Amine-functionalized mesoporous silica for urea adsorption

Amine-functionalized mesoporous silica for urea adsorption

Materials Chemistry and Physics xxx (2016) 1e7 Contents lists available at ScienceDirect Materials Chemistry and Physics journal homepage: www.elsev...

2MB Sizes 13 Downloads 113 Views

Recommend Documents

No documents
Materials Chemistry and Physics xxx (2016) 1e7

Contents lists available at ScienceDirect

Materials Chemistry and Physics journal homepage:

Amine-functionalized mesoporous silica for urea adsorption Wee-Keat Cheah a, Yoke-Leng Sim b, Fei-Yee Yeoh a, * a b

School of Materials and Mineral Resources Engineering, Universiti Sains Malaysia Engineering Campus, 14300 Nibong Tebal, Penang, Malaysia Department of Chemical Science, Universiti Tunku Abdul Rahman Perak Campus, Jalan Universiti, Bandar Baru Barat, 31900 Kampar, Perak, Malaysia

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 Urea adsorption of silica is better than carbon despite the higher surface area.  Surface modification e surface architecture and functional group modification.  Surface modification e improve adsorption capacity, thus improve overall design.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 November 2015 Received in revised form 2 March 2016 Accepted 5 March 2016 Available online xxx

The limitations of the present hemodialysis have led towards the development of several wearable artificial kidney prototypes. The most important component of the miniaturized model, as compared to the hemodialysis, is the closed-system dialysate. Typically, these models utilize activated carbon as adsorbents. With superior adsorbents as replacements, the amount of dialysate required to remove uremic toxins could be reduced and regenerated. Mesoporous silica could perform better than generic activated carbon since the silica surface could be functionalized for target specific adsorption. This paper evaluates the performance of mesoporous silica SBA-15 and amine-functionalized SBA-15 for the removal of major uremic toxin constituent urea. Results are promising as mesoporous silica could potentially replace activated carbon in artificial kidney applications due to the larger urea adsorption capacity and improved adsorption kinetics. Functionalization of mesoporous silica further improved its urea adsorption capacity. Though activated carbon has theoretically higher surface area, its low affinity towards urea remains a drawback. © 2016 Published by Elsevier B.V.

Keywords: Microporous materials Adsorption Chemisorption Surfaces

1. Introduction One of the major shortcomings of the current hemodialysis system is intermittent treatment, as opposed to the continuously functional human kidney. In order to solve this existing problem, several wearable artificial kidney models have been developed

* Corresponding author. E-mail address: [email protected] (F.-Y. Yeoh).

[1,2]. A wearable artificial kidney is actually a scaled down version of a conventional hemodialysis, though these prototypes are in clinical trials at best. The key component of a wearable artificial kidney model is the dialysate regeneration system. In order for an artificial kidney to be wearable, the bulk volume of dialysate has to be reduced and regenerated within a closed system. These systems typically utilize activated carbon to adsorb uremic toxins from the spent dialysate [3,4]. Although activated carbon is a universal adsorbent material, its affinity towards urea is rather low [5]. Urease is commonly applied in such systems to catalyze the 0254-0584/© 2016 Published by Elsevier B.V.

Please cite this article in press as: W.-K. Cheah, et al., Amine-functionalized mesoporous silica for urea adsorption, Materials Chemistry and Physics (2016),


W.-K. Cheah et al. / Materials Chemistry and Physics xxx (2016) 1e7

decomposition of urea into ammonia and carbon dioxide [3,4]. The drawback of this enzyme is the generation of carbon dioxide gas, which directly complicates the closed fluid system. Thus, a separate carbon dioxide gas removal system would be required [6]. Application of solid adsorbent based urea removal could simplify the entire wearable artificial kidney device setup. With reference to the long list of uremic toxins, urea is the largest uremic toxin constituent [7]. The amount of activated carbon required to adsorb urea (without the inclusion of urease enzyme) is more than that required for other uremic toxins such as creatinine and uric acid. Based on this consideration alone, from a material selection standpoint, an alternative material should be used in place activated carbon for urea adsorption. Thus, a complementary adsorbent material which targets urea should be included to improve the overall uremic toxin adsorption capacity of the entire artificial kidney system. The concentration of urea (H2NCONH2) in a healthy human adult is less than 6700 mM, while the concentration in a kidney failure patient is 38,000 ± 18,000 mM [7]. The molecular dimension of urea is estimated to be 0.56 nm  0.63 nm  0.30 nm [8]. Apart from conventional activated carbon, based on high surface area alone, zeolite [9] and mesoporous silica are two possible candidate materials for urea adsorption. The pore size of mesoporous silica could be varied on a larger scale compared to zeolite. Additionally, various surface functional groups could be introduced on mesoporous silica, an advantage which zeolite does not possess. The fact that functional groups, such as amine, could easily be introduced on the surface of mesoporous silica is worth investigating [10e12]. The presence of these functional groups would enhance the interaction between adsorbent and adsorbate, thus improving the overall adsorption capacity. Additionally, it is worth mentioning that silica nanoparticles have been successfully functionalized for hemocompatibility in a bloodesilica interface [13]. Yet again, the flexibility of mesoporous silica functionalization could be further exploited to improve its biocompatibility in applications such as artificial kidney systems. Common variants of mesoporous silica include MCM-41 [14], FSM-16 [15] and SBA-15 [16]. Although mesoporous silica (such as MCM-41) with smaller pore diameter possesses higher surface area for adsorption, the size of adsorbate should be the main consideration. Thus, SBA-15 is commonly selected due to its larger pore diameter. SBA-15 was selected for this current work due to its large pore diameter, SBA-15 with larger pore diameter could be functionalized without risking the closure of pores. The introduction of surface functional groups would result in thicker pore walls and smaller pore diameter compared to the original SBA-15. Two common methods to introduce functional groups on the surface of mesoporous silica are post-synthesis grafting and co-condensation method [17,18]. Between these two methods, post-synthesis grafting produces mesoporous silica with stronger pore walls and more ordered structure [19,20]. A stable pore structure is more desirable for adsorption applications. Thus, based on this consideration, post-synthesis grafting method is preferred for the introduction of functional groups on the surface of synthesized mesoporous silica SBA-15. In this current work, mesoporous silica SBA-15 and aminefunctionalized mesoporous silica were synthesized and characterized. The performance of four adsorbents for urea adsorption are evaluated; i.e. commercial dense silica, mesoporous silica SBA-15, amine-functionalized SBA-15 and commercial activated carbon. A better packing of adsorbed urea could be achieved through crystallization of urea, initiated via a seeding process. The single oxygen atom of urea would form hydrogen bonding with two NeH groups (from amine group of urea) during crystallization [21]. Amine group is chosen as the surface functional group for grafting

to reproduce a similar crystallization process, which would greatly improve the capacity of the entire urea adsorption process. 2. Experimental 2.1. Synthesis of mesoporous silica SBA-15 In a standard synthesis using soft template route, 4 g Pluronic P123 (BASF) was mixed with 30 ml deionized water and 120 ml 2 M HCl (Merck) at room temperature, with constant stirring. The temperature of the mixture was then increased to 40  C in a water bath. 8.5 g of tetraethyl orthosilicate (ACROS), abbreviated as TEOS, was subsequently added dropwise into the solution while the solution was stirred vigorously. Once TEOS dripping was completed, the solution was stirred at 120 rpm for 20 h. Next, the suspension was transferred into a Teflon bottle and hydrothermally aged at 100  C for 24 h in an oven. After the aging process, the solution was then cooled to room temperature and the resulting solid was recovered through centrifugation process. The solid product was washed and dried at 100  C overnight. The dried powder was calcined at 500  C for 5 h to burn off the Pluronic P123 template to reveal the formed pores. Resulting SBA-15 mesoporous silica is denoted S. The synthesized SBA-15 was then functionalized with amine group via post-grafting synthesis method. 1 g of synthesized powder was mixed with 3 ml (3-aminopropyl)triethoxysilane (APTES) (Aldrich), refluxed at 80  C for 20 h. The specific amine group introduced on the mesoporous silica surface is 3-aminopropyl. The solid product was filtered, washed and dried overnight. The amine-functionalized SBA-15 is denoted N. Aside from synthesized samples S and N, two other adsorbents were used as a control for subsequent urea adsorption test and materials characterization techniques, i.e. commercial activated carbon (denoted AC) and commercial dense silica powder (denoted C). The commercial coal derived granular activated carbon was obtained from I-Chem Solution Sdn Bhd. 2.2. Characterization of synthesized SBA-15 The surface functional groups of as-synthesized samples were determined by Fourier Transform Infrared Spectroscopy (FTIR) technique with Perkin Elmer Spectrum One using KBr pellet method. The formed pores within the as-synthesized samples were observed using High-Resolution Transmission Electron Microscope (HRTEM) 200 kV with Field Emission (FEI Tecnai G2 20 S-Twin). The surface area and pore characteristics of the as-synthesized and commercial samples (S, N, C, AC adsorbents) were determined by nitrogen adsorption at 77 K using apparatus Quantachrome autosorb iQ. The surface area of various adsorbents was determined using BrunauereEmmetteTeller (BET) model. The pore diameter and pore size distribution were obtained using density functional theory (DFT) and BarretteJoynereHalenda (BJH) models. The pore diameter was only measured up to 0.5 nm since diameters smaller than 0.5 nm could not be utilized for urea adsorption (urea size 0.56 nm  0.63 nm  0.30 nm). 2.3. Urea adsorption test This test was carried out to evaluate the performance of urea adsorption by various adsorbents. Aside from samples S and N, 2 other samples, C (commercial dense silica) and AC (commercial activated carbon), were introduced as controls for the urea adsorption test. Solution concentration was determined using UVeVis Spectroscopy. Initially, a calibration curve was developed by plotting the intensity against concentration for a series of urea

Please cite this article in press as: W.-K. Cheah, et al., Amine-functionalized mesoporous silica for urea adsorption, Materials Chemistry and Physics (2016),

W.-K. Cheah et al. / Materials Chemistry and Physics xxx (2016) 1e7

solution with known concentrations. The characteristic peak of urea lmax was taken at wavelength 200 nm in the UV region. Urea solution with concentration 38,000 mM was prepared by dissolving 231 mg of urea powder into 100 ml of distilled water. This value corresponds to the mean/median urea concentration of a patient suffering from uremia [6]. 150 mg of sample S was added into a conical flask containing the prepared urea solution. The conical flask was then shaken at 100 rpm using an orbital shaker (Lab Companion SK-300 Benchtop Shaker). 3 ml urea solution was sampled using a filter syringe at various time intervals. Sampled urea solutions were diluted with 6 ml deionized water to readjust the concentration to fit into the linear range of previously obtained calibration curve. The amount of adsorbed urea against time was plotted based on Equation (1), where Q ¼ amount of adsorbed urea (mg/g), Ci ¼ initial concentration (mg/L), Ct ¼ final concentration (mg/L), V ¼ volume of the solution and m ¼ mass of adsorbent (g).


 Ci  Cf V m


3. Results and discussion Prior to the urea adsorption test, samples C (dense silica), S (SBA-15), N (amine-functionalized SBA-15) and AC (commercial


activated carbon) were subjected to nitrogen adsorption analysis. Fig. 1 shows the adsorptionedesorption isotherm of C, S, N and AC. Commercial dense silica C exhibits Type III adsorption isotherm (IUPAC classification) since this sample is in dense form with no apparent porosity. Sample AC exhibits a Type I adsorption isotherm (IUPAC classification) thus revealing that this sample is microporous in nature. Sample S and sample N exhibit typical Type IV adsorption isotherm (IUPAC classification), indicated by the notable occurrence of capillary condensation. The observation on S and N isotherms indicates the formation of mesopores. Table 1 summarizes the surface area, DFT pore volume and mode pore diameter of samples C, S, N and AC. The high surface area of activated carbon makes it a generic choice as a universal adsorbent, such as in the application of wearable artificial kidney. Mesoporous silica samples S and N have a lower surface area in comparison. However, the pore volume of mesoporous silica S is much larger than that of AC, which is mainly contributed by the larger pore size of S. Hence, mesoporous silica could possibly adsorb higher amount of urea compared to activated carbon if all the pores were completely filled. The presence of amine functional groups on the surface of sample N reduces the effective surface area. The pore diameter of sample N is relatively smaller than that of bare SBA-15 sample S due to the presence of additional grafted amine groups on the external and internal pore surface. The regular pore structure of bare mesoporous silica collapses as more functional groups were introduced

Fig. 1. Nitrogen adsorption isotherm for (a) C, (b) S, (c) N and (d) AC.

Please cite this article in press as: W.-K. Cheah, et al., Amine-functionalized mesoporous silica for urea adsorption, Materials Chemistry and Physics (2016),


W.-K. Cheah et al. / Materials Chemistry and Physics xxx (2016) 1e7

Table 1 BET surface area, DFT pore volume, pore diameter, nominal rate constant (k), urea adsorption capacity (A) and normalized surface area adsorption capacity (M/a) of C, S, N and AC. Sample

BET surface area (m2/g)

DFT pore volume (cc/g)

Pore diameter (nm)

k (mg/g/s)

A (mg/g)

M/a (molecules/nm2)


158 488 576 0.5 a

0.283 0.800 0.320 e

5.2 5.7 0.52 e

0.6199 0.1924 0.1630 0.1252

542.6 498.6 348.8 109.6

34 10 6 *a


* M/a for C is disregarded due to the exceptionally low BET surface area.

on the inner pore surface [22]. For a better understanding of pore characteristics of sample S and N, the pore size distribution was plotted using the BJH (desorption) model, with reference to Fig. 2. Mesoporous silica with a narrow pore diameter distribution was obtained from the synthesis method reported in this paper. As a result, mesoporous silica of consistent pore characteristics could be controlled so long as the synthesis parameters are unchanged. The BJH pore size distribution for the mesoporous silica samples are consistent with the BET and DFT results obtained in Table 1, whereby the pore volume of sample N is lower than that of sample S. Similarly, as observed in Fig. 3, the pore diameter of N (5.6 nm) is smaller than that of S (6.5 nm). As mentioned, the presence of amine functional groups on the surface effectively reduces the pore diameter of sample N and thus the pore

volume as well. The primary pore diameter (mode) of AC, which is 0.52 nm, contributes to a large fraction of the pore volume and surface area. In comparison to the size of urea, 0.56 nm  0.63 nm  0.30 nm, the primary pores with a diameter of 0.52 nm would not be able to accommodate the urea molecules within its pores. Urea molecules can fit only into the secondary pores of AC with diameters 0.78 nm and 1.16 nm, which are lesser than primary pores with diameter 0.52 nm in contrast. Thus, the pore volume contributed by the mode pore diameter would be insignificant. The molecular size of urea is one of the smallest among uremic toxins, especially when compared with larger medium and proteinbound uremic toxins. If such is the case for urea with activated carbons, the mismatch would apply for the rest of uremic toxins.

Fig. 2. TEM micrographs of (a) sample S and (b) sample N.

Please cite this article in press as: W.-K. Cheah, et al., Amine-functionalized mesoporous silica for urea adsorption, Materials Chemistry and Physics (2016),

W.-K. Cheah et al. / Materials Chemistry and Physics xxx (2016) 1e7


Fig. 3. BJH (desorption) pore size distribution of samples S and N; Inset: pore size distribution of AC.

Carbon is not the most suitable adsorbent when taking pore diameter against adsorbate molecular size into consideration. Thus, the selection of activated carbon as the main adsorbent material in a wearable artificial kidney system should be promptly reevaluated. Fig. 2(a) and (b) shows the TEM micrographs of samples S and N respectively. The synthesized mesoporous silica samples possess elongated particles which look like tubes under the TEM. Highly ordered mesopores are observed for both S and N. A rough visual estimate of the particle size of the tube is 150 nm in diameter and 1200 nm in length. The darker parallel regions observed within the mesoporous silica particles are the pore walls while the brighter regions are the mesopores. Visually, the mesopores for sample S are measured and estimated to be 5.1 nm in diameter while that of sample N is slightly smaller, 4.0 nm in diameter. Despite this small difference, the formation of pores (brighter region) is thus confirmed since the values are in agreement with the pore diameter previously determined using nitrogen adsorption (BJH), which is 5.7 nm for sample S and 5.2 nm for sample N. The pores of sample N are less obvious than highly ordered pore structure of sample S. From visual comparison between S and N micrographs, the presence of amine subsequent to the functionalization process has perturbed the ordered structure of the mesoporous silica to some extent [22]. In order to confirm the presence of amine functional groups subsequent to the grafting process, FTIR spectra of samples before and after are compared. Fig. 4(a) shows the FTIR spectra of samples S and N. From the enlarged FTIR spectra in Fig. 4(b), weak NeH bonds are observed at 690, 1490 and 1563 cm1, which corresponds to NH bending, symmetry NH2 bending and NH2 deformation respectively. However, the NeH stretching which is usually observed at 3380 and 3310 cm1 is not observed in sample N. The broad peaks at 3200 to 3500 cm1, which correspond to OH bonds, might have covered the NeH peaks (3380 and 3310 cm1). The intensity of peaks corresponding to the OH group at 1233 cm1 is observed to reduce subsequent to the post-grafting process indicating the partial replacement of the OH group (silanol) by the amine functional group (APTES). Fig. 5 shows the results obtained from the urea adsorption test. The plot of reactant concentration versus time show the adsorbed urea versus time by all four C, S, N and AC samples are very similar to that of a typical adsorption curve. The individual curves were fitted using the exponential fit exp1 from MATLAB. The general mathematical form of urea adsorption curve could be expressed as in Equation (2):

Fig. 4. (a) FTIR spectra of sample S and N; (b) enlarged FTIR spectra of sample S and N.

Fig. 5. Urea adsorption kinetics of C, S, N and AC.

  y ¼ A 1  ekt


Whereby y corresponds to the adsorbed urea, coefficient k is the gradient of the curve during initial adsorption and A is the adsorption capacity (plateau). Table 1 summarizes the coefficients k and A obtained from the MATLAB exponential fit exp1. Coefficient k governs the rate at which adsorption occurs. Both nominal rate constant and adsorption capacity increase in the following order: C < AC < S < N. Often,

Please cite this article in press as: W.-K. Cheah, et al., Amine-functionalized mesoporous silica for urea adsorption, Materials Chemistry and Physics (2016),


W.-K. Cheah et al. / Materials Chemistry and Physics xxx (2016) 1e7

for adsorption applications, the engineering design for the adsorbent material is more focused on the adsorption capacity, since higher adsorbate uptake yields better performance. Emphasis should also be placed on the kinetics study and rate of adsorption of a material as well, instead of giving full attention to the adsorption capacity. The dialysate within the wearable artificial kidney model is circulated in a dynamic closed cycle flowing through the adsorbent packed in the form of cartridges. Mesoporous silica N possesses a much higher nominal rate constant (k) compared to both S and AC, thus making N an ideal adsorbent for urea removal within the dynamic circulating dialysate system of the wearable artificial kidney. The performance of a wearable artificial kidney could be improved by increasing the dialysate flow rate (as compared to single AC adsorbent system) in which more uremic toxin (urea) could be cleared over a specified predetermined duration. Design consideration to increase the dialysate flow rate for better uremic toxin clearance would only be possible with the employment of a material with high nominal rate constant and fast adsorption kinetics, i.e. amine-functionalized SBA-15 in this study. Theoretically, nanoporous materials with a higher surface area should have a higher adsorbate uptake. However, despite possessing higher surface area, the urea adsorption capacity of AC is only slightly higher than that of dense silica C and much lower than that of samples S and N due to the carbon's naturally low affinity towards the chemically unreactive urea [5]. As previously discussed, the pore size of AC is smaller than the molecular size of urea. Thus, the high surface area of AC could not be fully utilized as the accessibility to the micropores has been limited due to the relatively smaller pore size in comparison to urea molecule. Though the effective surface area S is thrice as high as that of N, sample N performed better due to the presence of amine group, which has a higher affinity towards urea as compared to silanol or eOH. In actual wearable artificial kidney application, the adoption of amine-functionalized SBA-15 could potentially improve the performance of such device by approximately 50% capacity compared to AC judging from the adsorption capacity (A). The performance of the adsorbent in a miniaturized artificial kidney is critical since it dictates the entire scale of miniaturization and also the frequency of adsorbent replacement. The urea adsorption capacity normalized by surface area, or number of molecules adsorbed on 1 nm2 area (M/a), is derived from the calculation of adsorbed urea on the adsorbent surface. Assuming that a 1 nm2 unit area could accommodate approximately 3 urea molecules (based on the size of urea molecule, largest surface area 0.56 nm  0.63 nm), we can assume that multilayer adsorption (M/a > 3 molecule/nm2) occurred for all 3 nanoporous materials. This result suggests that the presence of amine functional group in sample N has enabled the adsorption of more layers

Fig. 6. Proposed adsorption mechanism of urea on (a) sample S and (b) sample N.

of urea on both the external surface and within the mesopores. Fig. 6 shows proposed adsorption mechanism of urea on sample S and sample N. Each oxygen from the silanol could provide up to two adsorption sites through hydrogen bonding. However, due to the arrangement of silanol functional groups on the planar surface of mesoporous silica S, as shown in Fig. 6(a), not all of these sites from the silanol groups are accessible to direct chemisorption through hydrogen bonding. Though the adsorption capacities (A) of both sample S and sample N are quantitatively close, the number of molecules adsorbed per unit area (M/a) of N is higher than that of S (Table 1). With reference to Fig. 6(a), urea molecules which had already adsorbed on the surface of sample S would cause a form of steric hindrance, preventing free urea molecules from reaching the surface. Urea molecules forming direct hydrogen bonding with the surface functional groups are subjected to chemisorption. Subsequent urea molecules could only form physisorption with underlying layers of adsorbed urea molecules, resulting in the formation of multilayer urea adsorption. Based on M/a value in Table 1, multilayer urea adsorption occurred, though at lower number of layers compared to that of sample N. The propyl hydrocarbon chain on the functionalized mesoporous silica N creates a structure which encourages better packing of chemisorbed urea, as shown in Fig. 6(b). The steric hindrance due to planar surface (sample S) could be reduced by the elevation of amine functional group by propyl chain (sample N). Urea is able to form hydrogen bonding with both amine and silanol surface functional groups with less obstruction between the adsorbed urea molecules. Although silanol is known to possess a higher polarity compared to amine, the multilayer adsorption on sample N is higher. One plausible explanation is the presence of amine group on sample N, which has a similar amine functional group as urea, might have caused the crystallization of urea. During the crystallization of urea, the oxygen center of urea forms hydrogen bonding with two other NeH groups from two separate urea molecules, creating a network of urea crystal [21]. The amine group on sample N could have acted as the seed for urea crystal growth. A similar crystallization mechanism may have been initiated with silanol as a seed for sample S. The seeding effect of sample S might not be as efficient as that of sample N due to poor accessible to silanol seeding sites. As a whole, the architecture of the surface functional groups could ultimately determine the efficiency of an adsorbent material. Samples of S and N after urea adsorption process was recovered and subjected to FTIR analysis for a second time. Fig. 7(a) shows the FTIR spectra of sample S before and after urea adsorption, with S þ Urea denoting sample S after urea adsorption. The presence of urea can be observed at 1631 cm1, which corresponds to NH2 ddeformation. At 1457 cm1, the peak corresponding to NeCeN stretching mode is observed. The free NeCeN stretching suggests that the carbonyl group of urea might have interacted with the eOH group on the surface of S. The small peaks observed at 614 and 571 cm1 correspond to CeC stretching, which might belong to the surface functional group from the APTES source. Fig. 7(b) shows the FTIR spectra of sample N before and after (N þ Urea) urea adsorption. The spectrum of sample N þ Urea is very similar to that spectrum of sample S þ Urea. The peak at 1628 cm1 is due to NH2 deformation while the peak at 1455 cm1 is due to NeCeN stretching. The absence of peak corresponding to the NeH group in post adsorption N þ Urea might indicate that amine from urea could have interacted with the surface functional group, forming stronger chemical bonding instead of pure physisorption. The amine carbonyl reaction (similar to amine ketone reaction) forms strong bonds, which could explain the unusually higher urea adsorption capacity as exhibited by sample N. Though sample N has relatively lower effective surface area as compared to

Please cite this article in press as: W.-K. Cheah, et al., Amine-functionalized mesoporous silica for urea adsorption, Materials Chemistry and Physics (2016),

W.-K. Cheah et al. / Materials Chemistry and Physics xxx (2016) 1e7


Fig. 7. (a) FTIR spectrum of sample S before and after urea adsorption; (b) FTIR spectrum of sample N before and after urea adsorption.

samples S and AC, the presence of amine functional groups on the surface compensates the slightly reduced surface area with higher urea uptake for the same unit of weight. 4. Conclusion In this study, urea adsorption by mesoporous silica has been evaluated for the first time. Mesoporous silica SBA-15 and aminefunctionalized SBA-15 are found to be much superior compared to the present activated carbon based adsorbent in all evaluated aspects, i.e. urea adsorption capacity and urea adsorption rate. The implementation of such uremic toxin specific adsorbents to complement the present activated carbon would lead to immediate improvement in the overall design of a wearable artificial kidney. An increased dialysate flow rate (result of higher adsorption rate) and reduced dialysate amount (through increased adsorption capacity) could significantly reduce the size as well as improve both portability and uremic toxin removal performance of the existing wearable artificial kidney model. Acknowledgment Authors would like to thank the Ministry of Education Malaysia (Grant nos: 6730072, 6071244, 6071202), Ministry of Science, Technology and Innovation, Malaysia (MOSTI), MyBrain15 and AUN/SEED-Net for the financial support provided. References [1] W.H. Fissell, S. Roy, A. Davenport, Achieving more frequent and longer dialysis for the majority: wearable dialysis and implantable artificial kidney devices, Kidney Int. 2 (5) (2013) 632e666. [2] V. Gura, A.S. Macy, M. Beizai, C. Ezon, T.A. Golper, Technical breakthroughs in the wearable artificial kidney (WAK), Clin. J. Am. Soc. Nephrol. 4 (2009) 1441e1448. [3] M. Wester, F. Simonis, N. Lachkar, W.K. Wodzig, F.J. Meuwissen, J.P. Kooman, W.H. Boer, J.A. Joles, K.G. Gerritsen, Removal of urea in a wearable dialysis device: a reappraisal of electro-oxidation, Artif. Organs 38 (2014) 998e1006. [4] J.W. Agar, Review: understanding sorbent dialysis systems, Nephrology 15 (2010) 406e411. [5] W.K. Cheah, K. Ishikawa, R. Othman, F.Y. Yeoh, Nanoporous biomaterials for uremic toxin adsorption in artificial kidney systems: a review, J. Biomed. Mater. Res. Part B (2015), [6] V. Gura, C.J. Ezon, M. Beizai, Carbon dioxide gas removal from a fluid circuit of

a dialysis device, U.S. Patent No. 8,034,161, 2011. s, U. Baurmeister, P. Brunet, [7] R. Vanholder, R. De Smet, G. Glorieux, A. Argile W. Clark, G. Cohen, P.P. De Deyn, R. Deppisch, Review on uremic toxins: classification, concentration, and interindividual variability, Kidney Int. 63 (2003) 1934e1943. € C. Vagner, P. Boulet,  [8] D. BergE-Lefranc, H. Pizzala, J.-L. Paillaud, O. SchAf, B. Kuchta, R. Denoyel, Adsorption of small uremic toxin molecules on MFI type zeolites from aqueous solution, Adsorption 14 (2008) 377e387. €f, H. Ghobarkar, R. Denoyel, Adsorption properties of ze[9] V. Wernert, O. Scha olites for artificial kidney applications, Microporous Mesoporous Mater. 83 (1) (2005) 101e113. [10] D. Fine, A. Grattoni, R. Goodall, S.S. Bansal, C. Chiappini, S. Hosali, A.L. van de Ven, S. Srinivasan, X. Liu, B. Godin, L. Brousseau, I.K. Yazdi, J. FernandezMoure, E. Tasciotti, H.-J. Wu, Y. Hu, S. Klemm, M. Ferrari, Silicon micro- and nanofabrication for medicine, Adv. Healthc. Mater. 2 (2013) 632e666. [11] Q. Wei, H.-Q. Chen, Z.-R. Nie, Y.-L. Hao, Y.-L. Wang, Q.-Y. Li, J.-X. Zou, Preparation and characterization of vinyl-functionalized mesoporous SBA-15 silica by a direct synthesis method, Mater. Lett. 61 (2007) 1469e1473. [12] G. Wu, P. Li, H. Feng, X. Zhang, P.K. Chu, Engineering and functionalization of biomaterials via surface modification, J. Mater. Chem. B 3 (10) (2015) 2024e2042. [13] C. Urata, H. Yamada, R. Wakabayashi, Y. Aoyama, S. Hirosawa, S. Arai, S. Takeoka, Y. Yamauchi, K. Kuroda, Aqueous colloidal mesoporous nanoparticles with ethenylene-bridged silsesquioxane frameworks, J. Am. Chem. Soc. 133 (21) (2011) 8102e8105. [14] C. Kresge, M. Leonowicz, W. Roth, J. Vartuli, J. Beck, Ordered mesoporous molecular sieves synthesized by a liquid-crystal template mechanism, Nature 359 (1992) 710e712. [15] S. Inagaki, Y. Fukushima, K. Kuroda, Synthesis of highly ordered mesoporous materials from a layered polysilicate, J. Chem. Soc. Chem. Commun. (1993) 680e682. [16] D. Zhao, Q. Huo, J. Feng, B.F. Chmelka, G.D. Stucky, Nonionic triblock and star diblock copolymer and oligomeric surfactant syntheses of highly ordered, hydrothermally stable, mesoporous silica structures, J. Am. Chem. Soc. 120 (1998) 6024e6036. [17] Y.-q. Xu, G.-w. Zhou, C.-c. Wu, T.-d. Li, H.-b. Song, Improving adsorption and activation of the lipase immobilized in amino-functionalized ordered mesoporous SBA-15, Solid State Sci. 13 (2011) 867e874. [18] L. Zhang, C. Yu, W. Zhao, Z. Hua, H. Chen, L. Li, J. Shi, Preparation of multiamine-grafted mesoporous silicas and their application to heavy metal ions adsorption, J. Non-Cryst. Solids 353 (2007) 4055e4061. [19] F.-Y. Yeoh, A. Matsumoto, Y. Iwase, T. Baba, Characterization of acidic property of sulfo-group functionalized microporous and mesoporous silica by adsorption microcalorimetry, Catal. Today 132 (2008) 46e51. [20] A. Matsumoto, F.-Y. Yeoh, S. Fujihara, K. Tsutsumi, T. Baba, Acidic character and catalytic activity of sulpho-functionalized porous silica with an ordered structure, Adsorpt. Sci. Technol. 24 (2006) 451e459. [21] P.D. Godfrey, R.D. Brown, A.N. Hunter, The shape of urea, J. Mol. Struct. 413 (1997) 405e414. [22] F.-Y. Yeoh, A. Matsumoto, Y. Iwase, T. Baba, Preparation and characterization of sulpho-functionalized ordered mesoporous silica, Adsorpt. Sci. Technol. 26 (2008) 479e489.

Please cite this article in press as: W.-K. Cheah, et al., Amine-functionalized mesoporous silica for urea adsorption, Materials Chemistry and Physics (2016),