Aminosilane-Functionalized Cellulosic Polymer for Increased Carbon

ARTICLE pubs.acs.org/IECR

Aminosilane-Functionalized Cellulosic Polymer for Increased Carbon Dioxide Sorption Diana M. Pacheco, J.R. Johnson, and William J. Koros* School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive NW, Atlanta, Georgia 30332-0100, United States ABSTRACT: Improvement in the efficiency of CO2 separation from flue gases is a high-priority research area to reduce the total energy cost of carbon capture and sequestration technologies in coal-fired power plants. Efficient CO2 removal from flue gases by adsorption systems requires the design of novel sorbents capable of capturing, concentrating, and recovering CO2 on a cost-effective basis. This paper describes the preparation of an aminosilane-functionalized cellulosic polymer sorbent with enhanced CO2 sorption capacity and promising performance for use in postcombustion carbon capture via rapid temperature-swing adsorption systems. The introduction of aminosilane functionalities onto the backbone of cellulose acetate was achieved by the anhydrous grafting of N-(2aminoethyl)-3-aminoisobutyldimethylmethoxysilane. The dry sorption capacity of the modified cellulosic polymer reached 27 cc (STP) CO2/cc sorbent (1.01 mmol/g sorbent) at 1 atm and 39 cc (STP) CO2/cc sorbent (1.46 mmol/g sorbent) at 5 atm and 308 K. The amine loading achieved was 5.18 mmol amine(nitrogen)/g sorbent. Exposure to water vapor after the first dry sorption cycle increased the dry sorption capacity of the sorbent by 12% at 1 atm, suggesting its potential for rapid cyclic adsorption processes under humid feed conditions. The CO2 sorbent was characterized in terms of chemical composition, density changes, molecular structure, thermal stability, and surface morphology.

’ INTRODUCTION Carbon dioxide (CO2), a key byproduct of fossil fuel combustion processes, has become a major environmental concern due to the gradual increase of its global atmospheric concentration in recent decades. After water vapor, CO2 is the largest contributor to the greenhouse effect by transmitting solar radiation through the atmosphere but strongly absorbing the infrared and nearinfrared radiation reflected from Earth’s surface. One of the major point sources of CO2 emissions is coal-fired power plants; therefore, the control of CO2 emissions from these facilities is an urgent concern that has attracted significant attention from governments, energy companies, and scientific researchers. Several techniques have been proposed as approaches for CO2 separation and recovery, including chemical absorption and adsorption, physical adsorption, and membrane separation. Pressure swing adsorption and/or temperature swing adsorption systems consisting of packed beds of novel solid sorbents are emerging technologies with potential for carbon capture in coalfired power plants. These approaches are suitable for low CO2 concentration gas streams with large volumetric flow rates, and emerging systems using packed beds of hollow fiber hybrid sorbents have low flue gas pressure drop and low regeneration thermal requirements. These characteristics can significantly reduce the high costs of carbon capture and sequestration (CCS) technologies.1 However, the success of this and other emerging adsorption technologies relies upon the creation of novel sorbents with high CO2 sorption efficiencies and low manufacturing costs that allow favorable process economics and the minimization of significant CO2 capture costs in coalfired power plants. Conventional pellet sorbents such as zeolite molecular sieves, activated carbons, silica gels, and carbon molecular sieves have not been adequate for CO2 capture from r 2011 American Chemical Society

low concentration gaseous mixtures. Therefore, significant scientific research is in progress to produce suitable sorbents for different CO2 emitting sources applications using other traditional as well as novel materials. Efforts are underway ranging from new approaches for the use of well-known sorbents as zeolites1,2 and activated carbons3 in the design of complex hybrid materials to the exploration of totally new materials such as periodic mesoporous silicas4 and metal organic frameworks (MOFs).5 Despite the active efforts, development of efficient, reliable, thermally stable, and cost-effective sorbents for CO2 capture from power plant flue gases remains a challenge that needs attention. In order to reduce the cost of CO2 CSS, low cost sorbents with high CO2 sorption and long-term regeneration capacities in power-plant flue gas environments, with rapid adsorption and desorption rates and low energy requirements for regeneration, are needed. High thermal stability and resistance to humid and acid gas environments, as well as low mass transfer resistance, are also desirable sorbent properties to achieve economical postcombustion CO2 capture with environmentally benign performance. The use of such novel sorbents in periodic cyclic adsorption processes is a potentially low-cost carbon capture alternative technology that could meet the present and future economic and environmental constraints placed on CO2 capture from coal-fired power plants. The objective of the present research has been to develop a new strategy to design a cellulose acetate-based sorbent for CO2 capture. Cellulose acetate, whose molecular structure is Received: September 9, 2011 Accepted: November 10, 2011 Revised: November 2, 2011 Published: December 09, 2011 503

dx.doi.org/10.1021/ie2020685 | Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

Figure 1. Cellulose acetate’s molecular structure.

presented in Figure 1, has potential to evolve into a promising CO2 sorbent if properly treated in order to enhance its sorption properties. Cellulose acetate is an amorphous polymer characterized by its degree of substitution or degree of acetylation, the degree of polymerization, and the distribution of its hydroxyl groups. The degree of substitution refers to the average number of acetyl groups per glucoside unit. The degree of polymerization refers to the average number of glucoside units per polymer repeat unit. On the other hand, the distribution of unesterified hydroxyl groups among the primary (sixth position) and secondary (second and third) carbon positions is also important in cellulose acetate characterization.6 The acetyl groups in cellulose acetate reduce the amount of intermolecular hydrogen bonding present in cellulose, suppress its crystallinity, and increase its flexibility and mobility.7 The most common form of cellulose acetate has acetate groups on 2
2.5 of the 3 available positions for substitution and is typically called cellulose diacetate. Cellulose triacetate is a complete acetylated polymer in which at least 92% of the hydroxyl groups are acetylated.8 The presence of both acetyl and hydroxyl functionalities in cellulose acetate allows different types and degrees of intra- and intermolecular interactions that produce a polymer with useful properties. Among the main physical and chemical characteristics of cellulose acetate are the selective adsorption and removal of organic chemicals, its hydrophilicity, and its solubility in common organic solvents that allows the manufacture of fibers, membranes, or films. Commercial cellulose diacetate was chosen as the starting material for the development of the sorbent with enhanced CO2 sorption capacity considered here. This paper contains the essential elements of a chemical design strategy for transforming commercial cellulose diacetate into a useful aminosilane-functionalized cellulosic polymer sorbent with enhanced CO2 sorption capacity and promising performance for postcombustion carbon capture via rapid temperature-swing adsorption systems. The characterization of the sorbent in terms of chemical composition, density changes, molecular structure, thermal stability, and surface morphology are also presented. Transformation of the material into a fiber sorbent requires additional steps that are currently under study and will be considered in subsequent work; however, the current study demonstrates the attractiveness of this new type of sorbent material. Cellulose acetate has been used for adsorption of a variety of chemical species, including phenyl compounds and polynuclear aromatic hydrocarbons (PAHs) in membrane separations,9 N-methylcarbamates and lactose in ultrafiltration,10 and N-isopropylacrylamide and N-n-butyl-acrylamide in membrane-based biochemistry immunoassay,11 among many others.12 Its CO2 sorption properties and transport properties as CO2 separation membranes have been previously investigated and fully characterized.13 The results have shown great potential for this material to evolve into a promising CO2 sorbent if a proper chemical treatment is applied in order to enhance its sorption properties.

ARTICLE

’ BACKGROUND Organosilanes are compounds containing silicon bonded to a carbon atom. Aminosilanes are a particular type of organosilanes containing primary, secondary, and tertiary amines. The utilization of aminosilanes has been used extensively in surface modification, since they possess a hydrolytically sensitive center that reacts with hydroxyl groups from a surface to form stable covalent bonds. The organic substituent remains available for interaction with other species and allows permanent property modification of the substrate. The reaction with hydroxylated surfaces results in a substitution reaction at the silicon atom, with the formation of a silylated surface where the silicon is covalently bonded to the surface via an oxygen atom. The reaction can occur under hydrous or anhydrous conditions. Surface treatments with aminosilanes are influenced by the concentration of surface hydroxyl groups, the type of surface hydroxyl groups, the hydrolytic stability of the bond formed, and the substrate features, among others.14 Experiments have shown that solid sorbents with grafted amine functionalities are promising for CO2 sorption.2,4a
4c,4e
4h,15 Surface modification of cellulose acetate with aminosilanes can provide primary amino functionalities available for CO2 fixation. Therefore, the functionalization of cellulose acetate with aminosilanes was envisioned as a means for the improvement of its CO2 sorption properties. Grafting reactions are chemical procedures where one or more functionalities are connected to the main chain of a macromolecule as side chains, resulting in a polymer with different physical and chemical properties than the original. In this work, grafting of aminosilanes was used in order to introduce amine groups into the backbone of cellulose acetate and enhance its CO2 sorption properties. To the best of our knowledge, the method has not been tested yet on organic polymers like cellulose acetate for enhancement of its CO2 sorption properties. The chemical characteristics of the substrate, the amount of active hydroxyl groups available for aminosilane grafting, and the chemical properties of the silane modifier largely determine the course of the grafting. In addition, grafting conditions like water presence, temperature, aminosilane concentration, and reaction times are also critically important. Mixing of the aminosilane species with cellulose acetate in an adequate solvent allows the amine groups to catalyze the condensation of the silicon side of the molecule with C
OH groups. Silicone
oxygen
carbon bonds (Si
O
C) are formed with interaction of the polymer backbone in the absence of water for some aminosilanes or with an initial hydrolysis of the alkoxy groups for others. Therefore, the free hydroxyl groups are the main reactive centers for chemical bonding in the grafting reactions, whatever the reaction conditions may be. The larger the amount of free hydroxyl groups on the polymer backbone, the higher is the potential amount of aminosilane molecules grafted.16 Harlick et al. have proposed that the addition of a controlled amount of water prior to the grafting reactions can increase the amount of fixed amine functionalities on silica-based supports.4g The presence of water would increase the number of hydroxyl groups over the surface of the support material and start the hydrolysis of the unreacted aminosilane alkoxy groups with the free silanes still present in the solvent phase. Water may facilitate silane polymerization and generate a degree of silane condensation in the solution phase prior to the grafting reaction. 504

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

ARTICLE

This phenomenon may enhance the amount of certain types of aminosilanes grafted onto silica surfaces either directly or through the attachment of the silane polymers present in solution. In this regard, Harlick and co-workers suggest that some aminosilane molecules do not attach directly to the surface but rather to other grafted aminosilanes through Si
O
Si bridges. This phenomenon is acceptable as long as it increases the density of free amine functionalities available for CO2 capture and does not block the sites available for direct surface grafting. According to Engelhardt et al., it is necessary that some water be present at the surface of silica in order to form aminosilane layers using organic solvents.17 Caravajal et al. has showed that increasing the amount of water on a silica support surface decreases the amount of unreacted alkoxy moieties from the aminosilane molecule.18 A series of surface grafting experiments with different amounts of relative humidity and anhydrous support samples showed the elimination of unreacted alkoxy groups through the wet grafting procedure. They attributed the complete cleavage of the Si
O
R bond in aminosilanes to the presence of excess water. Therefore, different research groups have agreed on the importance of a desirable amount of surface water on silica-based supports to promote the cleavage of siloxyl bonds in the aminosilane molecule and allow further grafting to the support surface. However, as it is demonstrated in the present study, the presence of water in the grafting reactions of certain types of aminosilanes can actually prevent the fixation of amine functionalities onto the polymer chain. Among the additional grafting conditions that directly affect the amount of grafted amine groups on cellulose acetate are the amount of aminosilane species used and the temperature of the reactions. A series of amine grafting studies on mesoporous silica MCM-41 carried out by Harlick et al. showed that there is an optimum amount of aminosilane added to the reaction, and higher amounts did not increase the quantity of amine grafted or the CO2 sorption capacity of the sorbent.4g For the preparation of a smooth aminosilane layer over the support surface, Zhang et al. have found that a low aminosilane concentration and anhydrous solvents with a trace amount of water was desirable for silicon dioxide surfaces.19 On the other hand, the temperature of the reactions also plays an important role. It has been found that higher molecular mobility is favored by high temperature reactions in solution, overcoming hydrogen bonding interactions between amine groups and surface hydroxyl groups from silica surfaces.20 Vapor phase silanization has also been reported to produce a uniform aminosilane monolayer over silica surfaces.21 In our work, since the polymer matrix can swell and allow grafting beyond the strictly external surface, the situation is more complex than in strict surface attachment. Equilibrium sorption of a gas is usually described by an isotherm or the amount of sorbate on the sorbent as a function of its pressure at a constant temperature. The amount of sorbate is usually normalized by the mass or volume of the sorbent for convenience of comparison with other cases. For amorphous polymers like cellulose acetate above its glass transition temperature (189 °C), the equilibrium sorption of CO2 would follow the Henry’s Law sorption model given in eq 1: CD ¼ kD 3 p

However, amorphous polymers in the glassy state (below their glass transition temperatures) have a microheterogeneous molecular structure. These heterogeneities can be categorized in two components: the usual matrix and the microvoid region. Gas sorption into these different regions is therefore represented by different sorption mechanisms connected by local equilibrium. Henry’s Law sorption is the dominant mechanism in the equilibrium, well-packed regions of the matrix (CD), while Langmuir sorption describes uptake in microvoid domains in the glassy matrix related to nonequilibrium segmental packing (CH). The sum of both contributions is described by the dualmode sorption model (eq 2), where C is the total concentration of sorbed gas in the polymer, viz., 0

C ¼ CD þ CH ¼ kD p þ

CH bp 1 þ bp

ð2Þ

where C0 H is the Langmuir saturation or capacity constant and b is the Langmuir site affinity constant. The dual-mode sorption model provides a good description of the sorption process in most glassy polymers, like cellulose acetate.22 Puleo et al. have studied the gas transport and sorption behavior of CO2 in cellulose acetate polymers with different degrees of acetylation.13 Sorption isotherms measured at 35 °C fit the dual-mode sorption model and present a typical nonlinear concave shape characteristic of glassy polymers. Cellulose acetate films exposed to high CO2 pressures are plasticized or “conditioned”, affecting their sorption properties. Plasticization and swelling by CO2 exposure causes the rigid small microvoids between the polymer chains to redistribute. Sorption at extreme high gas pressures causes swelling of the polymer chains and a change in volume that may return to the original value only after very long times after removal of the swelling component. Therefore, swelling generates semipermanent changes in the cellulose acetate microstructure that result in the increment of the Langmuir capacity constant C0 H and the total sorption capacity. This is explained by the increase and redistribution of the sites where Langmuir sorption occurs. After CO2 desorption, the extremely slow relaxation time of the cellulose acetate prevents return to the polymer’s original state for long periods of time. In addition, the apparent Henry’s solubility coefficient kD increases after swelling due to changes in the packing efficiency of the polymer chains or to changes in the interaction with CO2 after conditioning. Therefore, the elevated uptake of CO2 in cellulose acetate makes it a highly sorptive penetrant.13 CO2 sorption into cellulose acetate chains is significantly affected by the acetyl content of the polymer. The higher the acetylation degree of cellulose acetate, the higher its intrinsic CO2 solubility and susceptibility to CO2 plasticization. Conditioning swells the polymer and modifies intermolecular interactions, increasing the mobility of functional groups and small-scale polymer chain motions. Cellulose diacetate shows higher increments in CO2 sorption capacity after swelling than is seen after swelling of low acetyl content films. Under equivalent CO2 exposure pressures, the CO2 sorption capacity of cellulose diacetate tends to increase with time due to slow relaxation processes in the glassy matrix.13 Intermolecular interactions of functional groups in cellulose acetate vary according to its degree of acetylation. In low-acetyl content cellulose acetates, the dominant effects are hydroxyl
hydroxyl group interactions due to hydrogen bonding. Hydroxyl
carbonyl interactions are weaker but still are of comparable

ð1Þ

where CD is the Henry’s Law concentration of CO2 sorbed in the polymer at equilibrium, kD is the Henry’s sorption coefficient, and p is the applied pressure of the system at equilibrium. 505

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

ARTICLE

Table 1. Typical Values of Relevant Properties and Molecular Structure of Aminosilanes Used

increased to 70 °C by placing it into a hot oil bath. Three milliliters of the corresponding aminosilane were injected, and the reaction was allowed to proceed during 24 h under a nitrogen atmosphere and constant temperature. After this time, the product was separated from the solvent using vacuum filtration and washed with 500 mL of anhydrous toluene and about 1 L of hexanes. The grafted samples were put into a vacuum oven at 70 °C overnight and stored afterward in capped vials until further use. Table 2 presents the various samples obtained by the anhydrous and wet grafting of aminosilanes on cellulose diacetate. To characterize the sorption of CO2 in the resulting sorbents as a function of CO2 pressure at a constant temperature, a pressure decay method developed for polymers was used.23 The system allows the determination of the sorption isotherms of sorbents, as well as the estimation of preliminary kinetic and equilibrium data. Accurate sorption kinetics measurements cannot be possible with this system since sorption generated heat effects are not taken into account. The time required for the sorbent to reach equilibrium can, however, be measured to provide an insight of the sorbent response time. The system is useful to test the sorption properties of the sorbents when exposed to pure gases only. Therefore, the effects of competitive adsorption and humidity in the feed stream could not be evaluated using this apparatus.1 Each sample was introduced and held inside a porous stainless steel filter and capped with aluminum foil. The foil cap was fastened in place using a tight stainless steel coil around it. The sorbent samples were introduced into a sample cell, which was tightly closed afterward. The closed sorption system was immersed into a constant temperature oil bath at 35 °C. After the sample was loaded, vacuum was pulled on the sample cell for one day to completely evacuate and dry the cell and the sorbent sample before the beginning of the test. After the drying step, CO2 was introduced to a reservoir cell at a controlled low pressure and allowed to equilibrate. After thermal equilibrium was achieved, the sample cell valve was opened rapidly to introduce the CO2 into the sample cell. As the stainless steel filter is porous, the gas easily reached the sample. The pressure decay over time was recorded until equilibrium was attained. When the sample reached sorption saturation at a particular pressure, the data recording was stopped and an additional amount of CO2 was introduced into the reservoir cell. The same procedure was repeated for this new pressure value. The recorded data allow the determination of the sorption isotherms

strength to hydrogen bonding. Acetyl
acetyl interactions are the weakest ones and the most common in cellulose diacetates. These interactions are more susceptible to being disrupted by a polarizable sorbate like CO2.13 Various IR studies have demonstrated that exposure to CO2 produces a shift in the carbonyl peak of the spectra, attributed to disruptions in the dipolar interaction between acetyl groups. In addition, the presence of CO2 in cellulose acetate also affects intermolecular and intramolecular hydrogen bonding interactions, so CO2 sorption can be promoted by the free hydroxyl functionalities.

’ MATERIALS AND METHODOLOGY All materials were used as received from the different suppliers. Cellulose acetate (CA) (MW 50 000 GPC, Sigma-Aldrich, St. Louis, MO) with 39.8 wt % acetyl content and 3.5 wt % hydroxyl content was used as the starting material for the improved CO2 sorbent. Toluene anhydrous (99.8%, SigmaAldrich) was used both as solvent in the grafting reactions and for rinsing purposes. Hexanes (mixture of isomers anhydrous, g99%, Sigma-Aldrich) were also used in the washing procedures. N-(2-aminoethyl)-3-Aminoisobutyldimethylmethoxysilane and aminopropyldimethyl-ethoxysilane (Gelest Inc., Morrisville, PA) were chosen as chain modifiers in the grafting reactions. For practical reasons, we will refer to the aforementioned aminosilanes by their molecular formulas, that is, C9H24N2OSi and C7H19NOSi. Typical values of relevant properties and the molecular structure of both aminosilanes are shown in Table 1. In order to introduce amine groups into the polymer backbone of pure cellulose acetate, the conventional wet and dry grafting techniques with aminosilanes were used as experimental approaches. In each run, 1 g of commercial, cellulose diacetate was dried overnight at 70 °C in a vacuum oven. All glassware used was dried overnight in a convection oven at 130 °C and heated externally with a propane torch right before its use. A 500 mL three-neck flask with a magnetic stirrer was connected to a condenser and, after introduction of the sample, the system was closed. A stream of dry nitrogen was run through the reaction system for 1 h. After this time, 150 mL of anhydrous toluene were injected into the system using a syringe. The powder and the solvent were allowed to mix for 30 min at room temperature under a nitrogen atmosphere. For wet grafting reactions, about 0.25 mL of water was incorporated at this point and the mixture was allowed to equilibrate for 3 h while mixing at room temperature with nitrogen flowing. For anhydrous grafting, the previous step was skipped and the temperature of the reaction solution was 506

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

ARTICLE

Table 2. Reaction Conditions for Samples of Cellulose Diacetate Subjected to Dry and Wet Aminosilane Grafting sample

cellulose acetate (g)

toluene (mL)

A

1

150

B

1

150

C

1

D

1

aminosilane species

aminosilane (mL)

H2O (mL)

reaction time (h)

reaction temperature (°C)

C9H24N2OSi

3

0.00

24

70

C9H24N2OSi

3

0.25

24

70

150

C7H19NOSi

3

0.00

24

70

150

C7H19NOSi

3

0.25

24

70

Table 3. EA of Pure and Aminated Cellulose Acetate Powder Samples through Dry and Wet Grafting Reactions elemental analysis sample

a

%C

%H

%O

%N

% Si 0.00

G

cellulose diacetate (CA)

48.21

5.93

45.60

0.00

A

CA powder anhydrous grafting with C9H24N2OSi

49.87

8.65

29.26

7.25

4.31

B

CA powder humid grafting with C9H24N2OSi

49.13

6.31

40.22a

2.53

1.81

C

CA powder anhydrous grafting with C7H19NOSi

48.66

6.72

40.35

2.01

1.98

D

CA powder humid grafting with C7H19NOSi

48.61

6.91

37.86a

3.09

3.53

Due to limitations in the sensitivity of the instrument, oxygen content could not be accurately measured; therefore, a value was calculated.

Table 4. Densities of Cellulose Diacetate and Aminated Cellulose Acetate

and preliminary sorption kinetics. The design of the system complicates the temperature control of the instantaneous local thermal heat effects inside the sample cell; however, unambiguous equilibrium results can be obtained easily by waiting for stabilization of cell pressures. Desorption was carried out by exposing the samples to temperatures of about 130 °C under vacuum for 24 h. The composition and properties of the modified cellulose acetate samples was characterized by elemental analysis (EA), density analysis, attenuated total reflection infrared spectroscopy (ATR-IR), thermogravimetric analysis (TGA), and scanning electron microscopy (SEM). Elemental analysis (Galbraith Laboratories Inc., Knoxville, TN) was used to determine the percentage weights of the main elements present in each sample and determine the amount of amine functionalities grafted to the polymer chains of cellulose acetate. Absolute density measurements (Micromeritics Analytical Services, Norcross, GA) were also performed in order to determine the change in mass per unit volume of the modified material and accurately evaluate their CO2 sorption capacity. Attenuated total reflection infrared spectroscopy tests (Shimadzu Scientific Instruments, Columbia, MD) were used to confirm the presence of surface amine functionalities grafted to cellulose acetate and to reveal details about the molecular structure of the new materials. Thermogravimetric analysis (Bruker Corporation, Billerica, MA) was also used to determine the thermal stability and adsorbed moisture content and to confirm the change in molecular structure of the modified cellulose acetate. In addition, scanning electron microscopy analyses (Leo Electron Microscopy, Cambridge, UK) were carried out with the aim to image the sample surface and investigate their changes in morphology.

sample

density (g/mL)

G

acetone-soluble cellulose acetate powder (CA)

1.4205

A

CA powder anhydrous grafting with C9H24N2OSi

1.1951

B

CA powder humid grafting with C9H24N2OSi

1.2931

C

CA powder anhydrous grafting with C7H19NOSi

1.3134

D

CA powder humid grafting with C7H19NOSi

1.2548

cellulose acetate powder under anhydrous conditions registered the highest loading of amine functionalities, with 7.25 wt % nitrogen. This amount is equivalent to grafting about 5.18 mmol amine(nitrogen)/g sample A. Estimated theoretical values of molecular weight composition shows that the grafting of one molecule of C9H24N2OSi to one hydroxyl group per glucoside unit of the polymer yields a nitrogen amount of 6.70 wt %. As the amount of nitrogen obtained from EA results is a bit larger, it is suggested that at least one molecule of C9H24N2OSi was grafted to each glucoside unit in cellulose acetate, and for some repeating units, more than one was actually attached. Clearly, by working with a lower degree of substitution CA or by decarboxylating the sample to make more hydroxyl units available, one could increase the amine loading by almost a factor of 2.5 in principle. The addition of a controlled amount of water to the grafting reaction drastically decreased the loading of amine functionalities and the sorption capacity of sample B. On the other hand, the sample with the least amount of grafted amines is C, corresponding to the anhydrous grafting of C7H19NOSi on cellulose diacetate powder. The amount of nitrogen in this sample was reduced to 2.01 wt % or about 1.44 mmol amine(nitrogen)/g sample C. Surprisingly, while the addition of water decreased the amount of amine fixation for the grafting reactions with C9H24N2OSi, humid conditions had the opposite effect when working with C7H19NOSi. The wet grafting of C7H19NOSi on cellulose acetate increased the amount of amine loadings and its CO2 sorption capacity. The reason for this dramatic difference in response to wet verus dry grafting for the two silanes will be discussed in the later analysis.

’ RESULTS AND DISCUSSION Elemental Analysis (EA). EA is a useful tool to analyze the effect that water and each type of aminosilane have on the amount of amines loaded into cellulose diacetate. Results from this characterization are provided in Table 3 for the EA of cellulose acetate before and after dry and humid grafting reactions with aminosilanes. The grafting of C9H24N2OSi on 507

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

ARTICLE

Figure 3. ATR-IR spectra of neat CA powder (black line) and humid grafting of CA with C9H24N2OSi, sample B (gray line).

Figure 2. ATR-IR spectra of neat CA powder (black line) and anhydrous grafting of CA with C9H24N2OSi, sample A (gray line).

Density Analysis. Densities of pure polymer and aminated samples are shown in Table 4. It was found that the densities of the cellulose acetate grafted samples decreased with increasing amines content. The bulky aminosilane molecules grafted to the cellulose acetate backbone in substitution of hydroxyl groups decrease the packing of the polymer chains and the total weight per unit volume of sample. Therefore, the higher the amount of grafted aminosilane functionalities, the larger the expected decrease in absolute density of a sample. In agreement with this expectation, sample A with the highest amine content presented the largest decrease in absolute density of about 15.8%. The sample with the lowest amount of grafted amines was sample C, which showed the smallest decrease in its density value (∼7.5% decrease). These results confirm the significant change in the morphology of the polymer chains due to aminosilane grafting. Attenuated Total Reflection Infrared Spectroscopy. Infrared spectra were obtained for commercial cellulose diacetate powder and for aminosilane-grafted cellulose acetate powder samples. The analyses were useful to investigate the main structural changes in the modified cellulose acetate molecules with respect to the pure polymer. The infrared analysis of pure cellulose acetate (sample G) is presented in Figure 2. Such spectra have been widely studied, and a complete interpretation of this spectrum has already been published.24 The weak absorption band at 900 cm
1 corresponds to amorphous regions in the polymer, while the strong doublet peaks at 1040 cm
1 and 1230 cm
1 are characteristic vibrations of acetate groups. At 1370 cm
1, a
CH3 deformation accounts for a peak of medium intensity, and at 1740 cm
1, the carbonyl CdO region shows an intense absorption band. The hydroxyl region (3400
3645 cm
1) displays a broad and weak absorbance band, corresponding to the free hydroxyl groups dangling from the polymer backbone of cellulose diacetate, as well as those involved in intra- and intermolecular hydrogen bonding. Grafting reactions are designed to introduce new functionalities into a macromolecule while conserving the basic molecular structure. The infrared spectrum of sample A, also displayed in Figure 2, shows that the molecular structure of the cellulose acetate backbone was well conserved after the grafting reaction. The Si
CH3 group has been reported to be easily recognizable by a strong, sharp band at about 1250 cm
1, together with one or more strong absorption bands in the range of 750
865 cm
1.25 Both characteristic absorptions bands are clearly visible in the spectrum of sample A, confirming the grafting of polysiloxane

Figure 4. ATR-IR spectra of neat CA powder (black line) and anhydrous grafting of CA with C7H19NOSi, sample C (gray line).

Figure 5. ATR-IR spectra of neat CA powder (black line) and humid grafting of CA with C7H19NOSi, sample D (gray line).

functionalities into the polymer backbone. In addition, the functionality [(CH3)2SiO]x, present in the molecule of C9H24N2OSi, is characterized by the strong band at 1020 cm
1, also visible in the present spectrum. The band of medium intensity at ∼1400 cm
1 corresponds to aliphatic amines, while the doublet of bands at 1550 and 1650 cm
1 are widely assigned to deformation vibrations of secondary amines.26 Kanan et al. have assigned 508

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

ARTICLE

Figure 6. SEM images of samples A, B, C, and D.

the 1650 cm
1 absorption band to free amine functionalities.27 The appearance of the broad band around 2800
2970 cm
1 may correspond to stretching C
H vibrations. Finally, the range from 3300 to 3550 cm
1 has been assigned to free amine functionalities.28 Therefore, it can be concluded that the anhydrous grafting of C9H24N2OSi worked very well under the reaction conditions applied and the present spectrum can be used as the reference for the characteristics of the desired sorbent. The spectrum of sample B, C9H24N2OSi-grafted cellulose acetate under wet conditions, is shown in Figure 3. This spectrum confirms the grafting of a smaller amount of aminosilane functionalities. The couple of absorption bands around 750
865 cm
1 and ∼1020 cm
1, characteristic of polisiloxanes, are weaker in comparison with the rest of this spectrum, as well as the bands at 1550 and 1650 cm
1, corresponding to amine vibrations. The decrease in free amine functionalities is more evident in the region of 3400
3500 cm
1. The comparison of the weaker absorption bands typical of organosilanes and amine functionalities with the rest of the spectrum suggest a smaller amount of C9H24N2OSi grafted in sample B under wet conditions. The same conclusions can be reached for the spectrum of sample C, grafted under anhydrous conditions with C7H19NOSi and presented in Figure 4. However, an increase in absorption intensities is registered by sample D, C7H19NOSi grafted under humid conditions. Its spectrum is presented in Figure 5. The absorption bands at 750
865 cm
1 indicate a higher presence of Si
CH3 bonds in the grafted molecule, and the strong intensities of the 1550 cm
1 and 1650 cm
1 peaks when compared to the rest of the spectrum suggest an increase in the amines grafting. Scanning Electron Microscopy. Scanning electron micrographs of the aminosilane grafted samples were performed in

order to observe if there were any significant changes in morphology. The images of aminosilane-grafted samples are presented in Figure 6 (as-received cellulose diacetate not pictured). SEM images for cellulose diacetate powder grafted with anhydrous or wet C9H24N2OSi (samples A and B) and anhydrous or wet C7H19NOSi (samples C and D) revealed a wide particle size distribution, which are similar to the as-received polymer. Surface morphologies of the particles are very heterogeneous and rough. This may be the results of multiple reaction steps affecting the original particle size and homogeneity of the commercial polymer. Regardless, we can clearly see that none of the reaction methods have any considerable change to the polymer morphology. Gas Sorption. Figure 7 provides the CO2 sorption isotherms obtained for samples A
D. Sample A, cellulose diacetate powder grafted with C9H24N2OSi under anhydrous conditions, presented a significant improvement in its CO2 sorption capacity after the grafting reaction. At an equilibrium pressure of about 1 atm, sample A was capable of adsorbing about 27 cc (STP) CO2/ cc sorbent, while pure cellulose acetate captured about 6 cc (STP) CO2/cc CA at very similar equilibration pressure. This represents an increment in the sorption capacity of about 4.5 times the sorption capacity of the unmodified polymer at low pressures. The increment in sorption capacity is equivalent to the capture of additional 0.78 mmol CO2 per gram of sample A or additional 0.15 mol CO2 per mole of amines grafted. At higher pressures of about 5 atm, sample A increased its sorption capacity up to 39.2 cc (STP) CO2/cc sorbent, capturing an additional amount of 0.84 mmol CO2 per gram of sorbent. In addition, its sorption equilibration time during the first dry sorption cycle is about 2 h, very similar to that of pure cellulose acetate. At 1 atm, 509

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

ARTICLE

Figure 9. TGA curves of neat CA (G) and anhydrous (A) and wet (B) grafting of CA with C9H24N2OSi and postreaction water exposure of sample A (F). Figure 7. CO2 adsorption isotherms for samples A (9), B (O), C (1), D (3), and G (b). All isotherms represent data measured at 35 °C.

This effect requires further study and is beyond the scope of the current exploratory work to identify preferred grafting agents and conditions. Sample A, which presented the highest increment in CO2 sorption capacity, was subjected to complete desorption of the CO2 captured from its first exposure by increasing its temperature to about 120 °C under vacuum for 24 h. Half of the desorbed sample was subjected to a second CO2 sorption cycle right after its desorption and was labeled as sample E. The other half of sample A was exposed to water vapor for 3 days and labeled as sample F. This was performed by spreading the sample over a clean and dry Petri dish and introducing it into the top of a closed glass chamber with 500 mL of deionized water at its bottom. The water in the chamber remained heated to a temperature of about 50 °C during the experiment. After this time, hydrated sample F was removed and put into a vacuum oven at 70 °C overnight. Afterward, it was loaded into the pressure decay sorption system in order to test the sorption capacity on a second sorption cycle after its exposure to water vapor. Figure 8 presents the isotherm for the first exposure to CO2 of sample A already presented above, the isotherm for sample E recorded at its second CO2 sorption cycle, and the isotherm for sample F on a second sorption cycle after hydration with water vapor, followed by mild drying in a vacuum oven at 70 °C. Sample E, obtained after a 24 h-desorption process at 120 °C applied to sample A, presented a CO2 equilibrium sorption capacity of 26 cc (STP) CO2/cc sorbent at an equilibrium pressure of about 1 atm, slightly below the equilibrium sorption capacity of sample A at similar pressure (about 27 cc (STP) CO2/cc sample A). This means that, in its second dry sorption cycle, sample A’s sorption capacity decreased by about 4% at low pressures. However, when desorbed sample A was exposed to water vapor for about 3 days, subsequently dried at 70 °C overnight, and loaded again into the sorption system, sample F presented an increase of about 11% on the sorption capacity with respect to its first sorption cycle. That is, the new sorption capacity of sample F is about 30 cc (STP) CO2/cc sorbent at 1 atm and 44.5 cc (STP) CO2/cc sorbent at 5 atm. It was able to capture an additional 0.11 mmol CO2 per gram of sample F at low pressures and about 0.20 mmol CO2 per gram of sample F at 5 atm. On the other hand, it was observed that sample E had an

Figure 8. CO2 adsorption isotherms for samples A (3), E (O), F (1), and G (b). All isotherms represent data measured at 35 °C.

sample B grafted with C9H24N2OSi under humid conditions was able to sorb about 11 cc (STP) CO2/cc sample B, an increment of about 1.8 times the sorption capacity of the pure polymer. This means that sample B adsorbed additionally 0.17 mmol CO2 per gram of sample at low pressures. On the other hand, both samples C and D, cellulose acetate grafted with C7H19NOSi under anhydrous and humid conditions, respectively, decreased its CO2 sorption capacity when compared to pure cellulose acetate. Perhaps the most surprising result is the lower CO2 sorption in sample D, since a clear grafting has occurred on the basis of the FTIR and EA results. At this point, we do not have an unambiguous explanation for the result; however, it is in marked contrast to the expected and observed results for sample A, which has a high degree of grafting of the silane with two functional groups. The effect is clearly complex, since the CO2 sorption level is actually suppressed even below the ungrafted control. 510

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

ARTICLE

the sample lost its thermal stability and registered a degradation step where it lost about 80% its weight. Therefore, the thermal stability of samples A, B, and F is similar and slightly reduced from that of cellulose acetate. For comparison, Figure 10 presents the TG curves of cellulose acetate samples grafted with C7H19NOSi under dry and humid conditions. The anhydrous grafting reaction registered a weight loss of 80% at a degradation temperature of 170
180 °C and about 15% of the sample remained as char residue at the end of the test. The cellulose acetate sample grafted under humid conditions showed a smaller decrease in its thermal stability, starting to degrade also at about 170
180 °C with a weight loss of about 76%. The rate of decomposition was slightly faster than the dry grafted sample and yielded residual char of about 21%. Overall, the thermal stability of all amine-grafted samples was affected to a similar extent by the introduction of aminosilane molecules in its molecular structure. Figure 10. TGA curves of neat CA (G) and anhydrous (C) and wet (D) grafting of CA with C7H19NOSi.

average equilibration time of 4 h, while the average time for sample F was about 15 h. It is important to note that these long equilibration times may have been caused by particle agglomeration, since after water exposure the average particle size appeared to be roughly as twice as big as the initial powders. Therefore, further systematic kinetic studies using films are needed in order compare the kinetics of samples with the same characteristic dimensions. Thermogravimetric Analysis. The thermal stabilities of pure and modified cellulose acetate sorbents were also investigated by thermogravimetry. TGA was performed for samples A
D, F, and G to determine their change in weight with respect to changes in temperature. A two-step dynamic temperature program was run in order to analyze the thermal stability and adsorbed moisture content of the samples. The measurements were performed using a heating rate of 10 K/min. The first step consisted in raising the temperature to 100 °C for about 60
70 min. Then, the furnace temperature was elevated to 750 °C, holding it for about 30 min. Helium was used as purge and protective gas at a rate of 30 mL/min. The thermogravimetric (TG) curves for aminosilane grafted cellulose acetate powder with C9H24N2OSi under anhydrous and humid conditions are shown in Figure 9. The thermal stability of both amine grafted samples decreased after the grafting reaction. Sample A, which presents the highest improvement in CO2 sorption capacity, decreased its thermal stability from 290 to 300 °C to about 170
180 °C, as evidenced by onset of weight loss. The mass loss during this degradation step was about 83%, with a slower rate of thermal decomposition than commercial cellulose acetate. After 140 min and at a temperature of about 760 °C, the residual char amounted to about 15% of the original sample weight. Similarly, cellulose acetate powder grafted with the same aminosilane species but under humid conditions (sample B) presented a decrease in thermal stability to about 170
180 °C. The weight loss step was about 72%, and the amount of char residue was 25% of the original weight. The TG curve for sample F, CO2-desorbed sample A after one sorption cycle and exposed to water for 3 days, was introduced into the TGA instrument right after its removal from water exposure. It registered a weight loss of about 7% its initial weight of 85 mg; that is, a mass of around 5.95 mg of water was removed from the sample at the beginning of the program. Around 170
180 °C,

’ DISCUSSION High-acetyl content, acetone-soluble cellulose acetate was subjected to a grafting reaction with two different types of aminosilanes, in the presence and absence of water. A diaminosilane (C9H24N2OSi) with a single methoxy group and a monoaminosilane (C7H19NOSi) with a single ethoxy group were reacted and characterized under the same reaction conditions in order to allow the comparison of amine loadings and improvement of CO2 sorption capacity. The presence of two amine functionalities in C9H24N2OSi was useful to investigate the effect of multiple amine groups per aminosilane molecule on the cellulose acetate sorption capacity and overall stability. The existence of a single methoxy and ethoxy group in each molecule, respectively, should facilitate the controlled grafting of aminosilane functionalities. In addition, the grafting reactions were run under dry and humid conditions in order to determine the effect of water on the successful aminosilane grafting of the support. The type and amount of solvent (anhydrous toluene) and the dose of aminosilane species employed were based on literature values. The temperature of the reaction (70 °C), the amount of water used in the wet grafting reactions, the drying procedures, and inert conditions applied in the anhydrous grafting reactions were selected from preliminary screening work in this research to indentify the most attractive conditions to study. The rigorous dry and inert conditions during the whole anhydrous grafting reactions were applied in order ensure the absence of any water other than that intentionally introduced. A long reaction time (24 h) was used in order to allow good diffusion of the reactants through the polymer chains. The solvent rinsing procedure was based on findings from the literature suggesting the considerable impact on the structure of grafted aminosilanes when water is used as washing species. It has been suggested that rinsing water can completely remove hydrogen bonded aminosilanes weakly attached to the supports before the drying step. Moreover, the drying procedures have also been found to be critical for ensuring covalent bond formation by condensation of hydrogen bonded silanol groups.20 The described conditions were applied to as received commercial cellulose acetate with a degree of substitution of about 2.5, in which acetate groups occupy an average of 2.5 out of 3 available positions for acetyl substitution in each glucoside unit. The remaining positions are occupied by hydroxyl groups, averaging 3.5 wt % of the polymer. Elemental analysis 511

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

ARTICLE

toward hydroxyl groups is conserved as long as one alkoxy functionality remains free. However, its use for improvement of CO2 sorption capacity in cellulose acetate is not recommended, since the formation of aminosilane multilayers grafted to the polymer chains can lead to increased amine hydrogen bonding and slow CO2 diffusion through the aminosilane layers, thereby suppressing sorption due to steric hindrance effects. Water presence in the reaction and subsequent polymerization of aminosilanes in solution can cause the formation of a number of possible grafted structures, including both covalent and hydrogen bonds. Covalently attached aminosilane molecules can be positioned in various orientations and local environments due to the various positions of hydroxyl groups along the chains. When the interactions between cellulose acetate hydroxyl groups and aminosilane molecules are due to hydrogen bonds only, the aminosilane molecule is weakly attached to the surface and can easily be lost. In addition, the already grafted amine functionalities can hydrogen bond to another hydroxyl or be attracted to oxygen atoms in the cellulose acetate surface by dipole intermolecular forces. These types of intricate networks can also presumably hinder CO2 diffusion and sorption onto free amine functionalities.20 These findings highlight the importance of understanding the chemical behavior of the selected aminosilane species for the grafting reactions in the presence or absence of water. Several literature reports using triethoxy and trimethoxy-aminosilanes remark on the benefits of the presence of water during aminosilane grafting reactions on silica-based supports.4d,g,15a,29 In these reactions, not all the three alkoxy ligands from the aminosilane species completely react with the surface hydroxyl groups from the silica-based support, so the presence of water molecules helps to consume the unreacted alkoxy ligands and complete the surface coverage. It has been suggested that water increases the surface density of hydroxyl groups on the supports or initiates the hydrolysis of unreacted alkoxy groups with free aminosilanes still present in solution, starting a polymerization process that enhances the surface coverage when free alkoxy functionalities from these polymers finally get linked to the silica surface. However, in reactions where aminosilanes with single methoxy or ethoxy functionalities are used as grafting agents, the polymerization effect that may result from the presence of water in the reaction consumes the only reactive alkoxy functionality and prevents further useful grafting of the aminosilanes polymers into the substrate. Sample A, cellulose diacetate powder grafted with C9H24N2OSi under anhydrous conditions, achieved the highest amine loading and the highest improvement in CO2 sorption capacity. The amine loading achieved in this sample was 5.18 mmol amine(nitrogen)/g sample A. Published work on amine-grafted silica-based sorbents prepared under anhydrous conditions and applied for CO2 sorption have reported values of amine content in the range of 1.20
2.57 mmol amine(nitrogen)/g final material for monoamino-trialkoxysilanes, 3.07
3.76 mmol amine(nitrogen)/g final material for diamino-trialkoxysilanes, and 3.86
6.11 mmol amine(nitrogen)/g final material for triamino-trialkoxysilanes.4a
c,g,15,30 Water-aided amine-grafted silica CO2 sorbents prepared with monoamino-trialkoxysilanes have reached amine contents in the range of 2.61
3.99 mmol amine(nitrogen)/g final material. When a diamino-trialkoxysilane has been used in humid grafting procedures, a range of 2.64
4.61 mmol amine(nitrogen)/g final material has been reported, and the use of triamino-trialkoxysilanes has yielded amine contents of 5.80
7.98 mmol amine(nitrogen)/g final material.4d,g,15a,29 Therefore, the amine loading achieved by sample A surpassed

Scheme 1. Polymerization of Aminosilanes in the Presence of Water

results demonstrate that, when working at anhydrous conditions, diaminosilane C9H24N2OSi accommodated a higher amount of amine functionalities into the polymer backbone than the monoaminosilane used. However, working under humid conditions, C7H19NOSi grafted a higher amount of amine groups. This can be explained by the fact that under anhydrous conditions the only available reactive hydroxyl groups are those dangling from the polymer backbone. Therefore, the reactivity of the alkoxy functionalities from the aminosilane molecule greatly influences the success of the reaction. According to Gelest, ethoxy silanes are essentially nonreactive in water-free environments and require catalysis for suitable reactivity.14 The amine functionality in aminosilanes acts as a catalyst and enhances its reactivity in grafting reactions. The influence of the amine group is such that the reactivity of hydroxylated species with organo-functional silanes decreases in the order Si
NR2 > Si
Cl > Si
NH
Si > Si
O2CCH3 > Si
OCH3 > Si
OCH2CH3. However, the presence of only one amine group per C7H19NOSi molecule decreases its overall amine loading. The amine functionalities can hydrogen bond with the hydroxyl groups from the polymer chain or with other amino groups, decreasing its availability for interaction with CO2. The formation of a stable five-membered ring intermediate is believed to be a mode of amine catalysis. Aminosilanes with long chains of alkyl linkers can favor the formation of this cyclic structure between NH2 end groups and hydrogen and oxygen centers.20,27 When this type of cyclic structures occurs, the number of available amine groups for CO2 sorption can decrease considerably. On the other hand, Gelest also suggests that only methoxysilanes are capable of reacting under anhydrous conditions.14 Therefore, the highly reactive methoxy functionality in C9H24N2OSi and the presence of two amine functionalities per aminosilane molecule are responsible for the high amine loading achieved by C9H24N2OSi in the anhydrous grafting reactions. On the contrary, the presence of water in the reaction catalyzes the reactivity of the ethoxy functionality in C7H19NOSi and increases it fixation into the polymer matrix. However, the opposite effect occurs when working with C9H24N2OSi under wet conditions, since the higher reactivity of the methoxy functionality increases its sensitivity to polymerization. Under humid conditions, the alkoxy groups of the aminosilane are hydrolyzed to silanols (Si
OH) that can combine to form a siloxane bond (Si
O
Si) between two aminosilane molecules in solution, with the production of a water molecule.16 This phenomenon is illustrated in the reactions presented in Scheme 1. Therefore, the presence of water can cause uncontrolled polymerization and/or oligomerization in the grafting reaction solution. In this case, aminosilanes like C9H24N2OSi and C7H19NOSi, containing only one alkoxy functionality per molecule, lose all their reactivity toward cellulose acetate hydroxyl functionalities after polymerization, remaining in solution and being washed out at the end of the grafting reaction. Aminosilanes containing multiple alkoxy groups per molecule are more likely to polymerize in the presence of water, but its reactivity 512

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research by about 38% the highest amine content reported for silica-based sorbents grafted with diamino-trialkoxysilanes under anhydrous conditions. Likewise, it outranged by 12% the highest amine loading reported in the literature for wet grafting of diaminotrialkoxysilanes on silica-based supports. In addition, the amount of loaded amines on sample A is in good correspondence with the theoretical value of amine content resulting from the graft of one molecule of C9H24N2OSi per glucoside unit (∼6.70 wt %). At a pressure of about 1 atm, sample A was capable of sorbing about 27 cc (STP) CO2/cc sorbent or about 1.01 mmol CO2/g sorbent. At about 5 atm, sample A was able to sorb about 39.2 cc (STP) CO2/cc sorbent or 1.46 mmol CO2/g sorbent. These sorption capacities are very competitive when compared to literature values of dry sorption capacity from aminated silicabased supports. Dry sorption capacities reported for anhydrous grafting of monoamine-trialkoxysilanes grafted silicas are reported to achieve capacities in the range of 0.41 to 1.59 mmol CO2/g sample. Diamino-trialkoxysilane species grafted with similar procedures have reported sorption capacities of 0.87
1.66 mmol CO2/g sample. Triamino-trimethoxysilane-grafted silicas sorption capacities for anhydrous grafting have a sorption capacity range of 0.97 to 1.55 mmol CO2/g sample. On the basis of these trends, it appears that the addition of water to the grafting reactions in silica-based supports has provided sample specific advantages in terms of sorption capacities, since monoamine-trialkoxysilanes grafted silicas have reported values of sorption capacity of 0.10
0.66 mmol CO2/g sample, while diamino-trialkoxysilane grafting has sorption capacities of 0.57
1.36 mmol CO2/g sample. Triamino-trimethoxysilane-grafted silicas have increased their capacities under wet conditions up to 1.58
2.65 mmol CO2/g sample.4a,c,d,g,15a
15c,29,30 The preliminary sorption equilibration time of sample A was about 3.5 h, which is of the same order of magnitude as the 2 h time for the pure cellulose acetate powder. Sample A captured an additional amount of CO2 on the order of 0.78 mmol CO2 per gram of sample, achieving an amine efficiency of 0.15 mol CO2 per mole of amines grafted. Amine
CO2 interactions during the sorption tests are affected by the possibility of hydrogen bonding between amine groups or with the remaining hydroxyl groups, leading to a reduced amount of free amine functionalities available for CO2 sorption. Amine efficiency values reported in the literature for anhydrous grafting of monoaminetrialkoxysilanes on silica-based adsorbents are in a range of 0.20
0.69 mol CO2/mol amine, 0.23
0.46 mol CO2/mol amine for diamino-trialkoxysilane grafted silicas, and 0.17
0.31 mol CO2/ mol amine for triamino-trimethoxysilane graftings. On the other hand, water-aided graftings of monoamine-trialkoxysilanes on silicabased adsorbents has showed CO2/N ratios of 0.02
0.25 mol CO2/mol amine, 0.22
0.30 ratios for diamino-trialkoxysilane grafted silicas, and 0.25
0.33 mol CO2/mol amine for triaminotrimethoxysilane graftings.4a,c,d,g,15a,15b,30,31 Therefore, although there are other CO2 sorbents with equal or higher sorption capabilities, it is believe that the improvement in cellulose acetate sorption capacity achieved in this study is promising. The simplicity of the preparation method and the low cost of the reactants make this procedure an attractive approach worthy of being further developed.

’ CONCLUSIONS The objective of this research was the design and development of a suitable experimental strategy to improve the CO2 sorption capacity of cellulose diacetate. The goal was achieved through the

ARTICLE

grafting of aminosilanes onto cellulose diacetate powder under a set of optimum reaction conditions derived from previous experimental approaches of the present study and from the literature. The use of a diamino-monomethoxysilane species as grafting agent under total anhydrous conditions greatly favored a high loading of amines into the polymer structure. A significant increase in the CO2 sorption capacity was registered for the sample with highest amine loading. In addition, exposure of the sorbent to water vapor did not negatively affect its sorption capacity; rather, it enhanced it. The aminosilane functionalized cellulose acetate sorbent in powder form was tailored to adsorb up to 27 cc (STP) CO2/ cc sorbent at 1 atm and about 39.2 cc (STP) CO2/ cc sorbent at 5 atm. Its amine loading was about 5.18 mmol amine(nitrogen)/g sorbent. The exposure of the sorbent to water vapor slightly increased its sorption capacity on a second CO2 sorption cycle. The new capacity of the sorbent reached 30 cc (STP) CO2/cc sorbent at 1 atm. This sorbent represents a very promising option for further investigation. Moreover, in the presence of humidity, further increases in sorption capacity are expected to occur on the basis of higher amine efficiencies known to occur in humid environments. For commercial cellulose acetate with acetyl content in the range of 39.8
40 wt %, the optimum grafting conditions yielded an amount of grafted amines of about 7 wt % for the C9H24N2OSi silane. The chemical reactivity of the aminosilane species under anhydrous or humid grafting conditions has an important impact in the amounts of amine loadings. Aminomethoxysilanes are among the most reactive organosilanes under anhydrous conditions, while aminoethoxysilanes have limited reactivity in the absence of water. Therefore, special attention should be paid to the aminosilane species selected for the grafting reactions and the corresponding grafting conditions used. The aminosilanefunctionalized cellulose acetate sorbents exhibit a relative short sorption equilibration time of about 2 h in the first sorption. The equilibration time can presumably be reduced using a highly porous fiber as opposed to dense particles or films. Future work is recommended in the following areas: (1) tuning up the optimal reaction conditions to find out further improvements in the CO2 sorption capacity; (2) investigation of the effects on the sorbent sorption capacity of a higher number of CO2 sorption cycles; (3) investigation of the sorbent sorption capacity when exposed to humid CO2 cyclic sorption tests; (4) research about the effectiveness of the aminosilanes grafting reaction in cellulose acetate fiber sorbents at optimal reaction conditions. Success of this novel material will depend on the effective improvement of the reaction conditions to identify suitable pathways to transform the modified sorbent into useful substrate forms, like films or fibers, and its efficient performance over humid CO2 cyclic sorption systems. Alternatively, of course, if highly porous precursor cellulose diacetate fibers can be treated to graft silanes on the already formed fibers, this could simplify the ultimate application of aminofunctionalized fiber sorbent technology.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT This publication is based in part on work supported by Award No. KUS-I1-011-21, made by King Abdullah University of 513

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514

Industrial & Engineering Chemistry Research

ARTICLE

Science and Technology (KAUST). The authors would also like to thank the Mexico’s National Council on Science and Technology (CONACYT) for their financial support in this research.

Biofunctionalization of Cellulose Acetate with Thermostable Chimeric Avidin. ACS Appl. Mater. Interfaces 2011, 3 (7), 2240. (13) Puleo, A. C.; Paul, D. R.; Kelley, S. S. The Effect of Degree of Acetylation on Gas Sorption and Transport Behavior in Cellulose Acetate. J. Membr. Sci. 1989, 47, 301–332. (14) Gelest, Inc.; Arkles, B.; Larson, G. Silicon Compounds: Silanes and Silicones: A Survey of Properties and Chemistry; Gelest, Inc.: Morrisville, PA, 2008. (15) (a) Raucher, D.; Sefcik, M. D. Gas Transport in Glassy Polymers as Predicted by the. Matrix Model. Polym. Prepr. 1983, 24, 87. (b) Hestekin, J. A.; Bachas, L. G.; Bhattacharyya, D. Poly(amino acid)-Functionalized Cellulosic Membranes: Metal Sorption Mechanisms and Results. Ind. Eng. Chem. Res. 2001, 40, 2668–2678. (c) Biermann, C. J.; Narayan, R. Grafting of Poly(ethylenimine) onto Mesylated Cellulose Acetate, Poly(methyl methacrylate) and Poly(vinyl chloride). Carbohydr. Polym. 1990, 12, 323. (d) Arockiasamy, D. L.; Nagendran, A.; Shobana, K. H.; Mohan, D. Preparation and Characterization of Cellulose Acetate/Aminated Polysulfone Blend Ultrafiltration Membranes and their Application Studies. Sep. Sci. Technol. 2009, 44 (2), 398–421. (16) Lapenko, V. L.; Slivkin, A. I.; Smirnov, D. N.; Zakharova, O. V.; Suntsova, N. S.; Dmitriev, L. A. Preparation of Amine-containing Cellulose Derivatives using Triethylenetetramine Copolymer. USSR Patent SU 1548184, Mar 7, 1990. (17) Engelhardt, H.; Orth, P. Alkoxy Silanes for the Preparation of Silica Based Stationary Phases with Bonded Polar Functional Groups. J. Liq. Chromatogr. 1987, 10 (8
9), 1999–2022. (18) Caravajal, G. S.; Leyden, D. E.; Quinting, G. R.; Maciel, G. E. Structural Characterization of (3-aminopropyl)triethoxysilane-modified Silicas by Silicon-29 and Carbon-13 Nuclear Magnetic Resonance. Anal. Chem. 1988, 60 (17), 1776–86. (19) Zhang, F. S.; M., P. Self-Assembled Molecular Films of Aminosilanes and Their Immobilization Capacities. Langmuir 2004, 20 (6), 2309–2314. (20) Smith, E. A.; Chen, W. How to Prevent the Loss of Surface Functionality Derived from Aminosilanes. Langmuir 2008, 24 (21), 12405–9. (21) Joensson, U.; Olofsson, G.; Malmqvist, M.; Ronnberg, I. Chemical Vapor Deposition of Silanes. Thin Solid Films 1985, 124 (2), 117–23. (22) Tsujita, Y. Gas Sorption and Permeation of Glassy Polymers with Microvoids. Prog. Polym. Sci. 2003, 28, 1377–1401. (23) Koros, W. J.; Paul, D. R. Design Considerations for Measurement of Gas Sorption in Polymers by Pressure Decay. J. Polym. Sci., Part B: Polym. Phys. 1976, 14, 1903. (24) Suthar, J. N.; Patel, M. J.; Patel, K. C.; Patel, R. D. Studies on Structural Aspects of Cellulose Acetate. Angew. Makromol. Chem. 1985, 130, 125–136. (25) Launer, P. J. Infrared Analysis of Organosilicon Compounds: Spectra Structure Correlations. Silicon Compounds Register and Review [Online]; Petrarch Systems: Bristol, PA, 1987, 100
103. (26) Smith, B. Infrared Spectral Interpretation: A Systematic Approach; CRC Press: Boca Raton, 1998. (27) Kanan, S. M.; Tze, W. T. Y.; Tripp, C. P. Method to Double the Surface Concentration and Control the Orientation of Adsorbed (3-Aminopropyl)dimethylethoxysilane on Silica Powders and Glass Slides. Langmuir 2002, 18 (17), 6623–6627. (28) Lawrence-Berkeley National Labs, Characteristic IR Band Positions. http://infrared.als.lbl.gov/content/web-links/60-ir-band-positions (Accessed 2009). (29) Daus, P. A.; Pauley, C. R. Process for Separating Natural Gas and Carbon Dioxide. Patent US 6128919, October 2000. (30) Harlick, P. J. E.; Sayari, A. Applications of Pore-Expanded Mesoporous Silica. 3. Triamine Silane Grafting for Enhanced CO2 Adsorption. Ind. Eng. Chem. Res. 2006, 45, 3248. (31) Pusch, W.; Tanioka, A. Structure Investigations of Homogenous Cellulose Acetate Membranes by Gas Permeation Experiments. Desalinisation 1983, 46 (1
3), 425.

’ REFERENCES (1) Lively, R. P.; Chance, R. R.; Kelley, B. T.; Deckman, H. W.; Drese, J. H.; Jones, C. W.; Koros, W. J. Hollow Fiber Adsorbents for CO2 Removal from Flue Gas. Ind. Eng. Chem. Res. 2009, 48, 7314–7324. (2) Su, F.; Lu, C.; Kuo, S. C.; Zeng, W. Adsorption of CO2 on Aminefunctionalized Y-type Zeolites. Energy Fuels 2010, 24, 1441–1448. (3) (a) Plaza, M. G.; Pevida, C.; Arias, B.; Casai, M. D.; Martin, C. F.; Fermoso, J.; Rubiera, F.; Pis, J. J. Different Approaches for the Development of Low-cost CO2 Adsorbents. J. Environ. Eng. 2009, 135, 426–432. (b) Czakkel, O.; Onyesty
ak, G.; Pilatos, G.; Kouvelos, V.; Kanellopoulos, N.; L
aszll
o, K. Kinetic and Equilibrium Separation of CO and CO2 by Impregnated Spherical Carbons. Microporous Mesoporous Mater. 2009, 120, 76–83. (4) (a) Huang, H. Y.; Yang, R. T.; Chinn, D.; Munson, D. L. AmineGrafted MCM-48 and Silica Xerogel as Superior Sorbents for Acidic Gas Removal from Natural Gas. Ind. Eng. Chem. Res. 2003, 42, 2427–2433. (b) Hiyoshi, N.; Yogo, K.; Yashima, T. Adsorption of Carbon Dioxide on Aminosilane-modified Mesoporous Silica. J. Jpn. Pet. Inst. 2005, 48 (1), 29–36. (c) Knowles, G. P.; Delaney, S. W.; Chaffee, A. L. AmineFunctionalised Mesoporous Silicas as CO2 Adsorbents. Stud. Surf. Sci. Catal. 2005, 156, 887. (d) Zheng, F.; Tran, D. N.; Busche, B. J.; Fryxell, G. E.; Addleman, R. S.; Zemanian, T. S.; Aardahl, C. L. EthylenediamineModified SBA-15 as Regenerable CO2 Sorbent. Ind. Eng. Chem. Res. 2005, 44, 3099. (e) Yue, M. B.; Chun, Y.; Cao, Y.; Dong, X.; Zhu, J. H. CO2 Capture by As-prepared SBA-15 with an Occluded Organic Template. Adv. Funct. Mater. 2006, 16, 1717–1722. (f) Kn€ ofel, C.; Martin, C.; Hornebecq, V.; Llewellyn, P. L. Study of Carbon Dioxide Adsorption on Mesoporous Aminopropylsilane-Functionalized Silica and Titania Combinig Microcalorimetry and in Situ Infrared Spectroscopy. J. Phys. Chem. C 2009, 113, 21726–21734. (g) Harlick, P. J. E.; Sayari, A. Applications of Pore-Expanded Mesoporous Silica. 5. Triamine Grafted Material with Exceptional CO2 Dynamic and Equilibrium Adsorption Performance. Ind. Eng. Chem. Res. 2007, 46 (2), 446–458. (h) Kim, S.; Ida, J.; Guliants, V. V.; Lin, Y. S. Functionalised Mesoporous Silica Membrane for the Separation of Carbon Dioxide. Int. J. Environ. Technol. Manage. 2004, 4 (1
2), 21. (5) (a) Couck, S.; Denayer, J. F. M.; Baron, G. V.; Remy, T.; Gascon, J.; Kapteijn, F. An Aminefunctionalized MIL-53 Metal-organic Framework with Large Separation Power for CO2 and CH4. J. Am. Chem. Soc. 2009, 131, 6326–6327. (b) Yazaydin, A. O.; Snurr, R. Q.; Park, T. H.; Koh, K.; Liu, J.; LeVan, M. D.; Benin, A. I.; Jakubczak, P.; Lanuza, M.; Galloway, D. B.; Low, J. J.; Willis, R. R. Screening of Metal-organic Frameworks for Carbon Dioxide Capture from Flue Gas Using a Combined Experimental and Modeling Approach. J. Am. Chem. Soc. 2009, 131, 18198–18199. (6) Tanghe, L. J.; Genung, L. B.; Mench, J. W. Methods in Carbohydrate Chemistry; Academic Press: New York, 1963; Vol. 3. (7) Kamidem, K. S. M. Adv. Polym. Sci.: Biopolymers 1987, 83, 1. (8) Fischer, S.; Th€ummler, K.; Volkert, B.; Hettrich, K.; Schmidt, I.; Fischer, K. Properties and Applications of Cellulose Acetate. Macromol. Symp. 2008, 262 (Structure and Properties of Cellulose), 89–96. (9) Kiso, Y.; Kitao, T.; Nishimura, K. J. Appl. Polym. Sci. 1998, 71, 1657. (10) (a) Takahashi, T. Adsorption of N-methylcarbamates on Membrane Filters. Hokkaidoritsu Eisei Kenkyushoho 2000, 50, 75–77. (b) Yartseva, N. M.; Ryabukhova, T. O.; Okisheva, N. A.; Ramazaeva, L. F.; Surkova, A. N. Adsorption on Porous Polymer Sorbents in Ultrafiltration. Fibre Chem. 1998, 30 (4), 260. (11) Miura, M.; Cole, C. A.; Monji, N.; Hoffman, A. S. Temperaturedependent Adsorption/desorption Behavior of Lower Critical Solution Temperature (LCST) Polymers on Various Substrates. J. Biomater. Sci., Polym. Ed. 1994, 5 (6), 555–68. (12) Heikkinen, J. J.; Riihim€aki, T. A.; M€a€att€a, J. A. E.; Suomela, S. E.; Kantomaa, J.; Kulomaa, M. S.; Hyt€onen, V. P.; Hormi, O. E. O. Covalent 514

dx.doi.org/10.1021/ie2020685 |Ind. Eng. Chem. Res. 2012, 51, 503–514