SYMMETRICALLY POLYSULFONE MEMBRANES OBTAINED BY SOLVENT EVAPORATION USING CARBON NANOTUBES AS
ADDITIVES. SYNTHESIS, CHARACTERIZATION AND APPLICATIONS
A.C. NECHIFOR, V. PANAIT, L. NAFTANAILA, D. BATALU, S.I. VOICU*
Politehnica University of Bucharest, Department of Analytical Chemistry and Environmental Protection, 1 Polizu St., 011061 Bucharest, Romania,
This paper presents the synthesis and characterization of a symmetric polysulfone membrane obtained by solvent evaporation using carbon nanotubes as additives in polymer solution. The membranes were synthesized from N, N' - dimethyl formamide with dispersed amino carbon nanotubes. The obtained symmetric membranes by solvent evaporation were characterized by Scanning Electron Microscopy, Fourier transformed infrared Spectroscopy, Thermal Analysis and were used as membrane reactors for carbon nanotubes functionalization. The composite membrane with functionalized carbon nanotubes were characterized by NMR spectrometry.
(Received April 12, 2013; Accepted June 14, 2013)
Keywords: Polysulfone symmetric membrane, Carbon nanotubes, Membrane reactor
1. Introduction
In the last decades, the polymeric materials have been intensively studied due to their application possibilities [1-3] Since the begging of this century, when the first microporous membranes were obtained by Bechold [4], was intuitively recognized the potential of synthetic membranes for separation of species dissolved in a solvent, but industrial applications of membranes remained limited until the development of the first asymmetric membranes by Loeb and Sourirajan [5] in 1962. These membranes present a practical value due to an extremely thin active layer with a thickness of approximately 0.1 to 0.5 µm from the surface membrane. This layer is supported by microporous structure of the membrane with an average thickness between 0.1 and 0.2 mm, which provide mechanical stability to the membrane. Due to this very thin active layer, the hydrodynamic resistance of a asymmetric membrane is very low compared with symmetric membranes with the same selectivity [5, 6].
Synthetic polymeric membranes can be obtained by several methods, the most versatile being the synthesis by phase inversion. Most commercial membranes are produced by different phase inversion techniques. The concept of phase inversion was introduced by Kesting [6, 7] in the 70s and can be summarized as follows: a homogenous polymer solution is transformed into a two-phase system in which the solid phase is the polymer that will form the membrane, while the liquid phase (which is in a very small amount in polymer structure) will occupy and lead to pore formation.
In order to prepare membranes it is necessarily to use medium or high concentrated polymer solutions. In such systems, the crystallization time becomes important because the key steps for the membrane formation are nucleation and the growth of the crystal structure. Moreover,
* Corresponding author: [email protected]
where transfe nuclea a limit mostly succes flow th contac Previo immob reactor in two carbon lie surf membr The m nanotu Scanni were u nanotu
formam to the polym of 250 membr carbon
Amino with c hydroc
crystallizatio er rate and th ation rate wil
A too quic ted ability to The symm y used as m ssfully been
hrough the p ct allows for ous work [
bilization of r. Based on t groups. The nyl and/or hy face of the n
This paper rane, by the membranes w
ubes. The ob ing Electron used as mem ubes were co
2. Exper 2.1 Synthe Polysulfon mide (DMF) formed poly meric film wit 0 µm and th ranes were w n nanotubes t
2.2 Functi The nanotu o crown ethe cyanuric chl chloric acid r
F
on is induced he nucleatio l induce a lo ckly mass tra grow the po metric membr
membrane r applied as m pores of a m
r high catal [11] reporte f enzymes by the reactivity e first one is ydroxyl funct
anotubes [11 r presents the use of a me were synthesi
btained sym Microscopy mbrane reacto valent functi
rimental esis of polysu ne (Ultrason ) at a concent ymer solution
th dispersed he membeane
washed conse traces and im ionalization
ubes were fu er (Sigma A loride in dio resulted in re
Fig. 1 Reaction
d by a chang on rate will a w crystalliza ansfer will le olymer structu
ranes presen reactors. Cat microstructur membrane wh lytic activity ed the fun y the use of y of carbon n based on che tions. The se 1, 12].
e synthesis o ethod (solven
zed from N, mmetric mem
y, Fourier tra ors for carbo ionalized wit
ulfone-amin S3010 prov tration of 15 n at 1% relat
carbon nano es were synt equently with mpurties from
of amino ca functionalized Aldrich) and oxan in the eaction) [13,
n equations fo
ge in the syst appear. A lo ation phenom ead to the for
ure [5].
nt a lower im talytic mem red reactors hich have be
y with negl nctionalizatio f a polysulfo nanotubes, th emical oxida econd group of a symmetr nt evaporatio N' - dimeth mbranes by s ansformed in on nanotubes
th crown eth
o carbon na vided by B 5% mw. Ami ated to polym otubes was de thesized by th iso-propan m membrane arbon nanotu
d with crow aza-crown
presence o 14]. The rea
or the synthesi
tem composi ow mass tran mena and a m rmation of m mportance fo mbranes with [8, 9]. If th een impregna ligible mass on of carb one asymme he functional ation of the n
is based on ric polysulfon
on) witch le hyl formamid
solvent evap nfrared Spect s functionali hers derivativ
anotubes com ASF) was d ino carbon na mer and dispe eposited onto solvent evap nol and deion surface remo ubes.
wn ethers usi ether (Sigma of triethylene action equatio
is of crown eth
tion, a comp nsfer rate wi membrane can many nucleati
or separation hout separat he reactant m
ated with cat transport r on nanotub tric membra ization react nanotubes lea
direct additi ne-carbon na ads to asym de with dispe poration wer troscopy, Th zation. The ves.
mposite mem dissolved in
anotubes (Na ersed by son o a glass at a poration at 6 nized water f
oval.
ng cyanuric a Aldrich) w e amine (fo ons are prese
hers derivativ
petition betw ill compared nnot be obta ion centers b n processes, tive function mixture is fo
talyst, the in resistances [ bes with c ane as a mem tions can be d
ading to carb ion to the gra anotubes com mmetric mem
ersed amino re characteri hermal Analy used amino
mbranes n N, N' - di
anocyl) were nication for 1 a standard th 60oC. The o for the polym
chloride as were functio or neutraliza ented in Fig.
ves
ween the d to the ained.
but with but are n have orced to ntensive [9, 10].
ovalent mbrane divided boxylic, aphitic- mposite mbranes.
carbon ized by ysis and
carbon
imethyl e added 1h. The ickness btained mer and
linker.
onalized ation of
1.
For the functionalization of amino carbon nanotubes was performed using the membrane as a reactor. Free parts of nanotubes in membrane pores were functionalized with crown ether derivatives in a mixture of water and dioxan at 70oC. The mixture of water with organic solvent is necessarily because the pure dioxan can affect the polymer by dissolving it. The reaction equations are presented in Fig. 2.
Fig. 2 Reaction equations for the functionalization of carbon nanotubes
2.3 Characterization techniques
The membranes were characterized by Scanning Electron Microscopy (FEI Instrument, the samples were covered with a thin layer of gold [15-17]), thermal analysis (TGA-DSC) using an Universal V4.5A TA instrument and FT-IR spectroscopy analysis (Bruker Tensor 27 with ATR annex) and NMR spectrometry Bruker 400 MHz.
3. Results and discussions
Usually, by phase inversion technique, asymmetric membranes are obtained (a big difference between pore sizes on each side of the membrane). If the phase inversion is realized using a bath with a non solvent for the polymer, the speed formation for the membrane is high (not more than 2 min), but if the membrane is synthesized by solvent evaporation, the speed formation for the membrane is very low (several hours or even days, depending by the boiling point of the polymer solvent and the environmental temperature).
The formation mechanism and obtained membranes were investigated by several analysis methods which will be explained below. The main absorption bands are 1150 cm-1 (SO2 group from polysulfone), 2870 cm-1 (attributed to CH3 groups from polysulfone), 3210 cm-1 (attributed to aromatic rings from polysulfone), and 1620 cm-1 (attributed to NH2 group from amino carbon nanotubes) [11]. In the FT-IR spectrum (Fig. 3) we can observe that on polysulfone-carbon nanotubes composite membrane a band at 1647 cm-1 disappears. This the typical band for the plane specific vibration of C=O bond from DMF solvent (detailed in Fig. 4). This can be explained by the presence of carbon nanotubes and to the fact that they cannot be wetted by solvents. In the structure of a polysulfone membrane, a significant amount of polymer solvent remains in the membrane pores (enough quantity to give specific absorption in the IR spoectrum). In the presence of carbon nanotubes, this does not happen due to the delocalized electrons on the surface of nanotubes which makes impossible the interaction of nanotubes polar solvent molecules. This explain the disappearance of the solvent from membrane structure. More than that, this must be correlated with the fact the evaporation temperature for DMF is 180oC, so the membrane formation is a very slow process and this is also a reason for the shape and dimensions of
membrane pores. In the FT-IR spectrum it also can be observed a slight shift of absorption bands to the left with 1-2 cm-1 (due to the electrical properties and influence of carbon nanotubes through polymer).
Fig. 3 The FT-IR spectrum of polysulfone and polysulfone-carbon nanotubes composite membranes
These chemical shifts suggests a non-covalent interaction between carbon nanotubes and polymer. Another interesting aspect is the fact in the membrane with carbon nantubes, the water is not present at all, specific bands for hydrogen bond associations (2800-3600 cm-1) being absent.
This observation should also be correlated with the inability of carbon nanotubes to be wetted by any solvent. The formation mechanism for these membranes is similar to the phase inversion in the presence of a solvent, but in this case, the non-solvent is represent by the water vapors from atmosphere.
Fig. 4 Details of polymer bands absorptions on FT-IR spectrum on polysulfone (left) and polysulfone-carbon nanotubes composite membrane (right)
From Scanning Electron Microscopy micrographs (Fig. 5) we can observe two membrane surfaces. We cannot say that one is the active surface and the other one is the porous surface because both exhibits pores with shape and dimensions comparable. But, the membrane surface
which was not in contact with the substrate (glass) present two types of pores: one large with dimensions comparable to those on the other side and some smaller, surrounding the large pores.
The presence of these small pores can be explained by the slow coagulation process during solvent evaporation. The large pores are determined by the mass transfer which occurs during the evaporation of the main amount of solvent and the smaller ones can be explained by a post- evaporation process of solvent traces from polymer block. Clearly, these small pores are determined by a very small amount of solvent, evaporated later than the main quantity. The average dimension of pores is between 20 and 50 µm and the small ones has a diameter around 2-4 µm. These pore dimensions recommend these membranes for classic filtration or clarification processes (separation of particles larger than 50 µm).
a b
c d
Fig. 5 Scanning Electron Microscopy of one side of the membrane (a - x500 and b - x2000) and of the other side of the membrane (c - x500 and d - x2000)
The thermal behaviour of the synthesized membranes was analyzed by Thermal Gravimetric Analysis (Fig. 6). Due to the small amount of carbon nanotubes, the thermal properties are not very different, but a small difference in the thermal degradation can be observed.
The weight loss for polysulfone-carbon nanotubes composite membrane is slightly and at the begging of the degradation the allure of the curve suggests the absence of water from membrane structure which is in accord with FT-IR data.
realize propan presen The m is negl This c conseq suppor temper
Fig
In order t ed. It was te nol at room nted in Fig. 7 membranes pr
ligible which can be expla quence the in
rted by the rature, the w
Polysu g. 6 The Therm
to evaluate t ested the me temperature 7 and for wa resents high
h suggests a ained by a c
ncrease of m water flu water flux pre
Polys
ulfone - carbo mal Gravimetr
the membra embrane perm e and atmosp
ater permeat fluxes due to a very stable
cross-linking mechanical p uxes at diff sents a high
sulfone mem
on nanotubes c tric Analysis c
ane performa meability fo pheric press tion at differ o the pore di e behaviour o g effect of c properties of ferent tempe increase due
brane
composite mem curves for obta
ances, the s or deionized sure. The da
rent tempera mensions. T of polymer s carbon nanot the compos eratures. No e to pore dila
mbrane ained membra
olvent perm water, meth ta for alcoh atures are pre
he variation structure to tubes to me ite material.
ormally, wit atation [18]. H
anes
meation teste hanol, ethan hols permeat esented in T of fluxes wi solvent cros embrane, hav
This conclu th the incre Here, the var
es were ol, iso- ion are Table 1.
ith time ss-flow.
ving as usion is ease of
riations
are not so high due to the membrane stability. This behaviour is important due to the requirements for thermal stability if the carbon nanotubes from membrane structure are involved in chemical reactions.
Fig. 7 Solvents permeation (methanol, ethanol, iso-propanol) through polysulfone-carbon nanotubes composite membrane
Table 1. Water fluxes at different temperatures
Temperature (oC) 20 30 40 50 60 70
Water Flux (106 L/m2 x h)
0,2435 0,2439 0,2441 0,2631 0,2698 0,2732
Previous work demonstrated the possibility to functionalized carbon nanotubes based on the fact that parts from carbon nanotubes are free in membrane pores. The used carbon nanotubes in this research presents free amino groups at surface (5%) in enough quantity to make possible the functionalization process (Fig. 8).
Fig. 8 Schematic representation of functionalization reaction in membrane pores Solvents permeation
0,4 0,41 0,42 0,43 0,44 0,45 0,46 0,47
0 0,5 1 1,5 2 2,5 3 3,5
Time (h)
Flux (10^6 L/m^2 x h)
Methanol Ethanol Iso-propanol
The access of reagents to reaction centres is realized by diffusion mechanism through membrane pores. The main advantage of this method is the homogenous functionalization (even if the free reaction centres are not in high number in membrane pores, there are many available than in a solution). Cyanuric chloride is a widely used linker for immobilization of different chemical species due to the characteristics of aromatic ring. After the first reaction at a chloride atom, the aromatic ring is deactivated and on this way the reactions in which two molecules from the same species can react will be avoided. After this deactivation the possibility to perform a second reaction at a second chloride atom is the increase of reaction temperature [11].
NMR spectra of polysulfone-carbon nanotubes functionalized with amino-benzo crown ether
NMR spectra of polysulfone-carbon nanotubes functionalized with aza crown ether Fig. 9 NMR spectrum of composite membranes with functionalized carbon nanotubes
The NMR analysis was performed by dissolving the membrane in deuterated chloroform.
The chemical shift at 1.845 ppm is attributed to NH2 protons from amino carbon nanotubes, 6.941 ppm, 6.969 ppm, 7.006 ppm, and 7.033 ppm are attributed to H from aromatic ring and 1.376 ppm is attributed to proton from –CH3. The chemical shifts at 2.975 ppm (in the case of amino-benzo- crown ether) and 2.977 ppm are attributed to the proton from -NH- which demonstrate the functionalization of carbon nanotubes. The small intensity of the signal can be explained by the small number of amino groups which can be involved in functionalization reaction.
For different applications, the membrane can be used as it is (for example for the separation of different cations or membrane sensors), or the carbon nanotubes can be recovery from the membrane structure, by dissolving the membrane in a large quantity of polymer solvent [11]. Functionalized carbon nanotubes can be used for the manufacturing of sensors or biosensors.
4. Conclusions
The synthesis and characterization of a polysulfone-carbon nanotubes composite membrane was presented in this paper. The membrane was synthesized by solvent evaporation (procedure which usually leads to asymmetric membranes), but symmetric material was obtained.
This was explained by the presence of carbon nanotubes and the observation was supported by analysis methods. The membrane was used as a membrane reactor for carbon nanotubes functionalization with crown ethers derivatives.
Acknowledgements
Authors recognize financial support from the European Social Fund through POSDRU/89/1.5/S/54785 project: Postdoctoral Program for Advanced Research in the field of nanomaterials.
References
[1] M. Ulbricht, Polymer 47(7), 2217 (2006).
[2] R. Ianchis, D. Donescu, M.C. Corobea, C.L. Nistor, C. Petcu, R. Somoghi, R.D. Fierascu, Optoelectron. Adv. Mater. - Rapid Commun. 5(12), 1352 (2011).
[3] G. Nechifor, S.I. Voicu, A.C. Nechifor, S. Garea, Desalination 241, 342 (2009).
[4] H. Bechold, Biochem. Z. 6, 397 (1907).
[5] S. Loeb, S. Sourirajan, Advan. Chem. Ser. 38, 117 (1962).
[6] R.E. Kesting, Synthetic Polymer Membranes, McGrow Hill, New York, 1972.
[7] J.G. Wijmans, Preparation of asymmetric membranes by the phase inversion process, Ph.D.
Thesis, Twente Enschede, The Netherlands, 1984.
[8] A.G. Dixon, Int. J. Chem. React. Eng. 1 (R6), 1 (2003).
[9] T. Westermann, T. Melin, Chemical Engineering and Processing: Process Intensification 48, 17 (2009).
[10] V.T. Zaspalis,W. van Praag, K. Keizer, J.G. van Ommen, J.R.H. Ross, A.J. Burggraaf, Appl.
Catal. 74(2), 205 (1991).
[11] S.I. Voicu, A.C. Nechifor, O. Gales, G. Nechifor, Proceedings of SPIE - The International Society for Optical Engineering, 8068, art. no. 80680Y, (2011).
[12] B. Bhushan, (ed.), Springer Handbook of Nanotechnology, Springer-Verlag Berlin Heidelberg, 3 (2004).
[13] F.D. Balacianu, A.C. Nechifor, R. Bartos, S.I. Voicu, G. Nechifor, Optoelectron. Adv.
Mater. – Rapid Commun. 3(3), 219 (2009).
[14] A.C. Nechifor, M.G. Stoian, S.I. Voicu, G. Nechifor, Optoelectron. Adv.
Mater. – Rapid Commun. 4(8), 1118 (2010).
[15] F. Miculescu, I. Jepu, C. P. Lungu, M. Miculescu, M. Bane, Digest Journal of Nanomaterials and Biostructures 6(2), 769 (2011).
[16] F. Miculescu, I. Jepu, C. Porosnicu, C. P. Lungu, M. Miculescu, B. Burhala, Digest Journal of Nanomaterials and Biostructures 6(1), 307 (2011).
[17] F. Miculescu, M. Miculescu, L. T. Ciocan, A. Ernuteanu, I. Antoniac, I. Pencea, E. Matei, Digest Journal of Nanomaterials and Biostructures 6(3), 1117 (2011).
[18] S.I. Voicu, A. Dobrica, S. Sava, A.Ivan, L. Naftanaila, J. Optoelecron. Adv. Mater.
14(11-12), 923 (2012).
