Low-Temperature Electrodeposition of Aluminium From Lewis Acidic 1-Allyl-3-Methylimidazolium Chloroaluminate Ionic Liquids
Low-Temperature Electrodeposition of Aluminium From Lewis Acidic 1-Allyl-3-Methylimidazolium Chloroaluminate Ionic Liquids
Low-Temperature Electrodeposition of Aluminium From Lewis Acidic 1-Allyl-3-Methylimidazolium Chloroaluminate Ionic Liquids
ZHENG Yong (郑勇)1,2, ZHANG Suojiang (张锁江)1,**, LÜ Xingmei (吕兴梅)1, WANG Qian
(王倩)1, ZUO Yong (左勇)1 and LIU Lian (刘恋)1,2
1
Beijing Key Laboratory of Ionic Liquids Clean Process, State Key Laboratory of Multiphase Complex Systems,
Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China
2
College of Chemistry and Chemical Engineering, Graduate University of Chinese Academy of Sciences, Beijing
100049, China
Abstract Lewis acidic 1-allyl-3-methylimidazolium chloroaluminate ionic liquids were used as promising elec-
trolytes in the low-temperature electrodeposition of aluminium. Systematic studies on deposition process have been
performed by cyclic voltammetry and chronoamperometry. The surface morphology and X-ray diffraction (XRD)
patterns of deposits prepared at different experimental conditions were also investigated. It was shown that the nu-
cleation density and growth rate of crystallites had a great effect on the structure of aluminium deposited. The crys-
tallographic orientation of deposits was mainly influenced by temperature and current density. Smooth, dense and
well adherent aluminium coatings were obtained on copper substrates at 10-25 mA·cm−2 and 313.2-353.2 K. More-
over, the current efficiency of deposition and purity of aluminium have been significantly improved, demonstrating
that the ionic liquids tested have a prospectful potential in electroplating and electrorefining of aluminium.
Keywords ionic liquid, low temperature, electroplating, electrodeposition, deposits
and responsible for the deposition. In recent years, an amber liquid at room temperature. Melting point:
many authors [29-31] reported that the structure of 290 K. Yield: 96.3%.
aluminium deposits was greatly influenced by the sur- 1-Butyl-3-methylimidazolium chloride ([Bmim]Cl)
face interaction in the interfacial solvation layer. The It was prepared according to Huddleston et al.’s work
cation type of IL had a vital effect on the interaction [36]. The final product was obtained as a white wax-like
strength of the interfacial layer. In recent years, some solid at room temperature. Melting point: 338 K. Yield:
bis(trifluoromethylsulfonyl)imide ([NTf2]−) ILs were 87.4%.
found to be capable of dissolving AlCl3 and nanocrys- Lewis acidic [Amim]Cl/AlCl3 and [Bmim]Cl/
talline aluminium could be electrodeposited from these AlCl3 ILs 1︰1.5 [Amim]Cl/AlCl3, 1︰1.8 [Amim]Cl/
solutions [32-34]. Nevertheless, these electrolytes are AlCl3, 1︰2.0 [Amim]Cl/AlCl3 and 1︰2.0 [Bmim]Cl/
inferior to the chloroaluminate ones in many key prop- AlCl3 were prepared by mixing precise molar quanti-
erties, such as conductivity, viscosity and liquid range ties of [Amim]Cl or [Bmim]Cl with anhydrous AlCl3
when they contain the same molar concentration of under dry argon atmosphere at room temperature. The
AlCl3. Moreover, the limited dissolving power of resulting liquids were filtered through G4 glass frit
AlCl3 and high costs of [NTf2]− ILs hinder their ad- and purified by pre-electrolysis before use. The final
vancement to an industrialization process. Consider- products were obtained as pale yellow liquids.
ing the properties and costs, chloroaluminate ILs have [Bmim][NTf2] (1-butyl-3-methylimidazolium
wide application potential and are worth of further bis(trifluoromethylsulfonyl)imide) The IL was
investigation. prepared and purified as a yellow liquid via an early
In this work, Lewis acidic 1-allyl-3-methylimi- method [37]. Yield: 95.5%.
dazolium chloroaluminate ILs ([Amim]Cl/AlCl3) were All ILs prepared have been characterized by nu-
first synthesized and used as electrolyte in the clear magnetic resonance (NMR). The peaks and cor-
low-temperature electrodeposition of aluminium. The responding chemical shifts obtained confirmed the
properties of [Amim]Cl/AlCl3 ILs, especially electri- structures of these ILs, and no impurity peaks were
cal conductivities have been discussed. Based on elec- found in the NMR spectra.
trochemical measurements, the electrodeposition
process of aluminium from [Amim]Cl/AlCl3 was 2.3 The measurement of physicochemical properties
studied in detail. The effects of deposition parameters,
including temperature, current density and molar ratio
of [Amim]Cl/AlCl3, on the surface morphology, X-ray The NMR experiments of ILs were performed on
diffraction (XRD) pattern and purity of deposits were a NMR spectrometer (av-400 MHz, Bruker, Switzer-
also investigated. These results were analyzed from land) at 298.2 K. In the measurement of NMR for
the perspective of microcrystalline nucleation and Lewis acidic [Amim]Cl/AlCl3 and [Bmim]Cl/AlCl3
growth. According to above studies, the optimum ex- ILs, acetone-D6 was used as an external standard to
perimental conditions for aluminium electrodeposition suppress the interference from other solvents. The
were finally determined. surface morphology and compositional analysis of
aluminium deposits were examined with emission
scanning electron microscope (SEM) (JSM-6700F,
2 EXPERIMENTAL JEOL Ltd., Japan) and energy dispersive analysis
2.1 Chemicals X-ray (EDAX) (XL30 S-FEG, FEI Co., USA) respec-
tively. And the X-ray diffraction (XRD) patterns of
aluminium deposits were collected on a X-ray dif-
All the chemicals used in this work were fractometer (X’Pert Pro MPD, PANalytical Co., Neth-
purchased commercially with analytic grade from erlands) with Cu Kα radiation (λ = 0.15405 nm). A
Alfa Aesar and Sinopharm Chemical Reagent Co. density meter (Anton Paar DMA5000, Anton Paar Co.,
N-methylimidazolium, allyl chloride, acetonitrile, Austria) with a resolution of 1 µg was employed to
1-chlorobutane and acetone were purified by distilla- determine the densities of ILs.
tion. Anhydrous aluminium chloride was used without
further purification. Prior to use, aluminium plate
(99.95%), aluminium wire (99.99%), platinum plate 2.4 Electrochemical experiments
(99.95%) and copper foil (99.95%) were polished with
emery paper, cleaned with acetone, treated with a di- All the electrochemical experiments were carried
lute hydrochloric acid/sulfuric acid mixture, finally out in an argon-filled glove box (Universal, MIK-
rinsed with deionized water and dried in air. ROUNA Co., China) where water and oxygen content
was both kept below 1×10−6. The data of electrical
2.2 Preparation and characterization of ILs conductivities were obtained through a conductivity
meter (DDS-307, Shanghai Precision & Scientific
Instrument Co., China) from 298.2 to 373.2 K. The
1-Allyl-3-methylimidazolium chloride ([Amim]Cl) other electrochemical measurements were performed
This IL was synthesized according to method described using an electrochemical workstation (CHI1140A, CH
in the literature [35]. The final product was obtained as Instrument, USA) controlled by CHI1140A software
132 Chin. J. Chem. Eng., Vol. 20, No. 1, February 2012
in a three-electrode cell. In the determination of cyclic relatively weak cation-anion interaction, high struc-
voltammograms, platinum working electrode and plati- tural symmetry of [Amim]+ cation and high molar
num plate counter electrode were used. A non-aqueous concentration of ions (8.89 mol·L−1, 298.2 K) [38, 39].
Ag+/Ag electrode (CHI112, CH Instrument, USA) Besides, [Amim]Cl and [Amim]Cl/AlCl3 are much
containing the 0.1 mol·L−1 solution of AgNO3 in easier to prepare and purify. The final IL products
[Bmim][NTf2]/acetonitrile (298.2 K) was employed as were obtained with good purity, which paved the way
reference electrode. As for the electrodeposition ex- for low-temperature electrodeposition of aluminium.
periments, the working, counter and reference elec- In the following work, Lewis acidic [Amim]Cl/AlCl3
trodes were copper foil, aluminium plate and alumin- ILs with molar ratio of 1︰2.0, 1︰1.8 and 1︰1.5
ium wire, respectively. And the electrodeposition of were used as electrolytes.
aluminium on copper substrates was conducted by
galvanostatic method. All the ILs were heated in an oil 3.1 Results of cyclic voltammetry
bath and the accuracy of temperature was ±0.5 K
(RET basic, IKA, Germany). After electrodeposition,
all the deposits were washed by acetonitrile, deionized To probe into the eletrodeposition behavior of
water and finally dried in air. aluminium, cyclic voltammetry has been used in this
work. Fig. 2 shows the cyclic voltammograms of
Lewis acidic [Amim]Cl/AlCl3 ILs recorded on a Pt
2.5 Calculation of current efficiency electrode at 303.2 K. The potential range was con-
trolled between −2.0 and 0.2 V to avoid the interfer-
The current efficiency (CE) in electrodeposition ence from the oxidation of [AlCl4]− anions and reduc-
was calculated by tion of [Amim]+ cations. It can be seen that a pair of
3 × 96485 × m peaks is obtained between −0.1 and −2.0 V in each
CE (%) = × 100 (1) sample. The reduction peaks at −2.0 V on the negative
27 × I × t scan are ascribed to the bulk electrodeposition of alu-
where the yield of aluminium (m) was determined minium, while the oxidation peaks around −0.4 V on
from the mass difference of cathode before and after the reverse scan reflect the stripping process of depos-
deposition. ited aluminium:
4[Al2 Cl7 ]− + 3e− ⎯⎯
→ Al + 7[AlCl4 ]− (2)
3 RESULTS AND DISCUSSION
7[AlCl4 ]− + Al ⎯⎯
→ 4[Al2 Cl7 ]− + 3e− (3)
Compared with the conventional ILs applied in In these ILs, the reduction of Al(III) begins to
the low-temperature electrodeposition of aluminium, occur around −1.2 V, which is much more positive
[Amim]Cl/AlCl3 and [Amim]Cl have more attractive than the standard potential of Al/Al(III) couple in
properties. As Fig. 1 illustrates, the electrical conduc- aqueous solutions (−2.47 V, vs. Ag+/Ag). The differ-
tivity of 1 ︰ 2.0 [Amim]Cl/AlCl3 is significantly ence in deposition potential of aluminium is probably
higher than that determined in the well-known 1︰2.0 attributed to the different coordination pattern of
[Bmim]Cl/AlCl3 [22]. It indicates that lower cell volt- Al(III) in [Amim]Cl/AlCl3 and aqueous solutions. As
age, higher current efficiency and higher deposition Fig. 2 shows, it can also be found that the reduction
rate can be expected in [Amim]Cl/AlCl3. Considering potential of Al(III) in ILs decreases in the order 1︰2.0
the relationship between ionic conduction and mo- [Amim]Cl/AlCl3>1 ︰ 1.8 [Amim]Cl/AlCl3>1 ︰ 1.5
lecular features of ILs, the high electrical conductivi-
ties of [Amim]Cl/AlCl3 are probably ascribed to the
[Amim]Cl/AlCl3. According to the values of molar It can be inferred that the electrodeposition of alu-
ratio and density, the molar concentration of [Al2Cl7]− minium from 1︰2.0 [Amim]Cl/AlCl3 is mainly in-
in above systems follows the same order. The predomi- fluenced by mass diffusion and kinetic limitations,
nant presence of anion can be expressed as [Al2Cl7]− such as electron transfer or surface chemical reaction
in 1 ︰ 2.0 [Amim]Cl/AlCl3 and 1 ︰ 2.0 [Bmim]Cl/ steps [23]. Similar conclusion has also been drawn
AlCl3 based on the analysis of 27Al NMR results (see from other Lewis acidic [Amim]Cl/AlCl3 ILs.
Table 1) [29]. Therefore, it is proper to state that alu-
minium is more prone to be electrodeposited from 3.2 Nucleation/growth process for electrodeposi-
Lewis acidic [Amim]Cl/AlCl3 ILs with higher molar tion of aluminium
concentration of [Al2Cl7]− anions under the same con-
ditions.
Chronoamperometric experiments have been
conducted to investigate the nucleation/growth proc-
Table 1 27Al NMR chemical shifts of 1︰2.0 ess of aluminium, which is critical to the study of
[Amim]Cl/AlCl3 and 1︰2.0 [Bmim]Cl/AlCl3 electrocrystallization mechanism and surface mor-
27
Al NMR chemical shift δ/ppm phology of aluminium deposits on copper substrates.
T/K For each experiment, the potential of working elec-
1︰2.0 [Amim]Cl/AlCl3 1︰2.0 [Bmim]Cl/AlCl3 trode was stepped from an initial value without reduc-
298.2 103.960 106.626 tion of Al(III) taking place to sufficiently negative
potentials to initiate the nucleation/growth process. A
series of typical current-time transients derived from
Additionally, cyclic voltammograms of [Amim]Cl/ different step potentials for the electrodeposition of
AlCl3 ILs at different scan rates have been determined. aluminium from Lewis acidic [Amim]Cl/AlCl3 ILs at
The cathodic limit was set to −2.1 V so that the current 313.2 K are shown in Figs. 4 and 5. Apparently, all the
density (ipc) of bulk deposition peak could be achieved. transients exhibit typical shape for a nucleation proc-
The experimental results for 1︰2.0 [Amim]Cl/AlCl3 ess. Owing to the charging of electrode double layer, a
are presented in Fig. 3. It is evident that both ipc and sharp decay of current can be observed after the ap-
the corresponding deposition peak potential ( ϕ pc ) of plication of step potential, especially at higher poten-
aluminium shift more negative when the scan rate (v) tials. The following increase in current is attributed to
increases. As shown in the inset of Fig. 3, the plots of the nucleation and growth of aluminium nuclei on
ipc against the square root of scan rate (v1/2) display copper electrode. The rising current finally reaches a
good linearity. However, the straight line obtained by maximum (im) at time (tm), when the discrete diffusion
plotting ipc vs. v1/2 does not pass through the origin as zones of growing crystallites start to overlap [22]. Af-
expected for a simple linear diffusion controlled process. ter that, the current transients decay as the deposited
(5)
To identify the nucleation model, a common
method is applied by the comparison between dimen-
sionless experimental current-time experiments and
dimensionless theoretical transients for each nuclea-
tion mechanism. Before an effective comparison can
be made, the experimental time has been corrected for
the potential-dependent induction time (t0), by refining
the time axis as t ′ = t − t0 and tm′ = tm − t0 . Estimation
Figure 4 Current-time transients resulting from potential
step experiments on Cu electrodes in 1︰2.0 [Amim]Cl/AlCl3 of t0 can be made according to the method described
at 313.2 K in [42]. The dimensionless plots of (i / im )2 vs. (t ′ / tm′ ) 2 ,
deposition potential/V: 1—−0.25; 2—−0.20; 3—−0.15; 4—
−0.10; 5—−0.05 along with theoretical curves obtained from Eqs. (4)
and (5) are presented in Fig. 6. As can be seen, the
electrodeposition of aluminium on copper substrates
fits well to the theoretical model for three-dimensional
instantaneous nucleation with diffusion-controlled
growth.
layer becomes thicker. The time required to reach im is Figure 6 Comparison of the dimensionless experimental
shorten when the applied potential is more negative in current-time transients, derived from the results in Fig. 4,
the same IL (see Fig. 4). Meanwhile, Fig. 5 indicates with the theoretical curves for instantaneous nucleation and
progressive nucleation
that im increases in the order 1 ︰ 1.5 [Amim]Cl/ instantaneous; progressive; ● −0.15 V; ▽ −0.20 V;
AlCl3<1 ︰ 1.8 [Amim]Cl/AlCl3<1 ︰ 2.0 [Amim]Cl/ ◆ −0.25 V
AlCl3, while tm changes in an opposite trend. These
results suggest that the nucleation density and rate of
aluminium are enhanced as the molar concentration of 3.3 Surface morphology of aluminium deposits
Al(III) increases in Lewis acidic [Amim]Cl/AlCl3 ILs.
The bulk electrodeposition of aluminium on for- It is necessary to explore optimum experimental
eign substrates often involves a three-dimensional conditions of aluminium deposition on copper substrates
nucleation/growth process. Some classic models have for industrial application. Consequently, the effects of
been proposed to describe this process, typically as main experimental parameters, including temperature
hemispherical diffusion-controlled growth of the nuclei (313.2-373.2 K), current density (5-25 mA·cm−2) and
[40]. This model has two limiting cases: instantaneous molar ratio of [Amim]Cl to AlCl3 [(1︰1.5)- (1︰2.0)]
nucleation and progressive nucleation. In the first lim- on the surface morphology of aluminium deposits and
iting case, all the nuclei are instantaneously formed at current efficiency have been investigated. The deposi-
the beginning of chronoamperometric experiments. tion time and stirring speed were kept as 30 min and
The second case refers to a situation where the nuclei 300 r·min−1, respectively.
are generated gradually as the deposition experiment In this work, it is found that temperature has a
proceeds. They are expressed respectively by [41] significant impact on the surface morphology of alu-
minium coatings. Bright, smooth and dense deposits
( i / im )2 = 1.9542 ( t / tm )−1 {1 − exp ⎡⎣ −1.2564 ( t / tm )⎤⎦}
2
were obtained in the temperature range from 313.2 to
353.2 K. However, the deposits turned dark and poorly
(4) adherent at the temperature higher than 353.2 K,
Chin. J. Chem. Eng., Vol. 20, No. 1, February 2012 135
particularly at 373.2 K. For illustration, the SEM nuclei were generated so that the deposits displayed
micrographs of aluminium deposits prepared from finer microstructures. After the nucleation density
1︰2.0 [Amim]Cl/AlCl3 under different temperatures reached a maximum value, the crystallites experienced
are shown in Fig. 7. When the temperature rises, the a rapid overlapping growth at 25 mA·cm−2, which
average grain size of crystallites increases from 1 to eventually resulted in larger grain size and rougher
10 µm approximately. The deposits produced at 313.2, deposits.
333.2 and 353.2 K all display dense and homogeneous As is well-known, the content of AlCl3 in chloro-
microstructures with spherical particles. By contrast, a aluminate ILs is another important parameter in elec-
relatively rough and incompact surface with large trodeposition of aluminium. Herein, experimental re-
clusters of platy-shaped crystals is observed on the sults reveal that the deposits increase in grain size and
sample obtained at 373.2 K. The difference in micro- become less compact when the molar ratio of
structures of aluminium deposits probably results [Amim]Cl to AlCl3 [(1︰1.5)- (1︰2.0)] rises under
from the variation of electrical conductivities and dif- the same conditions. For example, the particle size of
fusion rate of ions in IL at different temperatures. A resulting deposits in the studied ILs follows the order
higher temperature followed by nonuniform distribu- 1 ︰ 1.5 [Amim]Cl/AlCl3>1 ︰ 1.8 [Amim]Cl/AlCl3>
tion of current density and high deposition rate is 1︰2.0 [Amim]Cl/AlCl3 at 20 mA·cm−2 and 313.2 K.
prone to promote the dendritic growth, which often This trend of variation of surface morphology can be
leads to rough coatings. seen clearly from the SEM micrographs of aluminium
Searching for appropriate current density is key deposits shown in Figs. 9 and 7 (a). Based on the dis-
to the electrodeposition of aluminium. Figs. 8 and 7 (a) cussion above, it is proposed that the molar concentra-
show the SEM images of deposits prepared from tion of reducible [Al2Cl7]− anions and corresponding
1︰2.0 [Amim]Cl/AlCl3 with different current densi- nucleation density of aluminium on copper surface
ties at 313.2 K. It is apparent that the grain size of decrease as the molar content of AlCl3 increases in
aluminium crystallites decreases from about 3 to 1 µm [Amim]Cl/AlCl3 ILs. Consequently, uniform, compact
as current density increases in the range of 5-20 and dense aluminium coatings have been obtained
mA·cm−2. Nevertheless, the deposited crystals grow to from 1︰2.0 [Amim]Cl/AlCl3 more easily than 1︰1.5
a larger size when the current density increases to 25 and 1︰1.8 [Amim]Cl/AlCl3.
mA·cm−2. Among these samples, the deposits obtained In summary, the aluminium deposits prepared
at 15 and 20 mA·cm−2 have the most smooth and from Lewis acidic [Amim]Cl/AlCl3 ILs have finer
compact structures. It can be inferred from above microstructures than those produced in many conven-
phenomena that the nucleation density on copper elec- tional chloroaluminate-based ILs at the studied ex-
trode surface increased with current density rising perimental conditions [22-24]. Moreover, most of the
between 5 and 20 mA·cm−2. As a result, more aluminium deposited films adhered well to the substrates without
In this work, more than 95% current efficiency optimal efficiency was obtained as 99.6% from 1︰2.0
has been achieved for all the aluminium deposits. The [Amim]Cl/AlCl3 at 313.2 K and 20 mA·cm−2, while
Chin. J. Chem. Eng., Vol. 20, No. 1, February 2012 137
4 CONCLUSIONS
Figure 11 XRD pattern of the aluminium deposits ob- A systematic research on the low-temperature
tained from 1︰2.0 [Amim]Cl/AlCl3 at 20 mA·cm−2 electrodeposition of aluminium from Lewis acidic
deposition temperature/K: 1—313.2; 2—353.2; 3—373.2
[Amim]Cl/AlCl3 ILs is presented. These electrolytes
138 Chin. J. Chem. Eng., Vol. 20, No. 1, February 2012
[BMP]TFSA and [EMIm]TFSA ionic liquids with Au(III)”, J. Phys. and hydrophobic room temperature ionic liquids incorporating the
Chem. C, 113, 13266-13272 (2009). imidazolium cation”, Green Chem., 3, 156-164 (2001).
31 Armand, M., Endres, F., MacFarlane, D., Ohno, H., Scrosati, B., 37 Bonhôte, P., Dias, A., Papageorgiou, N., Kalyanasundaram, K.,
“Ionic-liquid materials for the electrochemical challenges of the fu- Grätzel, M., “Hydrophobic, highly conductive ambient-temperature
ture”, Nature Mater., 8, 621-629 (2009). molten salts”, Inorg. Chem., 35, 1168-1178 (1996).
32 Zein El Abedin, S., Moustafa, E.M., Hempelmann, R., Natter, H., 38 Zheng, Y., Xuan, X.P., Wang, J.J., Fan, M.H., “The enhanced disso-
Endres, F., “Additive free electrodeposition of nanocrystalline alu- lution of β-cyclodextrin in some hydrophilic ionic liquids”, J. Phys.
minium in a water and air stable ionic liquid”, Electrochem. Com- Chem. A, 114, 3926-3931 (2010).
mun., 7, 1111-1116 (2005). 39 Xuan, X.P., Guo, M., Pei, Y.C., Zheng, Y., “Theoretical study on
33 Zein El Abedin, S., Moustafa, E.M., Hempelmann, R., Natter, H., cation-anion interaction and vibrational spectra of 1-allyl-3-
Endres, F., “Electrodeposition of nano- and microcrystalline alu- methylimidazolium-based ionic liquids”, Spectrochim. Acta A, 78,
minium in three different air and water stable ionic liquids”, Chem- 1492-1499 (2011).
PhysChem, 7, 1535-1543 (2006). 40 Scharifker, B., Hills, G., “Theoretical and experimental studies of
34 Rodopoulos, T., Smith, L., Horne, M.D., Rüther, T., “Speciation of multiple nucleation”, Electrochim. Acta, 28, 879-889 (1983).
aluminium in mixtures of the ionic liquids [C3mpip][NTf2] and 41 Lee, J.J., Miller, B., Shi, X., Kalish, R., Wheeler, K.A., “Aluminum
[C4mpyr][NTf2] with AlCl3: An electrochemical and NMR spec- deposition and nucleation on nitrogen-incorporated tetrahedral amor-
troscopy study”, Chem. Eur. J., 16, 3815-3826 (2010). phous carbon electrodes in ambient temperature chloroaluminate
35 Zhang, H., Wu, J., Zhang, J., He, J.S., “1-Allyl-3-methylimidazolium melts”, J. Electrochem. Soc., 147, 3370-3376 (2000).
chloride room temperature ionic liquid: A new and powerful non- 42 Tsuda, T., Boyd, L.E., Kuwabata, S., Hussey, C.L., “Electrochemis-
derivatizing solvent for cellulose”, Macromolecules, 38, 8272-8277 try of copper(I) oxide in the 66.7-33.3 mol% urea-choline chloride
(2005). room-temperature eutectic melt”, J. Electrochem. Soc., 157, F96-F103
36 Huddleston, J.G., Visser, A.E., Reichert, W.M., Willauer, H.D., Broker, (2010).
G.A., Rogers, R.D., “Characterization and comparison of hydrophilic