Journal of Alloys and Compounds 648 (2015) 205e214

Contents lists available at ScienceDirect

Journal of Alloys and Compounds

journal homepage: http://www.elsevier.com/locate/jalcom

Magnetic properties of the magnetic hybrid membranes based on

various polymer matrices and inorganic fillers
Aleksandra Rybak a, *, Waldemar Kaszuwara b
a
Department of Physical Chemistry and Technology of Polymers, Faculty of Chemistry, Silesian University of Technology, Strzody 9, 44-100 Gliwice, Poland
b
Faculty of Materials Science and Engineering, Warsaw University of Technology, Woloska 141, 02-507 Warszawa, Poland

a r t i c l e i n f o a b s t r a c t

Article history: Magnetic hybrid membranes based on ethylcellulose (EC), poly(2,6-dimethyl-1,4-phenylene oxide) (PPO)
Received 20 March 2015 and various magnetic praseodymium and neodymium powder microparticles as fillers were obtained.
Received in revised form Permeability, diffusion and sorption coefficients of O2, N2 and synthetic air components were estimated
17 June 2015
for homogeneous and heterogeneous membranes using the Time Lag method based on constant pressure
Accepted 23 June 2015
Available online 30 June 2015
permeation technique. The microstructure studies and the phase analysis of magnetic membranes were
also performed using SEM and XRD. The influence of magnetic parameters, like coercivity, remanence
and saturation magnetization of created membranes on the gas transport properties was studied. The
Keywords:
Composite materials
results showed that their coercivity depended on composition and microstructure of the magnetic
Magnetic films powder. On the other hand, remanence and saturation magnetization increased with the increase of the
Diffusion powder addition in the membrane. It was found that the magnetic membrane's gas transport properties
Magnetic measurements were improved with the increase of membrane's remanence, saturation magnetization and magnetic
Magnetic hybrid membranes particle filling. The decrease in powder particle size and associated increase of the membrane's coercivity
Gas separation also positively influenced the gas transport and separation properties of investigated membranes. It was
observed that the magnetic ethylcellulose and poly(2,6-dimethyl-1,4-phenylene oxide) membranes had
higher gas permeability, while their permselectivity and solubility coefficient values were rather
maintained or slightly increased. The results also showed that the magnetic powder content enhanced
significantly gas diffusivity in EC and PPO membranes. It was also analyzed the dependence of the drift
coefficient w on the magnetic parameters of investigated membranes. The correlation between the
membrane selectivity, permeability and magnetic properties with their XRD characteristics was stated.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction controlling the composition, content and morphology of the par-

Polymer composites are defined as polymers in which some
amounts of micro- or nanometer size fillers are homogeneously
dispersed. The main idea of such composites is to integrate several
component materials and their properties in a single material.
These composites represent the very attractive class of materials,
which have become a center of many research projects throughout
the world. Their physical and chemical properties show sensitivity
to changes and have the ability to adapt to the environmental pa-
rameters like temperature, pressure, magnetic field, optical wave-
length, etc. Composites with polymer matrix are materials of great
interest for researchers because their properties can be adjusted by
* Corresponding author.
E-mail addresses: Aleksandra.Rybak@polsl.pl, alryx@wp.pl (A. Rybak).
ticle addition, different processing techniques or by modification of
the polymer matrix. Material and type of filler (particles, short or
elongated fibers) as well as of their amount in the composite have a
significant influence on the physical and mechanical properties
(thermal stability, optical and thermal properties as well as me-
chanical properties) [1,2].
Over the last decade, many researchers focused on a new type of
hybrid material based on polymers-matrix composites containing
dielectric or/and magnetic fillers. There are polymers with
extremely good thermal and chemical stability, an easy tenability of
their electronic properties and high levels of electromagnetic
shielding performances at microwave frequencies with a low mass
by unit of surface. The electromagnetic properties of polymers can
be modified by the addition of inorganic fillers. The inclusion of
magnetic particles may improve the magnetic and dielectric
properties of host materials. Therefore, polymers combined with

http://dx.doi.org/10.1016/j.jallcom.2015.06.197
0925-8388/© 2015 Elsevier B.V. All rights reserved.
206 A. Rybak, W. Kaszuwara / Journal of Alloys and Compounds 648 (2015) 205e214

magnetic particles provides materials exhibiting novel functional- During the gas permeation experiments was measured flow rate
ities [3]. Q, which in standard conditions is defined as [20].
Combination of polymers with other organic or inorganic ma-
terials provides new materials with tailored properties, suitable for TSTP p
QSTP ¼ Q (1)
various electrical and electromagnetic applications in aeronautic or TpSTP
aerospace areas, various sensors, microelectronics, photocatalysis,
magnetic devices, batteries, electrochemical display devices, where:
electrical-magnetic shields and microwave absorption materials,
coatings, biomedical, and powder metallurgy [2,4]. Q e measured flow rate [cm3/s],
Polymer composites have the benefit of easy process and p e atmospheric pressure [Pa],
preparation, several species of polymer with low price make the T e temperature [K],
composite materials to be used extensively [5]. There have been pSTP e standard pressure, pSTP ¼ 1.013$105 Pa,
used many polymer composites with enhanced electrical and TSTP e standard temperature, TSTP ¼ 293.15 K.
magnetic properties, like, for instance: non-conducting and con-
ducting polymers PANI with Fe3O4, nano-colloidal iron oxides, g- To avoid the necessity of taking into account the impact of
Fe2O3 and Fe3O4, NiZn ferrite, BaFe12O19, hollow Fe3O4ePANI possible changes of temperature and pressure on the measurement
spheres, PANI nanotubes containing Fe3O4 nanoparticles, the SiO2/ results, they were converted to standard conditions.
NdeFeeB magnet, ferrospinels of the type MFe2O4 (M is a divalent This flow rate QSTP was then recalculated into diffusive mass flux
metal cation), Fe3O4/rubber, MnFe2O4/polyaniline, NiFe2O4/poly- in a stationary state JS:
styrene and hexaferrite/polymer [3,4,6e10]. Authors focused not
only on ferrites, but they extended their research to other materials QSTP
such as titanium dioxide, zinc oxide, manganese oxide, vanadium Js ¼ (2)
A
pentaoxide, clinoptilolite, carbon nanotubes and silicon PANI/
Co1
xMgxFe2O4 nano-composites (x ¼ 0, 0.5, 1) [11]. It could be also where:
" #
used magnetic coreeshell nanocomposites using Zn0.5Co0.5Al0.5-
cm3STP
Fe1.46La0.04O4 as a core being surfactant with four different poly- QSTP e flow rate at standard condition s ,
meric shells (PVP, PVA, PVAc and PEG) [12]. A e membrane active area [cm2].
Because nowadays the interest in polymer composites is still
growing, they have found also the application in membrane sepa- Using this JS, permeation coefficient P was evaluated according
ration techniques. They have been implemented as so called mixed to the given equation [20,22]:
matrix membranes (MMMs). This is a new class of membrane
materials, that offers the significant potential in membrane sepa- JS l
P¼ (3)
ration technology. The successful application of MMMs depends Dp
mainly on the polymer matrix selection, the inorganic filler as well
as the interaction between these two phases. These membranes where:
combine the selectivity of fillers and the simplicity of polymer
cm3 $cm
membrane processing [13,14]. It was observed that dispersion of P e permeation coefficient [Barrer], Barrer ¼ cm2 $s$cmHg
STP
$10
10 ,
impermeable particles in polymers could improve the gas separa- L e membrane thickness [cm],
tion properties of membranes [15e19]. Dp e gas pressure difference at both sides of the membrane
Over the last few years, our research had concentrated on [cmHg], " #
cm3
magnetic hybrid membranes used for the air enrichment in oxygen. Js e diffusive mass flux in a stationary state cm2STP
$s
.
We based on the difference between magnetic properties of para-
magnetic oxygen and diamagnetic nitrogen. That has provided the An average diffusion coefficient D was calculated from the sta-
differentiated conditions (other than sorptionediffusion) in mass tionary state of permeation according to the formula:
transport of the separated gases and gives the opportunity for their
separation. Js $l
D¼ (4)
This paper is the continuation of our earlier work [20e22], Dc
where we have found that incorporation of magnetic micro-
powders (with granulation 20e32 mm and 32e50 mm) into the where
polymer matrix (EC, PPO) improved the gas transport properties of
membranes. In this work, we are going to exam the magnetic pa- Js e diffusive mass flux in stationary state (cm3STP/cm2.s),
rameters (coercivity, remanence and saturation magnetization) of l e thickness of membrane (cm),
various types of hybrid membranes (EC and PPO with dispersed Dc-concentration difference.
magnetic powders, like commercial MQP-14-12, MQP-16-7 and
MQP-B with various particle sizes). After integration of JS with respect to time, we have got a
downstream absorption permeation Qa (l, t) (total flow of pene-
2. Theoretical trant). Concentration difference Dc and Time Lag La were obtained
using the Time Lag method. The first one was obtained from an
2.1. The evaluation of gas transport coefficients intercept of the asymptote to the stationary permeation curve with
the Qa (l, t) axis. While the second one La was determined as an
We have collected experimental permeation data for the oxygen intercept of the asymptote to the stationary permeation curve with
and nitrogen, both for individual, pure gases, as well as components the time axis [20e22].
of air. We have analyzed the system, i.e. membrane and permeating Comparing D with the value of diffusion coefficient DL calculated
gas using appropriate description [20e22], supported by the per- from the downstream absorption Time Lag La [20e22] according to
centage of the air enrichment. the following equation:
 A. Rybak, W. Kaszuwara / Journal of Alloys and Compounds 648 (2015) 205e214 207

l2
DL ¼ (5)
6La ðlÞ

we can get some insight into the nature of the transport process.
Namely, if D ¼ DL , the diffusion is supposed to be an “ideal Fickian”
unless there are some hidden processes, like a drift, which are
beyond the reach of this simple protocol [20,22]. The drift coeffi-
cient measures the strength of a potential field that causes a
preferred direction of motion of the inspected molecules with
respect to the field. In the case of the magnetic membranes, the
presence of a drift is connected with the participation of magnetic
field in the orientation of gas particles (especially paramagnetic
oxygen) movement through the membrane, according to the field
lines.
Assuming that diffusional contribution to the overall flux is
Fig. 2. Demagnetization curves (II quarter of hysteresis loop) of magnetic hybrid
magnetic field independent, we can separate the drift and diffusion membranes with different polymer matrices.
in an overall flux.

(6)

JHi ¼ J0 þ wHi Dc0 (7)

where Hi indicates the magnetic field intensity of a membrane. The

drift coefficient wHi can be calculated for an imposed field intensity
Hi as:

JHi
J0
wHi ¼ (8)
Dc0

and its values for different H are shown in Fig. 8.
Next parameter, the solubility coefficient S is a measure of the
size of the sorption in membranes. It is characterized by division of
the penetrant between the membrane and the outer phase in
equilibrium. This parameter was calculated from the relation:
Fig. 3. Demagnetization curves (II quarter of hysteresis loop) of magnetic hybrid
membranes with different amount of magnetic filler.
Another parameter, like ideal selectivity coefficient aO2=N2 could
be calculated from the ratio of obtained permeation coefficients:

P PO2
S¼ (9) aO2=N2 ¼ (10)
D PN2

where: where:
" #
cm3STP
S e solubility coefficient cm3 cmHg
PO2, PN2 e permeation coefficient of pure components.

Fig. 1. Hysteresis loops of magnetic hybrid membranes: a) with two different particle sizes, b) demagnetization curves e II quarter of hysteresis loop for four different particle sizes
of MQP-14-12 magnetic powder.
208 A. Rybak, W. Kaszuwara / Journal of Alloys and Compounds 648 (2015) 205e214
Fig. 4. Demagnetization curves (II quarter of hysteresis loop) of magnetic hybrid
membranes with different magnetic powders.
Fig. 5. Dependence of diffusion and permeation coefficients on EC magnetic mem-
brane's coercivity, changing with the powder granulation.
3. Experimental
3.1. Materials
Ethylcellulose (EC) was supplied by Acros Organics. Poly(2,6-
dimethyl-1,4-phenylene oxide) (PPO) was supplied by Sabic
Fig. 6. Dependence of diffusion and permeation coefficients on magnetic membranes remanence for membranes with a) different powder granulation in EC matrix and b) various
magnetic powder content in PPO matrix.
Innovative Plastics. The neodymium and praseodymium magnetic
powders MQP-14-12, MQP-16-7 and MQP-B were used as an inor-
ganic filler. The magnetic powder MQFP-14-12 with 5 and 25 mm
granulation (d50 ¼ 5 mm and d50 ¼ 25 mm), MQFP-16-7 and MQFP-B
with 5 mm granulation (with AA4 coating; d50 ¼ 5 mm) are the
products of Magnequench company. The particle size distribution
data come from a producer. While the MQP-14-12 with a particle
size below 20 mm and above 20 mm were obtained during the
milling of powder MQP-14-12 supplied by Magnequench. They are
the isotropic magnetic powders, based on a patented Nd-Fe-B and
Nd-Pr-Fe-B alloy compositions. These materials are produced by
employing a proprietary rapid solidification process followed by a
milling process and heat treatment.
Ethanol, toluene and trichloroethylene (TCE) with 99% purities
were supplied by Sigma Aldrich (Poland). Oxygen (5.6), nitrogen
(5.0), helium (6.0) and synthetic air (5.0) compressed cylinders
were supplied by Linde Gas (Poland).
3.2. Membrane preparation
In our research, we have examined the homogeneous ethyl-
cellulose (EC) and poly(2,6-dimethyl-1,4-phenylene oxide) (PPO)
and heterogeneous magnetic membranes of thickness 80e205 mm
(depended on the amount of added magnetic powder). Homoge-
neous EC membranes were cast from 3 wt% EC solution in 40:60
ethanol/toluene, while PPO membranes from 2.5 wt% PPO solution
in trichloroethylene (TCE). Magnetic membranes were cast in the
external field of a coil (stable magnetic field intensity 32 kA/m) in
two stages. Firstly, from a mixture of appropriate polymer (EC or
PPO) solutions and the dispersed magnetic powders after sonicat-
ion (of the appropriate amount: 0.8e1.75 g and various granula-
tions) was poured the first layer. After that the second layer was
poured from the polymer solution. Membranes were removed from
Petri dish (with some distilled water) and dried in 40  C for at least
two days before any analysis. With this procedure, a series of
membranes with an inorganic content ranging from 66.0 to 82.0 wt
% was prepared. Application of the external magnetic field during
the membrane casting, enhanced the proper particle arrangement.
This has enabled us to obtain membranes with perfectly dispersed
filler particles in the polymer matrix. The permeation measure-
ments were performed, before and after magnetization by a short
magnetic impulse using an impulse inducteur previously charged
to an appropriate voltage, generating magnetic field intensity
approximately of 2280 kA/m. Measurements of membrane mag-
netic field intensity were carried out using Magnetic Field Strength
Meter Gauss-/Teslameter FH 54. The membranes were stored in a
desiccator under the vacuum conditions (p ¼ 400 pa).
 A. Rybak, W. Kaszuwara / Journal of Alloys and Compounds 648 (2015) 205e214 209

Fig. 7. Dependence of gas transport coefficients: a) diffusion and permeation and b) ideal selectivity aO2=N2 on PPO magnetic membrane's saturation magnetization and remanence.

Fig. 8. Dependence of drift coefficient on PPO magnetic membrane's a) remanence and b) saturation magnetization.

3.3. Experimental setup
Low-pressure gas permeation analyzer APG-1, described in the
papers [21] was used for measurement of the nitrogen, oxygen and
synthetic air permeability. This setup was furnished with a gas
chromatograph HP 5890A, which let us to measure the oxygen and
nitrogen concentration in permeate. Its main part is a diffusive
chamber, where analyzed membrane was put in the form of disc
(membrane effective area 19.63 cm2). Oxygen, nitrogen and syn-
thetic air (21% O2 and 79% N2) were supplied to the setup from
compressed gas cylinders. After installation of membrane in the
diffusive chamber, we have rinsed this chamber with an analyzed
gas. Then, between high- and low-pressure part of a chamber the
suitable pressure difference was established (0.3e0.9 MPa) and
controlled by an electronic pressure meter and controller EL-press
Pe506C. After that, the flow rate of permeate was obtained using
Flow-Bus flowmeter (with a range 0e3 ml/min) with a computer
data acquisition. All measurements were carried out in room
temperature. Transport coefficients were calculated, using these
flow-rate data and percentage of the air enrichment.
3.4. Characterization of magnetic membranes
The magnetic properties of membranes were examined in a
Lake Shore 7010 vibrating sample magnetometer (VSM) in which
the sample vibrates perpendicularly to the applied magnetic field.
The coercivity, remanence, and the saturation magnetization were
determined from the recorded magnetic hysteresis loops. The
maximum intensity of the magnetic field in which the measure-
ments were made was 2280 kA/m. As a result of the changes of the
magnetic flux due to the perpendicular vibration of the sample an
electromotive force, proportional to the magnetization, was
induced in the detection coils. Thanks to the special system with
which the device is equipped, the measured magnetic moment of
the sample does not depend on the frequency and amplitude of its
vibrations. The intensity of the external magnetic field was
measured with a Hall probe. In general, the measured values of
magnetization depend on the shape of the sample. The samples
were cut off from the membranes and placed in the magnetometer
perpendicularly to the direction of the magnetic field so that their
thickness was negligibly small compared to their other dimensions.
The microstructure studies were performed using a scanning
electron microscope HITACHI 3500 N.
The X-ray diffractometry (XRD) was used to perform the phase
analysis of the applied powders and obtained magnetic mem-
branes. The XRD scans were measured using the Rigaku MiniFlex II
diffractometer with Cu Ka radiation. Samples were scanned from
20 to 60 at scan rate of 0.3 /min.
The particle size of magnetic powders was determined using the
HORIBA Laser Scattering Particle Size Distribution Analyzer LA-950.
210 A. Rybak, W. Kaszuwara / Journal of Alloys and Compounds 648 (2015) 205e214
The results of this analysis are given in Table 1.
The median particle size value differs, to a certain extent, from
the values specified by the manufacturer, remaining within the
limit of the measurement error. At the same time, we observed the
presence of a small fraction of agglomerates up to the 200 mm,
especially in the case of powders with higher particle size distri-
bution (for example, powders obtained during the milling of MQP-
14-12).
4. Results and discussion
4.1. Permeation data
The main permeation data for magnetic membranes with
magnetic praseodymium and neodymium powders MQP-14-12,
MQP-16-7 and MQP-B with a particle size dm ¼ 7, 8, 20 mm,
18 mm and 21 mm MQP-14-12 were collected in Tables 2 and 3.
Presented data, are the mean values of six measurements with
standard deviations not greater than 15%.
4.2. Magnetic properties
Figs. 1e4 show the hysteresis loops and demagnetization curves
(II quarter of hysteresis loop for better presentation) of the mag-
netic hybrid membranes with various types, particle sizes, amounts
of added magnetic fillers and two different polymer matrices.
The magnetization under applied magnetic field for magnetic
hybrid membranes exhibits a clearly hysteretic behavior. Shape of
the loops evidences the hard magnetic character of the material.
Table 4 gives the magnetic parameters of obtained MMM mem-
branes, including coercivity, remanence and saturation magnetiza-
tion determined from the hysteresis loops measurements.
The results (Figs. 1e4, Tables 1 and 4) and earlier studies [8,9]
showed that the coercivity depended on a composition and
microstructure of the magnetic powder but not on the participation
of the powder (the differences are within the range of measure-
ment error). Remanence depends to a small degree on the micro-
structure. Remanence and saturation magnetization increased with
the increase of the powder addition in the membrane. Saturation
magnetization is linearly dependent on the magnetic filler additive.
These properties depend on the addition of the magnetic powder in
the membrane.
The coercivity, remanence and saturation magnetization results,
especially for membranes with a different particle size distribution
may be influenced by the fact that the powders of grain size
dm ¼ 7e11 and 22 mm are coated with an inorganic coating AA4.
Coercivity measured for magnetic membranes is within the
range specified by the manufacturer for the coercivity of the
applied magnetic powders. However, coercivity of the membrane
with powder granulation dm ¼ 36 mm is much smaller. It could be
caused by powder preparation process or by the fact that coercivity
is depended on surface anisotropy and interparticle interactions. So
the coating of magnetic powder by polymer will affect on the
Table 1
Results of particle size distribution analysis of magnetic powders.
Powder Rangea or medianb particle size [mm]
MQFP-14-12, 5 mm 5b
MQP-14-12, below 20 mm <20a
MQFP-14-12, 25 mm 25b
MQP-14-12, above 20 mm >20a
MQFP-16-7, 5 mm 5b
MQFP-B, 5 mm 5b
a
Sieves used by authors.
b
Given by producer.
participation of this mechanism to the net anisotropy. The reduc-
tion of effective magnetocrystalline anisotropy by the polymer
coating will decrease the coercivity.
It was found that the saturation magnetization Ms values of
magnetic membranes were lower than the bulk magnetic powder
particles. For instance for powders MQP-14-12 (granulation
dm ¼ 20 mm and 36 mm) the value Ms was 105.66 emu/g and 108.54
emu/g respectively, when for magnetic membranes with the
addition of these powders were 63.89 emu/g and 49.87 emu/g. A
similar dependency has been ascertained by D.L. Huber, J. Lee and
I.V. Beketov et al. [23e25] who has stated that the chemical envi-
ronment (polymers, surfactants) has an influence on the magneti-
zation of tested particles. However, they have indicated that the
real systems have a number of variables, like: polydispersity, shape
anisotropy, surface morphology, and bonding chemistry that are
not entirely determined.
Magnetic properties of the magnetic MMMs membranes
depend mainly on the type, the particle size and addition of the
magnetic powder, and to a minor degree on the type of polymer
matrix. That means that the total magnetic behavior of these
magnetic hybrid membranes could be tuned and tailored
depending on these parameters. And this will have a direct impact
on the separation and transport properties of investigated mem-
branes and will affect their potential use in the future.
Coercivity (Fig. 5) depends on the type of powder and its particle
size. The obtained results show the influence of the magnetic
particle size on the coercivity (Tables 1 and 4). Powders with
smaller granulation and particle size distribution achieve higher
coercivity. It can be associated with a lower probability of large
surface defects on smaller grains. Such defects and local demag-
netizing fields may be responsible for the reduction of the nucle-
ation field, that controls the coercivity value [8,9].
With decreasing particle size of the magnetic powder, coercivity
increases and also grow values of the diffusion, permeability and
separation coefficients, D, P and, a, while the sorption coefficient S
varies to a small extent. This trend is observed for pure oxygen and
nitrogen and mixtures thereof. Whereas the strongest tendency
was observed for the pure oxygen and oxygen in the air, and the
smallest for the pure nitrogen.
It should be noted that the observed effect of coercivity on the
transport and separation properties of the membrane is only
apparent. Because the coercivity depends on the remanence, and
this parameter largely affects the properties of the concerned
membranes, and, hence its influence is indirect.
Remanence (Fig. 6) depends on the magnetic powder content
and in a small degree on the microstructure. With decreasing
particle size of the magnetic powder (Fig. 6 a) its value increases,
and for gas transport coefficients (D, P, S and a) is observed a similar
behavior, as in the case of the coercivity.
Remanence also increases with the rise of magnetic powder
content (Fig. 6 b), while the values of transport coefficients grow
exponentially with it. However, the greatest increase is observed
for the paramagnetic oxygen.
dm mean particle size [mm] Standard deviation [mm] d10 d50 d90 Agglomerate [mm]
10.91 11.96 4.90 7.51 12.00 To 200
20.21 11.45 7.80 18.38 37.16 To 70
21.91 10.71 9.53 20.09 37.01 No
36.22 35.72 10.85 20.56 95.99 To 110
9.44 9.91 5.32 8.22 12.88 100e200
7.00 0.87 5.94 6.96 8.26 No
 A. Rybak, W. Kaszuwara / Journal of Alloys and Compounds 648 (2015) 205e214 211

Table 2
Mass transport coefficients of pure oxygen and nitrogen for various magnetic membranes.

Membrane aΟ2/N2 N2 pure O2 pure

D DL P S D DL P S
7 2 3 7 2
10 (cm /s) (Barrer) 10 ðcm3STP =cm3 cm HgÞ 10 (cm /s) (Barrer) 103 ðcm3STP =cm3 cm HgÞ

EC 1.45 3.70 3.70 0.80 0.22 11.00 10.00 1.16 0.11

EC, 1.5g, MQFP-14-12, 2.52 18.65 13.83 45.68 2.42 53.10 71.56 112.36 2.14
dm ¼ 11 mm
EC, 1.5g, MQP-14-12, 2.43 17.97 14.10 43.91 2.42 50.15 67.64 105.82 2.11
dm ¼ 20 mm
EC, 1.5g, MQFP-14-12, 2.44 17.52 13.22 42.82 2.42 48.31 64.91 101.80 2.11
dm ¼ 22 mm
EC, 1.5g, MQP-14-12, 1.99 14.51 11.81 34.93 2.40 35.20 47.65 72.69 2.07
dm ¼ 36 mm
EC, 1.5g, MQFP-16, 2.69 19.86 14.41 48.82 2.43 58.34 78.52 123.96 2.17
d m ¼ 9 mm
PPO 3.80 0.80 0.70 3.80 4.75 1.70 1.80 14.40 8.47
PPO, 1.0g, MQFP-16, 4.41 2.97 2.92 16.18 5.45 71.41 6.21 8.14 11.51
d m ¼ 9 mm
PPO, 1.5g, MQFP-16, 4.79 4.16 4.01 29.25 7.03 140.16 11.76 15.48 11.92
d m ¼ 9 mm
PPO, 1.75g, MQFP-16, 4.90 4.94 4.72 37.84 7.66 185.30 15.41 20.30 12.03
d50 ¼ 9 mm
PPO, 0.8g, MQFP-B, 4.12 2.59 2.57 11.99 4.63 49.39 4.43 5.79 11.16
d m ¼ 7 mm
PPO, 1.1g, MQFP-B, 4.50 3.14 3.08 18.07 5.75 81.36 7.01 9.20 11.61
d m ¼ 7 mm
PPO, 1.3g, MQFP-B, 4.67 3.60 3.49 23.07 6.41 107.69 9.14 12.01 11.79
d m ¼ 7 mm
PPO, 1.5g, MQFP-B, 4.77 4.03 3.89 27.81 6.90 132.57 11.15 14.67 11.89
d m ¼ 7 mm

Table 3
Mass transport coefficients of oxygen and nitrogen in air for various magnetic membranes.

Membrane aΟ2/N2 N2 in air O2 in air

D DL P S D DL P S
7 2 3 7 2
10 (cm /s) (Barrer) 10 ðcm3STP =cm3 cm HgÞ 10 (cm /s) (Barrer) 103 ðcm3STP =cm3 cm HgÞ

EC 1.53 9.00 9.00 15.00 2.00 15.00 15.00 23.00 1.52

EC, 1.5 g, MQFP-14-12, 2.74 83.72 94.04 154.14 1.89 159.84 184.28 401.99 2.67
dm ¼ 11 mm
EC, 1.5 g, MQP-14-12, 2.63 79.30 89.15 145.84 1.88 150.00 173.07 375.71 2.50
dm ¼ 20 mm
EC, 1.5 g, MQFP-14-12, 2.56 76.13 81.38 140.07 1.87 143.50 162.17 358.80 2.50
dm ¼ 22 mm
EC, 1.5 g, MQP-14-12, 2.34 56.73 61.58 103.52 1.80 99.97 113.83 242.23 2.33
dm ¼ 36 mm
EC, 1.5 g, MQFP-16, 2.94 91.57 102.73 168.89 1.92 177.30 204.18 448.63 2.76
d m ¼ 9 mm
PPO 3.04 0.80 0.80 4.82 6.02 1.80 1.90 14.65 8.14
PPO, 1.0 g, MQFP-16, 3.85 5.07 4.08 27.40 5.40 12.51 9.80 105.43 8.43
d m ¼ 9 mm
PPO, 1.5 g, MQFP-16, 4.25 8.53 6.85 49.93 5.85 23.75 19.44 212.00 8.93
d m ¼ 9 mm
PPO, 1.75 g, MQFP-16, 4.36 10.81 8.66 64.73 5.99 31.13 25.77 281.96 9.06
d50 ¼ 9 mm
PPO, 0.8 g, MQFP-B, 3.53 3.97 3.19 20.18 5.09 8.91 6.71 71.31 8.00
d m ¼ 7 mm
PPO, 1.1 g, MQFP-B, 3.94 5.57 4.48 30.66 5.50 14.14 11.19 120.86 8.55
d m ¼ 7 mm
PPO, 1.3 g, MQFP-B, 4.11 6.90 5.54 39.29 5.69 18.44 14.88 161.66 8.77
d m ¼ 7 mm
PPO, 1.5 g, MQFP-B, 4.22 8.15 6.54 47.45 5.82 22.51 18.37 200.24 8.90
d m ¼ 7 mm

As the amount of magnetic powder increases, so does the
saturation magnetization MS (Fig. 7). It was also observed linear
increase of gas transport coefficient's values with MS value.
Wherein the influence of remanence on the separation (coefficient
a), and transport properties of membranes (coefficient's P, D, S) is
much greater than this of saturation magnetization. The small
increase of a nitrogen diffusion coefficient with the rise of the
saturation magnetization, could be associated with the fact that the
saturation magnetization increases with the addition of magnetic
powder. This in turn causes an increase in the interphase bound-
aries' polymer-powder, which are probably the easy diffusion
paths. So, the diffusion coefficient of nitrogen can grow slightly due
212 A. Rybak, W. Kaszuwara / Journal of Alloys and Compounds 648 (2015) 205e214

Table 4
Magnetic properties of MMMs membranes.

Membrane Coercivity [kA/m] Remanence [emu/g] Saturation magnetization [emu/g]

EC, 1.5 g, MQFP-14-12, dm ¼ 11 mm 960 45.25 63.21

EC, 1.5 g, MQP-14-12, dm ¼ 20 mm 947 45.59 63.89
EC, 1.5 g, MQFP-14-12, dm ¼ 22 mm 950 35.43 50.07
EC, 1.5 g, MQP-14-12, dm ¼ 36 mm 911 34.84 49.87
EC, 1.5 g, MQFP-16, dm ¼ 9 mm 551 43.49 75.83
PPO, 1.0 g, MQFP-16, dm ¼ 9 mm 555 29.09 51.45
PPO, 1.5 g, MQFP-16, dm ¼ 9 mm 551 42.86 75.31
PPO, 1.75 g, MQFP-16, dm ¼ 9 mm 568 44.83 75.54
PPO, 0.8 g, MQFP-B, dm ¼ 7 mm 747 45.67 67.75
PPO, 1.1 g, MQFP-B, dm ¼ 7 mm 744 47.69 70.46
PPO, 1.3 g, MQFP-B, dm ¼ 7 mm 737 48.03 71.53
PPO, 1.5 g, MQFP-B, dm ¼ 7 mm 730 48.49 74.43

to the changes in the microstructure.
Analyzing the dependence of the drift coefficient's values w on
the magnetic properties of various investigated membranes (Fig. 8),
it was found that the increase of the remanence and saturation
magnetization causes an increase of a drift coefficient w, particu-
larly in the case of a paramagnetic oxygen. While the diamagnetic
nitrogen remains almost constant. For drift coefficient w of nitrogen
in the air, was found small changes, which may be caused by the
interaction of the mixture components (synthetic air) during their
permeation through the magnetic membrane.
On the basis of SEM images can be identified the presence of the
polymer upper layer (left side of images) and below the phase of
inorganic additive in the polymer matrix. It was concluded that the
powder particles with a greater particle size distribution (Fig. 9a)
form a heterogeneous structure, which consists of large particles
and the agglomerates of the smaller ones. In contrast to the
membranes with a smaller particle size and its distribution
(Fig. 9b), wherein the powder particles form a homogeneous
structure in the polymer matrix and an extensive system of the
magnetic channels (with appropriate magnetic properties, like
remanence and coercivity). This, in turn, has a direct influence on
the gas separation properties of obtained membranes (an increase
of selectivity coefficient aO2=N2 ). The idea of magnetic channels,
created by the magnetic powder particles in the polymer matrix
under the influence of the magnetic field was introduced in the
earlier papers. Attention was also paid to the role of these channels
in the separation process of oxygen and nitrogen (various magnetic
properties) which interacting differently with those magnetic
channels [20e22].
In Fig. 10 are shown XRD spectra of both magnetic membranes
(black lines) as well as of micropowders MQP-14-12, MQFP-16-7
(grey lines), which are the inorganic additives.
Comparing the XRD spectra of the membranes and magnetic
powders can be noted that the powder characteristic peaks overlap
with peaks from the magnetic membranes. In addition, it was also
found the presence of peaks characteristic for the polymer matrix
EC (20 ) and PPO (21,7 ) [26e28]. Ethylcellulose showed diffused
peaks indicating the amorphous nature of this polymer [26]. In
turn, it can be observed that the peak intensity is higher for the PPO
magnetic membrane. This indicates the higher density of crystal-
lites in this semi-crystalline polymer membrane [28]. The selec-
tivity and the permeability data for O2 and N2, obtained for these
membranes were correlated with the obtained membrane char-
acteristics from X-ray diffraction analysis. The permeability of both
O2 and N2 was higher for EC membranes in comparison with the
PPO membranes. In general the crystalline phase may be regarded
as impermeable, so that gas permeability in a semi-crystalline
polymer membrane, like PPO is substantially lower than in the
more amorphous membrane, like EC because of the reduced space
available for diffusion and the winding path around the crystallites.
Except that, it should be also taken into account the impact of the
added magnetic powders and their particle size.
XRD analysis shown also the presence of the crystalline phases,
like: the hard magnetic Pr2Fe14B (characteristic peaks 32,93 , 37,16
and 42,27 ), PrNdFe14B (characteristic peaks 37,18 and 42,23 ) and
Nd2Fe14B phases (characteristic peaks 37,32 , 42,34 and 43,02 )
and soft magnetic Nb2FeB2 (characteristic peaks 41,17 and 44,06 ),
Fe3B (characteristic peaks 43,12 and 48,20 ), Fe2B phases (char-
acteristic peaks 42,52 ). For some peaks was observed the phe-
nomenon of overlapping with peaks coming from the iron borides,
especially in the case of addition of powder MQFP-16-7. The largest
peaks originated from phases of RE2Fe14B type. Powder MQP-14-12
contains Nd as a metal RE, while the powder MQP-16-7 contains Pr
and Nd. Nd2Fe14B phases and the (Nd, Pr)2Fe14B have similar

Fig. 9. SEM images of cross sections for two various magnetic membranes: a) EC with 1.5 g of MQP-14-12 (dm ¼ 20 mm) and b) PPO with 1.5 g MQFP-B (dm ¼ 7 mm).
 A. Rybak, W. Kaszuwara / Journal of Alloys and Compounds 648 (2015) 205e214
Fig. 10. XRD spectra for two magnetic powders and membranes with various types and granulation of added magnetic powders: a) EC with 1.5 g of MQP-14-12 (dm ¼ 20 mm) and b)
PPO with 1.75 g MQFP-16-7 (dm ¼ 9 mm).
diffraction spectra. The characteristic peaks of the crystallographic
planes were marked in the figures as follows: 1 e (202), 2 e (212) 3
e (222) 4 e (311), 5 e (204), 6 e (214), 7 e (105), 8 e (313), 9 e
(224), 10 e (410), 11 e (411), 12 e (314), 13 e (413), 14 e (511), 15 e
(433) and 16 e (513).
The magnetic properties obtained for these membranes could
be correlated with their characteristics from X-ray diffraction
analysis. It was stated that partial substitution of neodymium by
praseodymium results in an improvement of MQFP-16-7 magne-
tization (alloy Nd-Pr-Fe-B). This significant change was caused by
the greater magnetocrystalline anisotropy of Pr2Fe14B phase. In the
case of MQP-14-12 (alloy Nd-Nb-Fe-B) niobium addition results in
the formation of Nb2FeB precipitates, which in turn anchor the
domain walls and lead to a significant increase in coercivity [29].
5. Conclusions
Magnetic hybrid membranes were successfully synthesized on
the basis of ethylcellulose and poly(2,6-dimethyl-1,4-phenylene
oxide) polymer matrices and various magnetic powder micropar-
ticles. Based on the analysis of the magnetic properties of mem-
branes, it was found that their coercivity depended on a
composition and microstructure of the magnetic powder but not on
the participation of the powder. Remanence depended to a small
degree on the microstructure. Remanence and saturation magne-
tization increased with the increase of the powder addition in the
membrane. Saturation magnetization was linearly dependent on
the magnetic filler additive. The values of magnetic membrane
saturation magnetization (for MQP-14-12 (dm ¼ 20 mm) 63.89 emu/
g) were lower than that of bulk magnetic powder particles (105.66
emu/g) what was connected with the influence of chemical envi-
ronment on particle magnetic properties. It was found that the gas
transport properties of the magnetic membranes were improved
with the increase of magnetic powder content, accompanied by an
increase in membrane's remanence and saturation magnetization.
Wherein the influence of remanence on the separation (coefficient
a), and transport properties of membranes (coefficient's P, D, S) is
much greater than this of saturation magnetization.
The decrease in powder particle size, accompanied by an in-
crease of the membrane's coercivity and remanence also positively
influenced the gas transport and separation properties (higher gas
permeability, permselectivity and to a lesser extent solubility) of
investigated magnetic hybrid membranes. The increase of magnetic
powder content enhanced significantly gas diffusivity in EC and
PPO membranes. It was also found that the increase of the
213
remanence and saturation magnetization cause an increase of a
drift coefficient w, particularly in the case of a paramagnetic oxy-
gen, while the diamagnetic nitrogen remains almost constant. The
microstructure studies and the phase analysis (SEM and XRD) of
magnetic membranes confirmed that the application of powders
with appropriate composition, smaller particle size and its distri-
bution allows to obtain membranes with a more homogeneous
structure and an extensive system of the magnetic channels. This,
in turn, has a direct influence on the gas separation properties of
obtained membranes (an increase of selectivity coefficient aO2=N2 ).
It was also stated the correlation between the membrane selec-
tivity, permeability and magnetic properties with their XRD
characteristics.
It can be concluded, that the magnetic properties of the mag-
netic MMMs membranes depend mainly on the type, the particle
size and addition of the magnetic powder, and to a minor degree on
the type of polymer matrix. That means that the total magnetic
behavior of these magnetic hybrid membranes could be tuned and
tailored depending on these parameters. And this will have a direct
impact on the separation and transport properties of investigated
membranes and will affect their potential use in the future.
Acknowledgments
The authors would like to thank Professor Z. J. Grzywna for
substantive support.
References
[1] R.M. Dukali, I. Radovic, D.B. Stojanovic, P.S. Uskokovic, N. Romcevic,
V. Radojevic, R. Aleksic, Preparation, characterization and mechanical prop-
erties of Bi12SiO20ePMMA composite films, J. Alloy. Compd. 583 (2014)
376e381.
[2] K.G. Chandrappa, T.V. Venkatesha, Generation of Co3O4 microparticles by so-
lution combustion method and its ZneCo3O4 composite thin films for corro-
sion protection, J. Alloy. Compd. 542 (2012) 68e77.
[3] B. Belaabed, J.L. Wojkiewicz, S. Lamouri, N. El Kamchi, T. Lasri, Synthesis and
characterization of hybrid conducting composites based on polyaniline/
magnetite fillers with improved microwave absorption properties, J. Alloy.
Compd. 527 (2012) 137e144.
[4] R.M. Khafagy, Synthesis, characterization, magnetic and electrical properties
of the novel conductive and magnetic polyaniline/MgFe2O4 nanocomposite
having the coreeshell structure, J. Alloy. Compd. 509 (2011) 9849e9857.
[5] Z. Song, J. Xie, P. Zhou, X. Wang, T. Liu, L. Deng, Toughened polymer com-
posites with flake carbonyl iron powders and their electromagnetic/absorp-
tion properties, J. Alloy. Compd. 551 (2013) 677e681.
[6] L. Zheng, W. Li, M. Zhu, L. Ye, W. Bi, Microstructure, magnetic and electrical
properties of the composite magnets of NdeFeeB powders coated with silica
layer, J. Alloy. Compd. 560 (2013) 80e83.
[7] S.-Y. Tong, Effect of Ni fillers on microwave absorption and effective
214 A. Rybak, W. Kaszuwara / Journal of Alloys and Compounds 648 (2015) 205e214

permeability of NiCuZn ferrite/Ni/polymer functional composites, J. Alloy. M. Sipek, Mixed matrix membranes based on hyperbranched polyimide and
Compd. 550 (2013) 39e45. mesoporous silica for gas separation, Desalin. Water Treat. 34 (2011)
[8] K. Pawlik, P. Pawlik, J.J. Wysłocki, W. Kaszuwara, Microstructure and magnetic 211e215.
interactions in PreFeeCoeZreNbeB magnets, J. Alloy. Compd. 536S (2012) [19] H. Maachou, K. Bal, Y. Bal, A. Chagnes, G. Cote, D. Aliouche, In vitro biomin-
S348eS353. eralization and bulk characterization of chitosan/hydroxyapatite composite
[9] M. Bystrzejewski, O. Łabed z, W. Kaszuwara, A. Huczko, H. Lange, Controlling microparticles prepared by emulsification cross-linking method: orthopedic
the diameter and magnetic properties of carbon-encapsulated iron nano- use, Appl. Biochem. Biotechnol. 168 (2012) 1459e1475.
particles produced by carbon arc discharge, Powder Technol. 246 (2013) [20] A. Rybak, Z.J. Grzywna, P. Sysel, Mixed matrix membranes composed of
7e15. various polymer matrices and magnetic powder for air separation, Sep. Purif.
[10] I. Gancarz, J. Bryjak, M. Bryjak, W. Tylus, G. Po zniak, Poly(phenylene oxide) Technol. 118 (2013) 424e431.
films modified with allylamine plasma as a support for invertase immobili- [21] A. Rybak, G. Dudek, M. Krasowska, A. Strzelewicz, Z.J. Grzywna, P. Sysel,
zation, Eur. Polym. J. 4 (2006) 2430. Magnetic mixed matrix membranes in the air separation, Chem. Pap. 68 (10)
[11] A.A. Farghali, M. Moussa, M.H. Khedr, Synthesis and characterization of novel (2014) 1332e1340.
conductive and magnetic nano-composites, J. Alloy. Compd. 499 (2010) [22] A. Rybak, G. Dudek, M. Krasowska, A. Strzelewicz, Z.J. Grzywna, Magnetic
98e103. mixed matrix membranes consisting of PPO matrix and magnetic filler in gas
[12] M.A. Ahmed, R.M. Khafagy, S.T. Bishay, N.M. Saleh, Effective dye removal and separation, Sep. Sci. Technol. 49 (11) (2014) 1729e1735.
water purification using the electric and magnetic Zn0.5Co0.5Al0.5Fe1.46- [23] D.L. Huber, Synthesis, properties, and applications of iron nanoparticles, Small
La0.04O4/polymer coreeshell nanocomposites, J. Alloy. Compd. 578 (2013) 5 (2005) 482e501.
121e131. [24] J. Lee, T. Isobe, M. Senna, Preparation of ultrafine Fe3O4 particles by precipi-
[13] P.S. Goh, A.F. Ismail, S.M. Sanip, B.C. Ng, M. Aziz, Recent advances of inorganic tation in the presence of PVA at high pH, J. Colloid Interface Sci. 177 (1996)
fillers in mixed matrix membrane for gas separation, Sep. Purif. Technol. 81 490e494.
(2011) 243e264. [25] I.V. Beketov, A.P. Safronov, et al., Iron oxide nanoparticles fabricated by
[14] D.Q. Vu, W.J. Koros, S.J. Miller, Mixed matrix membranes using carbon mo- electric explosion of wire: focus on magnetic nanofluids, AIP Adv. 2 (022154)
lecular sieves: I. Preparation and experimental results, J. Membr. Sci. 211 (2012) 1e24.
(2003) 311e334. [26] J. Desai, K. Alexander, A. Riga, Characterization of polymeric dispersions of
[15] J. Hradil, P. Sysel, L. Bro , J. Kova
zova rova
, J. Kotek, Heterogeneous membranes dimenhydrinate in ethyl cellulose for controlled release, Int. J. Pharm. 308
based on a composite of a hypercrosslinked microparticle adsorbent and (2006) 115e123.
polyimide binder, React. Funct. Polym. 67 (2007) 432e441. [27] M. Khayet, J.P.G. Villaluenga, M.P. Godino, J.I. Mengual, B. Seoanea,
[16] E. Minko, P. Sysel, M. Hauf, J. Brus, L. Kobera, Synthesis and properties of K.C. Khulbe, T. Matsuura, Preparation and application of dense poly(-
hyperbranched polyimides combined with silica (conference paper), Macro- phenylene oxide) membranes in pervaporation, J. Colloid Interface Sci. 278
mol. Symp. 295 (2010) 88e93. (2004) 410e422.
[17] D. Sieffert, C. Staudt, Preparation of hybrid materials containing copolyimides [28] K.C. Khulbe, T. Matsuura, G. Lamarche, A.-M. Lamarche, X-ray diffraction
covalently linked with carbon nanotubes, Sep. Purif. Technol. 77 (2011) analysis of dense PPO membranes, J. Membr. Sci. 170 (2000) 81e89.
99e103. [29] M. Leonowicz, J.J. Wysłocki, Wspo  łczesne Magnesy. Technologie, Mechanizny
[18] P. Sysel, E. Minko, M. Hauf, K. Friess, V. Hynek, O. Vopicka, K. Pilnacek, Koercji, Zastosowania, WNT, Warszawa, 2005.