Amine-based CO2 capture sorbents_ A potential CO2

Journal of CO₂ Utilization 26 (2018) 397–407

Contents lists available at ScienceDirect

Journal of CO2 Utilization journal homepage: www.elsevier.com/locate/jcou

Amine-based CO2 capture sorbents: A potential CO2 hydrogenation catalyst ⁎

T

Srikanth Chakravartula Srivatsa , Sankar Bhattacharya Chemical Engineering Department, Monash University, Wellington Road, Clayton, VIC-3800, Australia

A R T I C LE I N FO

A B S T R A C T

Keywords: FTIR CO2capture Amine coverage Carbamate Carbamic acid

Mechanism of CO2 adsorption on amine loaded SBA-15 sorbents with varying amine coverage has been assessed using thermogravimetric analysis (TGA) and in-situ Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS). TGA study showed that sorbent adsorption capacities increased by 122–153% with CO2 concentrations (5–80%) in the gas at 50/75 °C. The DRIFTS studies indicate that low amine loaded sorbents exhibit higher uptakes with an increase in CO2 concentrations which is attributed to the mode of CO2 adsorption i.e. 1:1 CO2 to amine forming carbamic acid. At higher amine loadings CO2 is adsorbed following the 1:2 CO2 to amine forming carbamate-ammonium ions pair resulting in lower adsorption capacities per amine site. The studies also showed an increase in carbamic acid formation with pressure from 100 to 500 kPa at low amine loadings and both carbamic acid and carbamates at higher amine loadings. The paper provides insights into the mechanistic understanding of CO2 adsorption behaviour of the varyingly covered amines on the SBA-15 support with a change in concentration and pressure of CO2. The current work presents the conditions to alter the CO2 adsorption mechanism on amine sites with potential application in CO2 conversion to chemicals.

1. Introduction Selective removal of CO2 from industrial gas streams have been reported to be carried out by physical adsorption of CO2 on materials like zeolites [1,2], metal-organic frameworks (MOF) [3,4], chemical absorption on to solvents [5–7], adsorption on solid amines [8–15] and membrane separations [16–18]. During the past decade, adsorption of CO2 using amines immobilised on solid supports gained prominence to capture the CO2 emission from coal-fired power plants. Amines immobilized on solid supports like Santa Barbara Amorphous, SBA-15 [19,20], Mobil Composition of Matter MCM-41 [21–23], SiO2 [12,24,25], hierarchically porous silica HPS [26,27], ion-exchange based resin [28] and carbonaceous materials [11,13,29] are promising candidates with high adsorption capacities. Although many aminebased sorbents were identified as potential sorbents for CO2 capture they either suffer from inferior CO2 adsorption efficiency, lack of longterm cyclic stability in the long term or high final cost of the sorbent to capture CO2. All the sorbents are tested for only for few cycles which does not give a true picture of the sorbent stability though might have high adsorption capacities. A recent work by Gadipelli et al. on amine loaded hyperporous graphene networks showed enormous promise with very high capture capacity of 7.5 mmol/g-sorbent [30]. The sorbents suffer degradation over long cyclic times due to the formation of urea in CO2 environment or formation of N]O due to degradation of primary amine in the air [28]. ⁎

Corresponding author. E-mail address: [email protected] (S. Chakravartula Srivatsa).

https://doi.org/10.1016/j.jcou.2018.05.028 Received 15 March 2018; Received in revised form 12 May 2018; Accepted 29 May 2018 2212-9820/ © 2018 Elsevier Ltd. All rights reserved.

However, the present work is focused on understanding the nature of CO2 adsorption mechanism under varying adsorption conditions such as gas concentration and pressure and amine loading. The amine efficiencies (CO2/N ratio) only remained < 0.4 in the majority of the sorbents suggesting the amines sites present in the sorbent remained significantly underutilized. Hence, it is important to identify possibilities which can allow these amines sites to participate in the CO2 adsorption process and achieve higher adsorption [31] capacities and open opportunities for chemical production by CO2 as reactant and amine site as an active site for CO2 activation. It is accepted that CO2 is adsorbed on amines and produce carbamates which require two amine sites for its adsorption and stabilisation resulting in only 50% capture efficiency. However, the nature of adsorbed species can be varied depending on the amine loading, type of amine, density of amine sites and the molecular size of amine immobilised on the solid support and importantly the concentration of CO2 or pressure used for the CO2 capture. In our previous studies [32] the effect of Tetraethylenepentamine (TEPA) loading on the CO2 adsorption on fresh and oxidative degraded TEPA/PEG/SiO2 sorbents by FTIR revealed the presence of weakly and strongly adsorbed CO2 and that their fractions depend on the amine loadings. At lower amine loadings, a higher amount of weakly adsorbed CO2 was present and was removed easily by the flow of inert gas. At higher loadings, the adsorbed CO2 was stabilised by the hydrogen bonding with adjacent amine molecules, thus producing a higher fraction of strongly adsorbed


S. Chakravartula Srivatsa, S. Bhattacharya

CO2 [32]. CO2 adsorption on degraded oxidative sorbents showed an increase in the carbamic acid formation due to the isolation of amine sites and physisorption of CO2 onto surface silanol groups [32]. When CO2 is adsorbed as a carbamic acid, it requires only one amine site per CO2 adsorption, which suggest that we can double the adsorption capacity of the sorbents by forming carbamic acid rather than carbamates. Hence, if the adsorption conditions are such that CO2 is adsorbed on single amine site as carbamic acid, the adsorption capacities of the amine-based sorbents can be increased to two folds. The adsorption of CO2 as carbamic acid on amine sites provides an opportunity to convert CO2 to value-added chemicals such as formic acid or methanol in presence of H2. To this extent, different amines sorbents were prepared and the adsorption conditions are varied to change the of adsorbed CO2 species. In the present work, we prepared four different amines with varying amine density are loaded onto the SBA-15 support. The sorbents were tested with varying CO2 concentrations in the flue gas and different pressures to investigate the adsorbed species using in situ FTIR studies. The work highlights how the adsorbed CO2 species on solid sorbents can be adjusted with varying amine loadings or change in theCO2 concentrations or pressures to increase the adsorption capacity of sorbents.

P/Po of 0.995. Brunauer-Emmett-Teller (BET) surface area was measured by 0-0.3 P/Po and total pore volume by at 0.995 from the amount of liquid nitrogen adsorbed. Amine loading on the sorbents was measured following an established procedure [33] by using Thermogravimetric analysis (TGA) (Model STA 449 F3 Jupiter, NETZSCH-Geratebau GmbH, Germany). 10 mg of the sample was pretreated at 105 °C for 30 min in the N2 atmosphere to remove any pre-adsorbed CO2 and water from the sample. The sample was then heated to 200 °C and 600 °C and held for 30 min respectively followed by heating to 900 °C where the N2 gas is switched to Air for 30 min. The weight loss from 105 °C is considered as the amine loading on the sample. Elemental analysis was performed using Perkin Elmer-2400 series to determine % N content in the sorbents by combusting the sample at 950 °C. Fourier Transform Infrared (FTIR) spectrum of the SBA-15 and sorbents was collected using Perkin Elmer Frontier FTIR with Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) accessory capable of high temperatures up to 850 °C and 34 bar pressure. The spectra were collected from 4000–1000 cm−1 with 16 co-added scans at a spectral resolution of 4 cm−1.

2. Experimental

2.2.1. Thermogravimetric analysis Measurement of the CO2 adsorption capacity of various sorbents was carried out by taking 20 mg of the sample in the ceramic crucible. The sample was pretreated at 130 °C for 15 min in 100 ml N2 flow and allowed to cool down to 50 °C in the same N2 atmosphere. Varying concentrations of 5, 15, 30, 50 and 80% CO2 balanced in N2 at a flow rate of 100 ml/min for 30 min for the saturation of the sample. The samples were then purged for 15 min using N2 flow at 100 ml/min, and Temperature programmed desorption (TPD) was performed by heating the sample from 50 °C to 130 °C at 5 °C/min and held at 130 °C for 15 min in N2 flow 100 ml/min to completely regenerate the sorbent.

2.2. CO2 adsorption/desorption

The following procedure was used to prepare the SBA-15 support used for the sorbent preparation. In a typical preparation, 7.8 g of Pluronic 123 surfactant (poly (ethylene glycol)-block-poly (propylene glycol)-block-poly (ethylene glycol) average MW ∼5800, Aldrich Chemicals) was dissolved in 146 ml water and 6.2 ml anhydrous acetic acid solution (Aldrich Chemicals). The solution was stirred until all P123 had dissolved and a consistent, translucent appearance was achieved. 24.5 g of tetraethoxysilane (TEOS, Aldrich Chemicals) was added dropwise, and the solution stirred continuously at 40 °C for 20 h. This mixture was then transferred into a reagent bottle capable of withstanding pressure and was heated in an oven at 105 °C for 24 h. The solution was then centrifuged, the solid product extracted and washed with ethanol and water and dried in the oven at 105 °C overnight. The P123 template was removed by calcining the material at 500 °C for 5 h in the air. The amines used in this study are 3-(aminopropyl) trimethoxysilane (APTS, Aldrich Chemicals), N-(3-Trimethoxysilylpropyl) diethylenetriamine (Triamine, Aldrich Chemicals), and polyethyleneimine (PEI, 50 wt% in water, MW = 1200 Aldrich Chemicals). Two grams of SBA-15 sample was treated in an oven at 100 °C for 12 h to remove moisture. The dehydrated sample was dispersed in a mixture of 3 mL of APTS or Triamine and 60 mL isopropanol and refluxed at 95 °C for 3 h. After refluxing, the suspended solid product was filtered and washed with anhydrous ethanol three times and then dried at 70 °C for two h in air to give APTS/SBA-15 and Triamine/SBA-15 sorbents. Sorbents with 10 and 25 wt% content of amine (PEI) on SBA15 support were prepared by mixing SBA-15 with 10 mL of an ethanol solution containing the requisite amount of PEI. The mixtures were dried in an oven at 100 °C until all the ethanol was evaporated. In specific for PEI10 sorbent 1.9 g of SBA-15 and 0.2 g of 50 wt% PEI in water was mixed with 10 ml of ethanol to disperse the amine on to the support and the excess ethanol and water was evaporated.

2.2.2. In-situ FTIR studies In-situ FTIR studies using Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) mode is performed to investigate the mechanism of CO2 adsorption on various amine sorbent. FTIR is an analytical technique used to measure absorption of infrared radiation by the sample versus wavelength. FTIR give characteristics peaks for each chemical bond in the molecule when they are polar without overlapping. The vibrations bands can be observed simultaneously and continuously giving us a real-time understanding the mechanism involved in the adsorption-desorption process The sorbents were pretreated by heating the sample from room temperature to 130 °C in the N2 flow of 100 ml/min and held for 15 min before cooling down to 50 °C to conduct CO2 adsorption/desorption cycles. The varying concentrations of CO2 gas were achieved by controlling the flow rates of CO2 and N2 using mass flow controllers. Similar to TGA profiles CO2 adsorption/desorption was carried out in FTIR-DRIFTS mode, except the CO2 adsorption time was kept for only 10 min while N2 purge for 5 min. The changes in the spectral features of the sorbent throughout the experiment were monitored in a continuous series scan single beam mode using Perkin Elmer Frontier FTIR with ten co-added scans at a resolution of 4 cm−1. The single beam spectra are normalized according to the procedure reported before proceeding to absorbance spectra to observe the spectral changes in the sample during CO2 adsorption [32]. Once all the spectra are normalized and absorbance spectra are obtained, the intensities against the desired wavenumbers with time can be plotted against to obtain the profile at each wavenumber. To study the effect of pressure on CO2 adsorption, the sorbents were initially saturated with 15% CO2 for 10 min. The chamber was then set to batch mode closing the inlet and outlet valves. The pressure of the chamber was maintained using a back-pressure regulator to the desired

2.1. Characterization N2 adsorption/desorption isotherms on the SBA-15 and amine loaded sorbents were performed using the ASAP 2020 (Micromeritics USA). 100 mg of the sample was loaded into the sample tube and pretreated at 80 °C for 2 h. Low-temperature pretreatment was chosen to prevent the loss of amine under high vacuum treatment. The N2 adsorption was performed at liquid nitrogen temperatures up to relative 398



Table 1 Characteristics of amine sorbents used in the present study. Sorbent

BET surface area (m2/g)

Pore size (nm)

Pore volume (cm3/g)

N content (wt %)a

% Amine loadingb

SBA-15 APTS Triamine PEI10 PEI25

790 232 202 364 150

5.99 5.47 5.37 4.55 6.11

0.97 0.32 0.38 0.49 0.29

– 3.1 6.9 6.3 12.4

– 16.3 19.5 18.0 30.1

a b

determined from elemental analyser. determined from TGA.

pressure range 100–500 kpa and pressurised gradually. The sample was maintained at each pressure for 5 min before taking it to the next set point. During the entire process, the spectra were collected using PE Time based software at one scan per 5 s. 3. Results and discussion 3.1. Characterization results

Fig. 2. FTIR spectra of SBA-15 and different sorbents after pretreatment at 130 °C for 15 min.

Table 1 shows the results of BET surface area, pore size, and pore volume of SBA-15 and amine loaded sorbents determined from the N2 adsorption isotherms. The results show a decrease in the surface area and pore volume and pore size indicating that amine is functionalized on the pores of the SBA-15 support. The decreases in surface area are in line with the amine loadings. Table 1 also shows the results of % N content and wt% amine loading determined from elemental and TGA analysis respectively. The TGA results suggest that the amine loading is slightly higher than what is intended to impregnate on the sorbent. Fig. 1 shows the TGA profiles of SBA-15 and amine loaded samples. Initially, the sample was heated to 105 °C and held there for 60 min to remove the adsorbed CO2 and moisture from the sorbent. TGA profile of the samples shows weight loss including SBA-15 during this period due to the removal of adsorbed CO2 and water vapour from the sample. In the next step when the sample was heated to 200 °C and held there for 30 min, amine loaded samples showed weight loss which is attributed to the loss of methoxy ligands in APTS and Triamine functionalized sorbents. However, on PEI loaded samples the weight loss is relatively high suggesting a possible leaching of the amine from the sorbents. The loss in sample weight from 200 °C to 900 °C is due to the breakdown of the amine present in the sorbent and reported in Table 1. It can be observed that the amine loading is slightly higher than expected in PEI loaded sorbents. Fig. 2 shows the IR absorbance spectra of SBA-15 material along with amines grafted and impregnated sorbents. The spectrum of SBA-15

shows features from free silanol groups SieOH at 3750 cm−1, hydrogen-bonded groups at 3660 cm−1 and broad hydrogen bonded water centred at 3450 cm−1. The spectra here are obtained after pretreatment of the samples at 130 °C in nitrogen flow for 15 min and cooling down to 50 °C. The other major functional group vibrations in SBA-15 are Si-O-Si skeletal vibrations at 1060 and 1165 cm−1. The spectra of the sorbents obtained by grafting of APTS and Triamine, and impregnation of 10 and 25 wt % PEI on to the SBA-15 support shows peaks due to corresponding compounds. Amine loaded samples shows the vibrations due to NH2 symmetric and asymmetric stretching vibrations at 3290 cm-1 and 3360 cm-1 respectively, NH2 deformation at 1601 cm-1, CH2/CH3 stretching vibrations at 2931, 2860 and 2815 cm-1 and CH2 deformation at 1458 cm-1. The detailed IR band assignments are presented in Table 2. The addition of amines on the support suppressed the peaks due to surface silanol groups suggesting APTS and Triamine were successfully grafted on the SBA-15 support. PEI10 sample shows a small peak of free SieOH groups suggesting the sample is yet to form a monolayer over the support surface whereas PEI25 did not show any sign of silanol groups suggesting monolayer coverage of amine is achieved on the support. The spectra also show the presence of hydrogen-bonded SieOH groups at 3660 cm-1 after functionalization with an amine. 3.2. CO2 adsorption using TGA Fig. 3 shows the thermogravimetric % weight change profile of different sorbents during CO2 adsorption/desorption for five cycles. The Table 2 IR Band assignments. Wavenumbers (cm−1)

Functional groups

Ref

3750 3650–3740 3660 3500–3400

Free SieOH groups hydrogen bonding SieOHehydrogen bonded OeH⋯OH hydrogen bonds of adsorbed H2 O NeH stretching CeH stretching Bending vibration of adsorbed H2O NH2 bending vibrations CH2 bending vibrations SieOeSi vibrations of SBA-15

[32,34] [19,34] [19] [35]

3360, 3290 2931, 2810, 2880 1630 1600–1605 1450–1460 1165, 1060

Fig. 1. Proximate analysis of SBA-15 and amine loaded sorbents to determine the % amine loading. 399

[12,19,24] [19,34] [24,32] [24,32,35] [19] [35]



The results in Fig. 3 and Table 3 show that there is a considerable increase in the adsorption capacity of the sorbents with an increase in CO2 concentration in the flue gas (CO2/N2) for all the sorbents used in the study. Triamine sorbent at 5–15 % CO2 concentration showed relatively lower adsorption capacity than expected, which could be either due to incomplete pretreatment of the sorbent. The samples during desorption at 5 and 15% CO2 concentration showed higher weight loss compared to weight gained during CO2 adsorption indicating the reason for low adsorption capacity. This also reflected in overall % increase in the adsorption capacity. When the amine loading was increased to 40 wt% PEI, no major change in the adsorption capacity with concentration, which is attributed to the diffusional limitation of CO2 to access the amine sites. PEI, being a three-dimensional polymeric molecule, at higher loadings, completely occupies the pores of SBA-15 support, and its high viscosity denies the access of CO2 molecules to the amine sites present inside the pores of the SBA-15 support. 40% PEI sorbents were not considered for further study. However, when the CO2 adsorption was performed at elevated temperatures at 75 °C the viscosity of PEI is decreased allowing the branches of PEI to expand and increase the accessibility of amine sites for CO2 adsorption. Similar results are reported in the literature [9,20]. The results show that at 75 °C the adsorption capacity of the PEI sorbents are higher compared to 50 °C and increase in CO2 concentrations further increased the adsorption capacity. The CO2 adsorption capacities of the sorbents reported in this work are smaller compared to the values reported in the literature on similar sorbents. However, the interest of this study is to see whether the sorbents show any change in adsorption capacities or adsorbed CO2 species by altering the adsorption conditions. Alternatively, the fresh sorbents are tested for CO2 capture at 80% CO2 for comparison. All the samples showed capture capacity close to the values of the sorbent gone through 5-cycles. The results suggest that the sorbents do not suffer major deactivation in the five cycles studied. The FTIR spectra of the sorbent after each adsorption cycle was compared (Supplementary S1) did not show or formation of new peaks due to deactivation. The increase in CO2 adsorption capacity on the amine sorbents with a CO2 concentration in the flue gas is hypothesised as follows. The adsorption of CO2 on the amines sites is assumed to follow a two-step mechanism in which CO2 is first adsorbed on amine site as carbamic acid which later transfers the proton to adjacent amine site to form carbamate (COO−) and ammonium ion (NH3+) pair to stabilise as represented in reaction 1 [36–39]

Fig. 3. Thermogravimetric analysis of CO2 adsorption behaviour of different sorbents with a varying CO2 concentration in the flue gas.

+RNH2

RNH2 + CO2 ↔ RNHCOOH ←⎯⎯⎯⎯→RNHCOO− + RNH3+

sorbents were tested for the adsorption capacities with varying CO2 concentrations from 5 to 80%. FTIR spectra of the samples did not show the formation of new peaks suggesting no appreciable degradation has happened to the sample during the 5 cycles of CO2 adsorption-desorption. The corresponding CO2 adsorption capacities and % increase in overall adsorption capacity with a change in CO2 concentration in the feed gas were reported in Table 3. While estimating % increase in adsorption capacity the values at 15% CO2, which is a typical CO2 concentration in power station flue gas, is taken as a baseline.

Carbamic acid

Carbamate

Reaction 1: CO2 adsorption on amines sites on solid sorbents. The equation suggests that for adsorbed CO2 to get stabilised two amine sites are required for carbamate, whereas carbamic acid requires only one amine site. Hence, if the adsorption conditions are maintained such that majority of the CO2 is adsorbed as carbamic acid, the amine efficiency can be doubled. Also, in the literature, it was observed that the maximum amine efficiency achieved on amine sorbents is > 0.35 suggesting that there are several amines sites present in the sorbent

Table 3 CO2 adsorption capacity of various sorbents determined using Thermogravimetric analysis (TGA). Sample

APTS Triamine PEI10 PEI25 PEI10 PEI25 a

Adsorption temperature (oC)

50 50 50 50 75 75

Adsorption Capacity (mmol/g)

% increase

5%

15%

30%

50%

80%

a

0.70 0.49 0.10 0.61 0.23 0.87

0.81 0.95 0.19 0.78 0.33 1.06

0.88 1.25 0.23 0.83 0.41 1.14

1.04 1.61 0.23 0.86 0.46 1.19

1.12 1.70 0.26 0.92 0.50 1.24

1.01 1.56 – – 0.65 1.16

represents the capture capacity on fresh sorbent at 80% CO2. 400

80% 160 177 136 149 153 141



hydrogen bonded carbamic acid at 1680 cm−1. The broad hump between 2800-2050 cm−1 represents the presence of adsorbed CO2 as zwitterions [32,40]. The spectra in Fig. 4 shows a gradual increase in the peak intensities of gas-phase CO2 overtones at 3700-3600 cm−1 with a concentration on APTS sorbent. As a result, there is a gradual increase in the peak intensities at 1700 and 1680 cm−1 corresponding to the free and hydrogen bonded carbamic acid. Similar observations are made on all the sorbents used in the study. However, the extent of increase in carbamic acid intensity varied with the amine loading on the sorbent. Fig. 5 shows the comparison spectra at varying CO2 concentrations after 600 s of CO2 adsorption on various sorbents. Depending on the density of amine sites the formation of carbamates and carbamic acids bands also varied with CO2 concentration. For example, APTS sorbents showed a decrease in the carbamate formation, which is also accompanied by a decrease in ammonium ion formation at 3030 cm−1. Triamine sample showed an increase in both carbamic acid and carbamates intensity. PEI10 sorbent with low amine loading showed a gradual increase in both the species; this is attributed to the nature of PEI molecule. PEI being a large three-dimensional molecule the amine sites are closely packed allowing the adsorbed CO2 to stabilise as carbamate with adjacent amine sites as well as an increase in the carbamic acid formation due to sudden exposure to high concentration of CO2 in the gas mix. This is explained as the amines on the terminal sites form carbamic acid whereas the amines present internally in the branches will be stabilised as carbamates. A similar observation is also seen in PEI25 sorbent. The schematic of adsorbed CO2 species with a change in amine loading and CO2 concentration is illustrated in Fig. 6. To further support the observations in Figs. 4 and 5 the absorbance intensities of carbamic acid peaks with time are plotted. Fig. 7 shows the absorbance intensity profiles of carbamic acid 1680-1710 cm−1 with CO2 adsorption time on various sorbents used in this study. The results show that there is an increase in the IR absorbance intensity of carbamic acid peak formation with CO2 concentration indicating that formation of carbamic acid is indeed is one of the causes for the increase in adsorption capacity of the sorbents. This observation is in line with the results observed from the TGA suggesting that the rise in adsorption capacity is also due to the adsorption of CO2 as carbamic acid and increased access to amine sites at higher CO2 concentrations. The carbamic acid formation is ideal to double the adsorption capacity. However, the amine density plays a significant role. At low amine loadings with amines spread out, the possibility of carbamic acid formation can double the adsorption capacity. APTS with only one amine site per moiety, when functionalized on SBA-15, is dispersed well over the surface which allows adsorption of CO2 on the individual amine sites to form carbamic acid first and then further gets stabilised as a carbamate. With the increase in the concentration of CO2 first step is fast achieved on a larger number of amine sites limiting the transfer of protons to adjacent amine sites to proceed to the second step, i.e., the formation of carbamate –ammonium ions pairs. This helped in the increase in adsorption capacity, i.e., an increase in efficiency of amine sites with an increase in CO2 concentration on APTS sorbent. However, the quantification of the ratio of carbamic acid/carbamate was not explored in this work yet. The intensity profiles for carbamate formation at 1560 cm−1 during CO2 adsorption on various sorbents are presented in Figure S2 in the supplementary information. The results indicate that there is a gradual decrease in the intensity of the profiles for APTS or PEI10. PEI25 sorbent did not show much variation in the carbamate formation due to the close proximity of amines at higher amine loading. However, Triamine sorbent shows a sharp increase in carbamate formation indicating the increase in CO2 concentration through increased CO2 capture is mainly from carbamate formation. Although, the carbamic acid formation is also observed to increase as seen from Fig. 7, due to the close proximity of amines carbamate has substantially increased. The intensity profiles for ammonium ion formation during CO2

which were not occupied during CO2 adsorption. Alternatively, the adsorption on amines can be represented as zwitterions [27,39,40] as shown in reaction 2

R − NH2 + CO2 ↔ RNH2+COO− Reaction 2: Zwitterion pair formation during CO2 adsorption It is established in the literature that in the presence of humidity in the stream the CO2 adsorption capacity of the adsorbent increases [22] and hypothesized to follow the mechanism in reaction 3, where one amine site is sufficient for adsorption of CO2 molecule [24]. However, the increase in adsorption capacity was never doubled in any study so far, but the presence of humidity helped in preventing the degradation of the sorbents [41] to an extent. However, in the present study, the effect of humidity on the adsorption capacity was not investigated but focused on changing the adsorbed CO2 to carbamic acid which has potential application towards converting CO2 to chemicals with aminebased sorbents are catalysts.

R1 R2 NH + CO2 + H2 O ↔ R1 R2 NH2+ + HCO3− Reaction 3: CO2 adsorption on amines sites on solid sorbents in humid conditions. Results in Table 3 show the increase in CO2 adsorption capacity of the sorbents with an increased CO2 concentration in the flue gas. The results can be attributed either due to access to a higher number of amines sites or due to adsorption of CO2 in the form of carbamic acid. The rate of adsorption on active sites is directly proportional to concentration/partial pressure of the adsorbate and is written as follows i.e., rate = Ka x CCO2 = Ka x PCO2 Where Ka is the rate constant for adsorption and Cco2 is a concentration of CO2 and PCO2 is partial pressure of CO2 in the reactor. In CO2 rich flue gas, there is a high probability of CO2 adsorbing on multiple numbers of individual amines sites at the same time which reduces the abundance of free amine sites for the transfer of a proton to adjacent amine site for the formation of NH3+ ions as these are already occupied with CO2. This phenomenon of increased competition for amine sites at higher CO2 concentrations resulted in the formation of a higher number of adsorbed CO2 molecules as carbamic acid, increasing the adsorption capacity and overall efficiency of the sorbents. To verify the above hypothesis a series of CO2 capture experiments like TGA experiments were performed on the powder sorbents with varying CO2 concentrations using in situ FTIR – DRIFTS to observe the of adsorbed CO2 species. The species of interest is the formation of carbamic acid with a change in the adsorption conditions are highlighted in the remainder of the work. 3.3. FTIR studies on CO2 adsorption Fig. 4 shows the IR absorbance spectra on APTS/SBA-15 sorbent during CO2 adsorption with varying CO2 concentrations 15, 30, 50, and 100% and time until 600 s. The absorbance spectra of APTS/SBA-15 with varying CO2 concentrations showed IR bands for gas phase CO2, ammonium ion (NH3+/NH2+), and adsorbed CO2 species. The band assignments because of adsorbed CO2 in the present study are summarised in Table 4. Adsorption of CO2 on amines resulted in the formation of NH3+/NH2+ with stretching vibrations at 3010-3030 cm−1 and NH3+/NH2+ bending at 1610-1630 cm−1. Spectra also exhibited a decrease in the IR intensities of amine stretching bands at 3360 and 3290 cm-1 coupled with suppression of CH2 stretching vibrations at 2810-2934 cm−1. The features in the fingerprint region are from asymmetric and symmetric stretching vibrations of COO- in carbamates at 1560-1570 cm−1 and 1500 cm−1, NCOO- skeletal vibrations are observed at 1410 cm-1 and1320 cm−1. The major IR bands of interest are the peaks due to carbamate ions at 1560-1570 cm−1 ammonium ions at 3010-3020 cm-1 and carbamic acid at 1700-1715 cm−1 and 401



Fig. 4. In situ FTIR absorbance spectra on APTS sorbent with varying CO2 concentrations during CO2 adsorption for 10 min.

adsorption on various sorbents are presented in Figure S2 in the supplementary information. The results indicate that there is a gradual decrease in the intensity of the profiles for APTS or PEI10 and PEI25 sorbents were observed. However, the intensity change is more clearly envisioned in the case of Triamine sorbent. With the increase in the amine loading, for example on Triamine sorbent, the increase in carbamic acid formation is not high as compared to the APTS. This observation suggests that increase in amine site density will allow the formation of ammonium ions resulting in stabilisation of the adsorbed CO2 as a carbamate. In other words, the terminal amine sites which are fast accessible for incoming CO2 at higher concentrations form carbamic acid. The sites present inside the chain moiety have relatively slower access which can transfer a proton to adjacent amine site and can be stabilised as a carbamate. Thus, at higher loadings, the increase in amine efficiency is not as high compared to low amine loadings at high concentrations of CO2 as shown in Table 3. The results also suggest that there is an increase in adsorption capacity of the sorbents with an increase in CO2 concentration in the flue gas resulting in higher access of amine sites which were not used up at lower CO2 concentrations. This observation is in line with the

Table 4 FTIR band assignments due to adsorbed CO2 species and ammonium ions. Wavenumbers (cm−1)

Functional groups

Ref

3730, 3625 3430–3435 3302, 3365 3010–3030 2360, 2349 2810–2980

Overtones of gas-phase CO2 NH stretching of carbamates NH2 suppression due to CO2 adsorption NH3+ stretching CO2 gas phase stretching CH2/CH3 suppression due to CO2 adsorption. Physisorbed CO2(v3) Zwitterions C]O carbamic acid C]O hydrogen bonded carbamic acid (dimer) NH3 + deformation (protonation by Si–OH) NH3+/NH2+ deformation (O]C]O)− stretching CH2 bending NCOO− skeletal vibration NCOO− skeletal vibration

[42] [32,43] [32,44] [12,44] [45] [32,44]

2267 2800-2050 1700 1680 1652 1620–1630, 1498 1560–1570, 1490–1500 1458 1410 1320–1325

[34] [32] [40,45,46] [32,40,46] [47] [32,45,48] [32,44,45] [19,44,45] [45] [32,47]

402



Fig. 5. IR absorbance comparison on different sorbents at a varying CO2 concentration at 10 min of CO2 adsorption.

be easily activated on amine sites at relatively lower temperatures.

reaction engineering principle, i.e., the rate of adsorption increases with an increase in concentration or partial pressure of the reactant. This indicates that on the amine sorbents at typical flue gas concentrations of 10–15% CO2, only a fraction of amine sites are used for CO2 adsorption. The sorbent still has a considerable amount of amine sites unused. Hence it is necessary to establish feasible alternative approaches to ensure that majority of the amine sites are utilised for CO2 adsorption which can improve the capture efficiency of the sorbents. The diffusional limitation of CO2 into the multilayers of amines was established in our earlier work, wherein the adsorption of CO2 on amines was observed in DRIFTS and ATR mode on thin films of TEPA with varying thickness [44]. The study revealed that the surface layer which is exposed to CO2 in the gas phase has broad bands due to close packing of amines resulted in the formation of zwitterion pairs. Whereas the layers underneath or bottom have limited access and hence stabilised as carbamates due to the transfer of protons from the CO2 adsorbed amine sites to the amine sites located at inner layers as observed in ATR mode [44]. The results so far indicate the lower amine loadings and higher CO2 concentrations are suitable for adsorption of CO2 as carbamic acid. Hence, at higher concentrations of CO2 and lower amine loadings, the following equation holds true and this suggests that CO2 molecules can

RNH2 + CO2 —→ RNH-COOH Reaction 4: CO2 adsorption on amines sites at lower amine density and higher CO2 concentrations. Between carbamic acid and carbamate, the former can be desorbed easily from the surface with less energy requirement < 100 °C. At these reaction conditions, the introduction of hydrogen in to the reaction mixture would force the formation of formic acid or methanol which requires further investigation. Alternatively, the change in pressure in the reactor could influence the adsorbed CO2 species which is discussed in the next section. 3.4. Effect of pressure on CO2 adsorption on sorbents The current work also investigated the effect of overall pressure on the CO2 adsorption using in situ FTIR experiments. The increase in CO2 concentration is directly related to the partial pressure according to ideal gas law PV = nRT P= (n/V) RT 403



Fig. 6. Schematic representation of CO2 adsorption with varying amine loading and CO2 concentrations on sorbents.

peaks. This can be attributed to the increase in amine site density which resulted in carbamic functional groups in proximity to be stabilised via hydrogen bonding. The shift in the OeH stretching vibration from 3450 to 3435 cm-1 confirms the presence of hydrogen bonding in carbamic acid functional groups PEI10 sorbent with multiple amine functionalities on each molecule and its three-dimensional structure, it is highly likely the adsorbed CO2 is stabilised through carbamate formation. With the increase in the pressure in the reaction chamber, the spectra showed an increase in the intensity of carbamate functional groups which in turn suggests the increase in CO2 adsorption on the sorbent. PEI25 sorbent with a higher number of functional groups also showed an increase in the formation of ammonium ions and carbamate functional groups with an increase in pressure. In other words, since the adsorbed CO2 were stabilised amine sites as carbamates during the equilibration during initial 10 min, the amines are not free to adsorb CO2 to form carbamic acid. However, increase in pressure allowed to access the amine sites which are not occupied at atmospheric pressure thus increasing overall capture capacity of the sorbents. The change in intensity of ammonium ion, carbamic acid and carbamate peaks with a change in pressure in the reaction cell during CO2 adsorption are shown in Figure S3 of supplementary information. The sorbents were already saturated with 15% CO2 after which the cell pressure is increased at 100 kPa intervals and equilibrated. The results on low amine loading sorbents APTS and PEI10 show a gradual increase in carbamate peaks until 200 kPa suggesting that increase in pressure allowed diffusion of gas molecules in to the layers and increased the access of amine sites for CO2 adsorption. Beyond 200 kPa there is a gradual increase in carbamic acid intensity indicating that increase in pressure will not only enhance the capture capacity by accessing the free amine sites but also adsorbing as carbamic acid. On the other hand, the sorbents with high amine loading Triamine showed a gradual and marginal increase of both carbamate and carbamic acid with pressure. The proximity of amines would help the adsorbed CO2 to stabilize more as carbamates. Interestingly, on PEI25 sorbent the change in intensity is marginal which is due to the large three-dimensional structure of PEI and high formation of multilayers of amine and highly viscous nature of PEI preventing the diffusion of CO2 in to the inner layers. The results from a change in CO2 concentration and overall pressure suggests that the solid amine sorbents used for CO2 capture have not

P = C RT where C = moles/volume i.e n/V. The above set of equation suggests that an increase in the concentration of CO2 is directly related to change in partial pressure of CO2 in the reaction chamber and should observe a similar effect as increasing the CO2 concentration of the flue gas. To verify the above hypothesis, in situ FTIR measurements are conducted with increasing the overall pressure of the DRIFTS chamber with 15% CO2. The sample was saturated with 15% CO2 for 10 min after which the pressure in the reaction chamber is increased from 100 to 500 kPa at an increment of 100 kpa and equilibrate for 5 min at each pressure. The spectral changes due to increase in the pressure in the DRIFTS chamber are recorded. Fig. 8 shows the change in the spectra of the sample with an increase in overall pressure in the reaction chamber on various sorbents. It is observed that with an increase in the cell pressure there is an increase in CO2 gas phase peaks at 3700-3600 cm−1, and 2350 cm-1, suggesting an expected increase in the overall CO2 partial pressure in the reaction chamber. Interestingly, increase in pressure of CO2 also observed some changes in the gas phase CO2 peak at 2368, 2328 cm-1 a doublet peak is associated with a shoulder peak at 2291 cm-1. This observation is more prominent in Triamine sorbent. These peaks are assigned to v3 hot band due to Lewis acid –base interactions between amine sites and CO2 [49]. These observations are also be attributed to the interaction of CO2 with the carbonyl groups formed such as carbamate and carbamic acid in the present study. However, this observation and its effect on CO2 adsorption capacity of the sorbents require further investigation. The spectra of APTS sorbent shows stronger depression in peak intensities of amine stretching vibrations at 3365 and 3302 cm−1 with pressure suggesting an increased adsorption of CO2 on amine sites. Similarly, the IR peaks due to carbamates at 1560-1320 cm−1 showed peak broadening and raise of IR bands at 1710 - 1675 cm−1 suggesting the formation of hydrogen-bonded and free carbamic acid functional groups on the sorbents. With the increase in carbamic acids, in proximity, they can be stabilised via hydrogen bonding with adjacent carbamic acids as shown in Fig. 6. Similar observations can be seen on Triamine sorbents with an exception that hydrogen-bonded carbamic acid formation is increased without any increase in free carbamic acid

404



Fig. 7. IR absorbance intensity of ammonium ion at 1700-1680 cm−1 during CO2 adsorption on various sorbents.

addition reactions for the Claisen condensation reaction [35]. Our next studies will be focused on the introduction of H2 in to the reaction mixture both in presence and absence of humidity for the in-situ production of chemicals from CO2 adsorption showing an alternative path for the thousands of solid-amine adsorbents researched so as catalysts for effective chemical production using the chemistry so far understood.

been used to their full potential with most amines sites not accessed at all during the adsorption process. Although the change in the pressure in the reaction chamber has increased the carbamic acid formation it is not as prominent as observed during an increase in CO2 concentration at atmospheric pressure. The likely reason for this observation is that the adsorbed CO2 was stabilized as carbamate before the reaction cell is being pressurized leaving less room for a carbamic acid formation which can be only formed by reversing the carbamate formation back to carbamic acid as discussed in Reaction 1. The close packing of amines will lead to decrease in CO2 adsorption as the adsorbed carbamic acid will stabilise as carbamate by transferring a proton to free amine sites present in the sublayers. Hence, amines dispersed on large surface area supports will allow the spacing between the amine sites, wherein not only all the amine sites are easily accessible but can be adsorbed as carbamic acid which can increase the efficiency of the sorbents. This observation is equivalent to having optimal metal loading on metal supported on various supports beyond which the metal crystallite size increases and leading to lower activity of the catalysts. It is also important to note that prolonged exposure of CO2 to amines leads to the formation of urea deactivating the sorbent. This can be prevented by the introduction of humidity in to the flue gas mixture which can reverse the formation of urea as reported [41]. In one of our previous studies, amine-based sorbents are used as catalysts for C-C

4. Conclusions The effect of CO2 concentration and overall pressure during the CO2 adsorption of differently loaded amine/SBA-15 sorbents are investigated. Amine-based sorbents were successfully prepared by grafting and impregnation onto the SBA-15 support as confirmed by TGA and FTIR. The thermogravimetric analysis establishes that PEI sorbents had higher adsorption capacities at 75 °C than at 50 °C due to the reduction of the viscosity of the three-dimensional molecules and allowing higher access of CO2 to the amine sites. The results also showed that the adsorption capacity of the sorbent increased linearly with CO2 concentration. FTIR studies showed an increase of carbamic acid IR peaks with a CO2 concentration in the gas mixture on low amine density APTS sorbent. The increase in carbamic acid formation is attributed to the faster saturation of amine sites with CO2 at highly concentrated CO2 inflow which did not allow the proton transfer to adjacent amine sites resulting 405



Fig. 8. Effect of pressure on adsorbed CO2 species on various sorbents.

Acknowledgement

in stabilisation of CO2 in carbamic acid form rather than the carbamate. At higher amine loadings the increase in carbamic acid formation is also observed, but smaller compared to APTS sorbents. This phenomenon is attributed to the availability of amines sites at inner layers for proton transfer stabilising them as carbamates. However, at both low and high amine loadings the increase in capture capacity was observed suggesting free amine sites are available even after adsorption with typical flue gas concentrations of 15% CO2. FTIR study on the effect of overall pressure in the reaction chamber showed an increase in carbamic acid formation on APTS and Triamine sorbents, whereas increased carbamate formation on PEI10 and PEI25 sorbents was observed. Thus, it can be concluded that the increase in partial pressure or overall pressure of CO2 the adsorption capacity of the amine sorbents can be enhanced. The increase in adsorption capacity is due to the access of free amines sites present in interlayers and increased adsorption of CO2 as carbamic acid which utilises one amine site instead of two sites in case of carbamate. The change in the CO2 adsorbed species with altering the reaction conditions can lead to the production of chemicals by the introduction of H2 in to the reaction mixture, amine-based sorbents as CO2 activation sites in the catalysts at relatively low temperature compared to conventional catalysts. However, the hypothesis requires further investigation and the work is underway.

Srikanth C.S. would like to thank the Department of Chemical Engineering, Monash University for the financial support. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jcou.2018.05.028. References [1] Q. Liu, A. Mace, Z. Bacsik, J. Sun, A. Laaksonen, N. Hedin, Chem. Commun. 46 (2010) 4502–4504. [2] Z. Liang, M. Marshall, A.L. Chaffee, Energy Procedia 1 (2009) 1265–1271. [3] Z. Liang, M. Marshall, A.L. Chaffee, Energy Fuel 23 (2009) 2785–2789. [4] Z. Liang, M. Marshall, A.L. Chaffee, Microporous Mesoporous Mater. 132 (2010) 305–310. [5] A.J. Sexton, G.T. Rochelle, Ind. Eng. Chem. Res. 50 (2010) 667–673. [6] W. Böttinger, M. Maiwald, H. Hasse, Fluid Phase Equilib. 263 (2008) 131–143. [7] P. Singh, J.P.M. Niederer, G.F. Versteeg, Int. J. Greenh. Gas Control 1 (2007) 5–10. [8] B. Aziz, N. Hedin, Z. Bacsik, Microporous Mesoporous Mater. 159 (2012) 42–49. [9] X. Wang, V. Schwartz, J.C. Clark, X. Ma, S.H. Overbury, X. Xu, C. Song, J. Phys. Chem. C 113 (2009) 7260–7268. [10] X. Wang, X. Ma, V. Schwartz, J.C. Clark, S.H. Overbury, S. Zhao, X. Xu, C. Song, Phys. Chem. Chem. Phys. 14 (2012) 1485–1492. [11] L. Zhao, Z. Bacsik, N. Hedin, W. Wei, Y. Sun, M. Antonietti, M.-M. Titirici, ChemSusChem 3 (2010) 840–845.

406



[12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30]

[31] [32] [33] [34]

J. Tanthana, S.S.C. Chuang, ChemSusChem 3 (2010) 957–964. D. Wang, C. Sentorun-Shalaby, X. Ma, C. Song, Energy Fuel 25 (2011) 456–458. K. Robinson, A. McCluskey, M.I. Attalla, Chemphyschem 12 (2011) 1088–1099. B. Li, B. Jiang, D.J. Fauth, M.M.L. Gray, H.W. Pennline, G.A. Richards, Chem. Commun. 47 (2011) 1719–1721. A. Car, C. Stropnik, W. Yave, K.-V. Peinemann, Adv. Funct. Mater. 18 (2008) 2815–2823. A.S. Kovvali, K.K. Sirkar, Ind. Eng. Chem. Res. 40 (2001) 2502–2511. J. Huang, J. Zou, W.S.W. Ho, Ind. Eng. Chem. Res. 47 (2008) 1261–1267. A. Danon, P.C. Stair, E. Weitz, J. Phys. Chem. C 115 (2011) 11540–11549. X. Wang, X. Ma, V. Schwartz, J.C. Clark, S.H. Overbury, S. Zhao, X. Xu, C. Song, Phys. Chem. Chem. Phys. 14 (2012) 1485–1492. X. Xu, C. Song, J.M. Andresen, B.G. Miller, A.W. Scaroni, Energy Fuel 16 (2002) 1463–1469. R. Serna-Guerrero, Y. Belmabkhout, A. Sayari, Chem. Eng. J. (Amsterdam, Netherlands) 158 (2010) 513–519. Y. Belmabkhout, A. Sayari, Adsorption 15 (2009) 318–328. C.S. Srikanth, S.S.C. Chuang, ChemSusChem 5 (2012) 1435–1442. A. Goeppert, S. Meth, G.K.S. Prakash, G.A. Olah, Energy Environ. Sci. 3 (2010) 1949–1960. G. Zhang, P. Zhao, Y. Xu, J. Ind. Eng. Chem. 54 (2017) 59–68. P. Zhao, G. Zhang, Y. Sun, Y. Xu, Energy Fuel 31 (2017) 12508–12520. Q. Yu, Jdl.P. Delgado, R. Veneman, D.W. Brilman, Ind. Eng. Chem. Res. 56 (2017) 3259–3269. A. Arenillas, K.M. Smith, T.C. Drage, C.E. Snape, Fuel 84 (2005) 2204–2210. S. Gadipelli, Y. Lu, N.T. Skipper, T. Yildirim, Z. Guo, J. Mater. Chem. A 5 (2017) 17833–17840.

[35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49]

407

K.P. Battjes, A.M. Barolo, P. Dreyfuss, J. Adhes. Sci. Technol. 5 (1991) 785–799. C.S. Srikanth, S.S. Chuang, J. Phys. Chem. C 117 (2013) 9196–9205. N. Mittal, A. Samanta, P. Sarkar, R. Gupta, Energy Sci. Eng. 3 (2015) 207–220. Z. Bacsik, N. Ahlsten, A. Ziadi, G. Zhao, A.E. Garcia-Bennett, B. Martin-Matute, N. Hedin, Langmuir 27 (2011) 11118–11128. E. Silva, M.S. Chakravartula, S. Chuang, Top. Catal. 55 (2012) 580–586. A. Sayari, Y. Belmabkhout, E. Da’na, Langmuir 28 (2012) 4241–4247. M.W. Hahn, M. Steib, A. Jentys, J.A. Lercher, J. Phys. Chem. C 119 (2015) 4126–4135. Ms.L. Pinto, Ls. Mafra, J.M. Guil, Jo. Pires, Jo. Rocha, Chem. Mater. 23 (2011) 1387–1395. G. Zhang, P. Zhao, L. Hao, Y. Xu, J. CO2 Util. 24 (2018) 22–33. J.B. Bossa, F. Borget, F. Duvernay, P. Theulé, T. Chiavassa, J. Phys. Chem. A 112 (2008) 5113–5120. A. Sayari, Y. Belmabkhout, J. Am. Chem. Soc. 132 (2010) 6312–6314. Y. Hu, Z. Liu, J. Xu, Y. Huang, Y. Song, J. Am. Chem. Soc. 135 (2013) 9287–9290. Z. Bacsik, R. Atluri, A.E. Garcia-Bennett, N. Hedin, Langmuir 26 (2010) 10013–10024. W.C. Wilfong, C.S. Srikanth, S.S. Chuang, ACS Appl. Mater. Interfaces 6 (2014) 13617–13626. U. Tumuluri, M. Isenberg, C.-S. Tan, S.S. Chuang, Langmuir 30 (2014) 7405–7413. A. Gómez-Zavaglia, R. Fausto, Phys. Chem. Chem. Phys. 5 (2003) 3154–3161. M.W. Hahn, J. Jelic, E. Berger, K. Reuter, A. Jentys, J.A. Lercher, J. Phys. Chem. B 120 (2016) 1988–1995. X. Wang, X. Ma, S. Zhao, B. Wang, C. Song, Energy Environ. Sci. 2 (2009) 878–882. S.G. Kazarian, M.F. Vincent, F.V. Bright, C.L. Liotta, C.A. Eckert, J. Am. Chem. Soc. 118 (1996) 1729–1736.