2

Shallow MSW 0.6062 0.3307 1.3270 7.9480 0.3845 0.1514 3.1210 0.1697 0.715 0.999
Middle MSW 0.5932 0.2455 1.2646 7.5960 0.4215 0.0755 3.4030 0.2532 0.667 0.999
Deep MSW 0.4607 0.1474 1.1906 4.6430 0.4580 0.0544 3.7060 0.2855 0.605 0.998

air pressure in the extractor increased, water in the macropores
was drained and, finally, some water-filled micropores be-
came unconnected to each other. The water retained in these
micropores could not be drained further and is regarded as a
part of the residual water content.
By putting the fitted parameters (Table 3) into Eq. (7), the
pore-size probability density was achieved. Then, the volume-
ratio curves were obtained as shown in Fig. 6. The Y-axis in
Fig. 6 is the volume ratio, i.e., the ratio of pore volume at a
certain pore radius to the total pore volume. As depth in-
creases, the volume ratio of macropores decreases, whereas
the volume ratio of micropores increases. The X-axis is the
pore radius. In total, 2000 radius values (from 0.0001 to 1
mm) were used to calculate the pore-size distribution and the
sum of the corresponding volume ratio equals to 100%. The
three curves in Fig. 6 are all bimodal. Each pore-size distribu-
tion curve can be divided into a macropore part and a micro-
pore part. The average macropore radii of shallow, middle,
and deep MSW samples are 0.193, 0.184, and 0.173 mm,
and the average micropore radii are 0.022, 0.011, and 0.008
mm, respectively. They both decrease with depth.
Pore-size distribution obtained by the NMR method
Based on the calculated ρ2 values, the pore-size distribution
can be determined from T2 by Eq. (10). The volume-ratio
distribution of MSW is shown in Fig. 7. The distribution
curves all show two-peak profiles. By comparison with the
three curves, it can be seen that the volume ratio of
macropores decreases with depth, and the volume ratio of
micropores increases with depth. The ranges of macropore
and micropore radii are close to that measured by the SWCC
method (Fig. 6). The lowest point between two peaks of each
pore-size distribution curve is taken as a distinction between
macropores and micropores, which all approximate a pore
radius of about 0.05 mm. By taking 0.05 mm as the limit,
the average macropore radii of shallow, middle, and deep
MSW are 0.213, 0.138, and 0.145 mm, and the average mi-
cropore radii are 0.013, 0.011, and 0.008 mm, respectively.
The average macropore and micropore radii all decrease with
depth.
Discussion
By comparison with Figs. 6 and 7, it shows that the results of
the SWCC method and the NMR method are the same in the
following aspects. First, the pore-size distribution curves both
show two-peak profiles, indicating the dual-porosity character
of MSW. Second, the variation trends of the macropore/
micropore volume ratio and the pore-sizes with depth are ex-
actly the same. Third, the measured pore ranges and the aver-
age macropore/micropore radii are close (Table 4).

Fig. 6 Volume-ratio distribution of pores determined from SWCC Fig. 7 Volume-ratio distribution of pores determined by NMR

Table 4 The average pore radii
for each type of waste determined (mm) Average radii of macropores
by the two methods
Shallow MSW Middle MSW
SWCC 0.193 0.184
NMR 0.213 0.138
The results of the two methods are quantitatively quite
different as volume ratio of macropores is concerned. The
discrepancy was mainly caused by different test principles
adopted by the two methods, and may also relate to some
extent with the difference in MSW samples due to the com-
plex components. Furthermore, the pores in MSW are flat in
shape in most cases, which are far from circular tubes assumed
in this study. Adopting the concept of equivalent radii is not to
reveal the actual pore shape or size, but to describe the flow
capacity, and to some extent make the study possible.
Variation of macropore/micropore limit in the SWCC method
will change the calculated tortuosity, τ, and hence change the
ranges of macropores. The ρ2 value of MSW has direct influ-
ence on the pore values. Although the variation of macropore/
micropore limit and the ρ2 value will not change the two-peak
profiles of the pore-size distribution curves, more attentions
and further validations are still needed.
Conclusions
Pore-size distributions of shallow, middle, and deep MSW
were investigated by a SWCC method and a NMR method
in this study. Complete SWCCs of MSW were achieved by a
pressure plate extractor and water breakthrough tests. The
pore-size distribution curve achieved from the complete
SWCC can be divided into a macropore part and a micropore
part. The SWCC method shows that average macropore radii
of shallow, middle, and deep MSWs are 0.193, 0.184, and
0.173 mm, and the average micropore radii are 0.022, 0.011,
and 0.008 mm, respectively. The average macropore radii
of shallow, middle, and deep MSWs obtained from the
NMR method are 0.213, 0.138, and 0.145 mm, and the
average micropore radii are 0.013, 0.011, and 0.008 mm,
respectively. The proportion of macropores decreases
with buried depth whereas the proportion of micropores
increases with depth. The results of the SWCC method
and the NMR method are qualitatively similar in two-
peak profile and the average sizes of macropores or mi-
cropores; however, they are quantitatively quite different
in volume ratio of macropores. The discrepancy between
the results is mainly attributed to the different test prin-
ciples adopted by the two methods, which is worthy to
be studied further.
Environ Sci Pollut Res
Average radii of micropores
Deep MSW Shallow MSW Middle MSW Deep MSW
0.173 0.022 0.011 0.008
0.145 0.013 0.011 0.008
Funding information This research was financially supported by the
National Natural Science Foundation of China under Grant Nos.
41772300 and 51478256.
Publisher’s note Springer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.
References
Aharoni I, Siebner H, Dahan O (2017) Application of vadose-zone mon-
itoring system for real-time characterization of leachate percolation
in and under a municipal landfill. Waste Manag 67:203–213
Bear J (1972) Dynamics of fluids in porous media. American Elsevier
Publishing Company, Inc
Behroozmand AA, Keating K, Auken E (2015) A review of the principles
and applications of the NMR technique for near-surface characteri-
zation. Surv Geophys 36:27–85
Burger CA, Shackelford CD (2001) Evaluating dual porosity of pelletized
diatomaceous earth using bimodal soil-water characteristic curve
functions. Can Geotech J 38:53–66
Coates GR, Xiao L, Prammer MG (1999) NMR logging principles and
applications. Halliburton Energy Services, Houston, USA
Daigle H, Dugan B (2009) Extending NMR data for permeability esti-
mation in fine-grained sediments. Mar Pet Geol 26:1419–1427
Fellner J, Brunner PH (2010) Modeling of leachate generation from
MSW landfills by a 2-dimensional 2-domain approach. Waste
Manag 30:2084–2095
Fellner J, Döberl G, Allgaier G, Brunner PH (2009) Comparing field
investigations with laboratory models to predict landfill leachate
emissions. Waste Manag 29:1844–1851
Feng SJ, Bai ZB, Chen HX, Cao BY, Lu SF, Chen YM (2018) A dual-
permeability hydro-biodegradation model for leachate recirculation
and settlement in bioreactor landfills. Enviro Sci Pollut R 25:14614–
14625
Van Genuchten MT (1980) A closed-form equation for predicting the
hydraulic conductivity of unsaturated soils. Soil Sci Soc Am J 4:
892–898
Jaeger F, Shchegolikhina A, Van As H, Schaumann GE (2010) Proton
NMR relaxometry as a useful tool to evaluate swelling processes in
peat soils. J Magn Reson 3:27–45
Kenyon WE, Day PI, Straley C, Willemsen JF (1988) A three-part study
of NMR longitudinal relaxation properties of water-saturated sand-
stones. SPE Form Eval 3:622–636
Kleinberg RL (2006) Nuclear magnetic resonance pore-scale investiga-
tion of permafrost and gas hydrate sediments. Geo Soc London
Special Pub 267:179–192
Kleinberg RL, Flaum C, Griffin DD et al (2003) Deep sea NMR: methane
hydrate growth habit in porous media and its relationship to hydrau-
lic permeability, deposit accumulation, and submarine slope stabili-
ty. J Geophys Res Solid Earth 108(B10):2508
Kuncoro PH, Koga K, Satta N, Muto Y (2014) A study on the effect of
compaction on transport properties of soil gas and water. II: Soil
pore structure indices. Soil Tillage Res 143:180–187
Environ Sci Pollut Res
Liu S, Noriyuki Y, Liu Q, Kiyoshi O, Hazarika H (2013) Bimodal and
multimodal descriptions of soil-water characteristic curves for struc-
tural soils. Water Sci Technol 67:1740–1747
Ma Y, Wang G, Ye G, Hu J (2018) A comparative study on the pore
structure of alkali-activated fly ash evaluated by mercury intrusion
porosimetry, N2 adsorption and image analysis. J Mater Sci 53:
5958–5972
Manzur SR, Hossain MS, Kemler V, Khan MS (2016) Monitoring extent
of moisture variations due to leachate recirculation in an ELR/
bioreactor landfill using resistivity imaging. Waste Manag 55:38–48
Meyer M, Buchmann C, Schaumann GE (2018) Determination of quan-
titative pore-size distribution of soils with 1H NMR relaxometry.
Eur J Soil Sci 69:393–406
Pape H, Arnold J, Pechnig R, Clauser C, Talnishnikh E, Anferova S,
Blümich B (2009) Permeability prediction for low porosity rocks
by mobile NMR. Pure Appl Geophys 166:1125–1163
Radulovich R, Solorzano E, Sollins P (1989) Soil macropore-size distribu-
tion from water breakthrough curves. Soil Sci Soc Am J 53:556–559
Romero E, Simms PH (2008) Microstructure investigation in unsaturated
soils: a review with special attention to contribution of mercury
intrusion porosimetry and environmental scanning electron micros-
copy. Geotech Geol Eng 26:705–727
Rosqvist NH, Dollar LH, Fourie AB (2005) Preferential flow in munic-
ipal solid waste and implications for long-term leachate quality:
valuation of laboratory-scale experiments. Waste Manag Res 23:
285–295
Saidian M, Prasad M (2015) Effect of mineralogy on nuclear magnetic
resonance surface relaxivity: a case study of Middle Bakken and
Three Forks formations. Fuel 161:197–206
Silletta EV, Franzoni MB, Monti GA, Acosta RH (2018) Influence of
exchange in NMR pore size/relaxation correlation experiments.
Microporous Mesoporous Mater 269:17–20
Solone R, Bittelli M, Tomei F, Morari F (2012) Errors in water retention
curves determined with pressure plates: effects on the soil water
balance. J Hydrol 470–471:65–74
Steinberg S, Kluitenberg GJ, Jones SB et al (2005) Physical and hydraulic
properties of baked ceramic aggregates used for plant growth medi-
um. J Am Soc Hortic Sci 130:767–774
Stoltz G, Tinet AJ, Staub MJ, Oxarango L, Gourc JP (2012) Moisture
retention properties of municipal solid waste in relation to compres-
sion. J Geotech Geoenviron 138:535–543
Tian H, Wei C, Wei H, Yan R, Chen P (2014) An NMR-based analysis of
soil-water characteristics. Appl Magn Reson 45:49–61
Zhang L, Chen Q (2005) Predicting bimodal soil water characteristic
curves. J Geotech Geoenviron 131:666–670
Zhang N, Zhao FF, Guo PY et al (2018) Nanoscale pore structure char-
acterization and permeability of mudrocks and fine-grained sand-
stones in coal reservoirs by scanning electron microscopy, mercury
intrusion porosimetry, and low-field nuclear magnetic resonance.
Geofluids 2018:1–20