Cross-Linking Amine-Rich Compounds into High Performing Selective CO 2 Absorbents

Transcription

1 Supplementary Information for Cross-Linking Amine-Rich Compounds into High Performing Selective CO 2 Absorbents Enrico Andreoli, Eoghan P. Dillon, Laurie Cullum, Lawrence B. Alemany, Andrew R. Barron* Correspondence to: arb@rice.edu This Supplementary Information includes: PEI-C Characterization CO 2 Absorption/Desorption at Atmospheric Pressure Effect of the Molecular Weight of PEI Identification of the Chemical Species Involved in the Absorption of CO 2 Figures S1 to S12 Tables S1 to S3 PEI-C characterization Thermogravimetric analysis: The results of the thermogravimetric analysis (TGA) of PEI-C are given in Figure S1. The absorbent was heated up from room temperature to 0 C at a rate of 2 C/min. During the scan the temperature was held at the two fixed values of and 3 C for 1 h and 12 h, respectively. This was done to allow the full completion of their corresponding processes which are the initial conditioning of the adsorbent in Ar and the decomposition of PEI in air, respectively. The weight decrease with time recorded during the second stage is given in the inset. 100 Weight (%) 74.0 % Weight (%) 35 3 o C / 12 h in Air Time (h).7 % 10 Ar Air 1.1 % Temperature ( C) Fig. S1. TGA characterization of PEI-C heated up from room temperature to 0 C at a scanning rate of 2 C/min. The temperature increase was halted and held at C for 1h and 3 C for 12 h. The inset shows the weight decrease associated with the second stage. After conditioning in Ar, the absorbent weight decreased to 74.0% of its initial value. At this point, air was allowed to flow in place of Ar. The weight of the absorbent dropped then to.7% after the PEI was decomposed at 3 C. A further drop to 1.1% was recorded on going from 3 to 6 C, corresponding to the decomposition of C. From these results, the portion of PEI is = 53.3% and that of C is = 19.6%. The PEI/C weight ratio can be thus calculated as 53.3/19.6. Of the 53.3 parts of PEI, 29.7 are of C and 17.3 of N, which, together with the C of C, give an overall C/N weight ratio of about 74/26. X-ray photoelectron spectroscopy: The X-ray photoelectron spectroscopy (XPS) survey and high resolution spectra of PEI-C conditioned in N 2 are presented in Figure S2. The two main elements detected are C and N, at about 285 and 0 ev, respectively. A third small signal at about 532 ev is related to the presence of a residual

2 amount of O. The C/N/O atomic concentration ratio of the three elements is 71.8/27.1/1.1, corresponding to a C/N weight ratio of about /. This value compares reasonably well with that obtained earlier for the bulk material with TGA, 74/26, the difference might be due to the fact that the XPS analysis is confined to a ~ 1 µm 3 volume portion on the surface of the absorbent. Fig. S2. XPS characterization of PEI-C after conditioning in N 2. The three peaks C1s, N1s and O1s of the survey spectrum are shown in high resolution in their respective insets. Surface area measurement: The surface area of PEI-C was characterized using N 2 adsorbate. In the relative pressure range of 5 to 0., two portions of the gas adsorption isotherm could be linearized using BET and Langmuir equations, as reported in Figure S3. The surface area of the absorbent is on the order of about m 2 /g. Fig. S3. Linearizations of the N 2 adsorption isotherm of PEI-C. The two linear fittings correspond to the BET and Langmuir models, respectively. Scanning electron microscopy: The scanning electron microscopy (SEM) images of as prepared PEI-C are given in Figure S4. The microporosity of the absorbent is evident at low magnification, Figure S4 (a), where free volume is present between small droplets and larger clots. The porosity is heterogeneously distributed with some regions characterized by pores smaller than a micrometer and others completely compact, as shown in the two images collected at higher magnification, Figures S4 (b) and (c). A close up of one of these last regions, Figure S4 (d), shows that at a scale below 1 µm the material is completely compact excluding the presence of mesopores. Fig. S4. SEM images of as prepared PEI-C showing the presence of non-homogeneously distributed microporosity (a, b and c). At the nanoscale level (d) the absorbent appears compact.

3 Atomic force microscopy: The results of the atomic force microscopy (AFM) characterization of PEI-C are shown in Figure S5. In this case the absorbent was probe sonicated in water and drop cast on a cleaved mica substrate as a suspension, due to the insolubility of PEI-C in solvents. As can be seen, the core of a large aggregate in the center of the image is composed of several small particles that are 1 nm in height, indicating that the C has been internalized and the PEI externalized in a micellar-like structure. The PEI spreads out on the surface due to the hydrophilicity of mica. Fig. S5. AFM image and sectional analysis of PEI-C showing internalized C and indicating a height for the central cluster of about 2 nm. Further evidence for the externalization of PEI comes from the SEM images of PEI-C in Figure S4. The SEM images reveal a highly interconnected network of polymer nodules and given that we used branched PEI with many primary amines, one can expect that one PEI chain would react with several nanocarbon molecules resulting in a highly interconnected network. CO 2 absorption/desorption at atmospheric pressure Thermogravimetric analysis: The CO 2 absorption performance of PEI-C at atmospheric pressure is shown in Figure S6. In this plot there are three gravimetric curves (lower) associated with the same temperature program (upper) used for each of the gravimetric curves. The gases used during the measurements are also indicated above the plots. Gas Ar 0.6 II CO 2 or N 2 wet CO 2 III Ar Temperature ( ο C) 0.5 Gas uptake (g/g) I dry CO wet N Fig. S6. Thermogravimetric characterization of the CO 2 capture performance of PEI-C. Initially, the weight of the absorbent falls with time while the temperature is increased in Ar. Then, the weight increases on exposure to wet or dry CO 2, or wet N 2. The absorption of CO 2 as itself, stage I, or with H 2 O, as the temperature is decreased, stage II, is followed by a last stage, III, relative to the regeneration of the absorbent in Ar. Three absorption conditions were investigated: wet CO 2 (red line), dry CO 2 (blue line), and wet N 2 (green line). Wet or dry gases were prepared at room temperature using bubblers filled either with deionized water or dry molecular sieves, respectively. Freshly prepared samples lost about g/g of their weight upon heating to C in Ar due to the release of moisture and adsorbed gases, i.e., CO 2. The carrier gas was then switched to either wet or dry CO 2 or wet N 2 and the weight increases determined while the temperature was lowered (Figure S6). Both wet and dry

4 CO 2 showed immediate absorption of 0.16 g/g and 0.14 g/g at C for wet and dry CO 2, respectively. In contrast, the sample exposed to wet N 2 showed no appreciable mass change until the temperature decreased below 75 C, at which point a maximum weight increase of 9 g/g was observed once the sample reached room temperature. The dry CO 2 sample shows no additional absorption with lowering of the temperature, but the wet CO 2 sample shows a second increase below 75 C to a maximum of about 0. g/g. It is easy to recognize, therefore, that the absorption of CO 2 went through two stages, I and II, in Figure S6. During stage I (between and 75 C) CO 2 is captured mainly in dry form (ca g/g), although a smaller portion associated with H 2 O (ca. 2 g/g) is observed. In stage II (75 C to room temperature) CO 2 is captured in combination with H 2 O and the total amount collectively absorbed during stage II can be calculated from the difference between wet and dry CO 2, i.e., = 0.46 g/g. If we consider a H 2 O uptake of 9 g/g, as from the wet N 2 absorption, the amount of CO 2 absorbed during stage II would be = 0.37 g/g. However, this value may be incorrect since the amount of H 2 O absorbed with N 2 or CO 2 could be different. This is evident in stage I where no H 2 O is absorbed with wet N 2 while it is with wet CO 2. A rationale for this difference could be that the driving force of H 2 O absorption for bicarbonate formation is greater than that of amine protonation. The exact amounts of absorbed CO 2 and H 2 O can be obtained from the elemental analysis of PEI-C conditioned in dry Ar and wet CO 2, as reported in Section 2.2. The regenerative performance of PEI-C is also presented in Figure S6. The portion of the wet CO 2 curve labeled as stage III shows that upon switching CO 2 to Ar the sample lost at room temperature % of the captured mass. Another % was lost below C. Furthermore, PEI-C maintained more than % of its starting absorption capacity after 100 absorption/desorption cycles at C, as shown in Figure S7. CO 2 uptake (g/g) CO 2 Abs. Perform. (%) Cycle Number Fig. S7. Thermogravimetric characterization of the CO 2 absorption/desorption performance of PEI-C at C. The first and last ten of a total of 100 cycles are presented. In the inset, the CO 2 absorption performance of the absorbent is shown as a function of the number of cycles. Elemental analysis: The total amounts of absorbed CO 2 and H 2 O can be calculated from the comparison of the elemental analyses of PEI-C conditioned in dry Ar and wet CO 2. The elemental analyses indicate that the chemical formula of PEI-C conditioned in dry Ar is CH 1.73 N 0.33, while that in wet CO 2 is CH 2. N 0.31 O The stoichiometric balance for the molecular components (C 2 H 5 N) x (C ) y (CO 2 ) z (H 2 O) w can be then calculated as x = 0.33 and y = 053 (z = w = 0), in dry Ar, and as x = 0.31, y = 0, z = 8 and w = 0.32, in wet CO 2. Details of the experimental results and calculations are given as follows. The PEI-C samples were conditioned in two environments: 1) dry Ar and 2) wet CO 2. The obtained conditioned samples were kept in containers filled with the same gas used for the conditioning and analyzed. The time per analysis was about min. Three measurements in sequence were performed per sample. The thermogravimetric profiles for the conditioning of PEI-C in dry Ar and wet CO 2 are reported in Figure S Ar 0.5 Gas uptake (g/g) dry Ar Temperature ( o C) Gas uptake (g/g) wet CO 2 Temperature ( o C) Fig. S8. Thermogravimetric profiles for the conditioning of PEI-C in dry Ar or wet CO 2.

5 The sample in dry Ar lost about % of its weight at the end of the conditioning, while that in wet CO 2 captured an overall amount of CO 2 +H 2 O of 0.55 g/g. The calibration lines for carbon, nitrogen and hydrogen obtained with acetanilide standard are presented in Figure S9. Response (mvs) k R 2 = k k 15k 10k 5k Response (mvs) 2.0k R 2 = k 1.0k Response (mvs) 10k R 2 = k 6k 4k 2k 0 0 Y = 74.95*X Y = 33.64*X Amount of C (mg) Amount of N (mg) Fig. S9. Calibration lines for carbon, nitrogen and hydrogen (R 2 > 0.999). Y = *X Amount of H (mg) The responses for the elemental analysis of PEI-C are given in Table S1. Sample weight (mg) Table S1. Responses for the elemental analyses of PEI-C conditioned in dry Ar and wet CO 2. dry Ar wet CO 2 Sample 1 Sample 2 Sample 3 Sample 1 Sample 2 Sample Element Response (mvs) Response (mvs) Response (mvs) Response (mvs) Response (mvs) Response (mvs) C N H These responses are converted to weights using the calibration lines. The weights are then converted into moles from which the stoichiometric ratios can be calculated as presented in Tables S2 and S3. From Table S2, the N/C stoichiometric ratio is reproducible among all samples, about The amount of H increases with time, most likely because of absorption of moisture since the samples were run in chronological order. For the same reason the difference between weighted and calculated weight (Diff%) also increases due to the oxygen of moisture. Table S2. Results from the elemental analysis of C -PEI conditioned in dry Ar. PEI-C dry Ar Sample mg C E H E N E mg Diff% 0.58 % Sample mg C E H E N E mg Diff% 4.41 % Sample mg C E H E N E mg Diff% 7.74 %

6 The sample most representative of PEI-C conditioned in dry Ar is the first less exposed to the air. The formula obtained for PEI-C conditioned in dry Ar is then C 1 H 1.73 N From this, it is possible to find the PEI/C weight ratio as follows: PEI = (aziridine) X = (C 2 H 5 N) X PEI-C = (C 2 H 5 N) X (C ) Y C 1 H 1.73 N 0.33 = (C 2 H 5 N) X (C ) Y = C 2X+Y H 5X N X 1 = 2X + Y ; 1.73 = 5X ; 0.33 = X X = 0.33, Y = 053 MW(C 2 H 5 N) = g/mol PEI = MW(C 2 H 5 N)*0.33 = g MW(C ) = 7.64 g/mol C = MW(C )*053 = 3.82 g PEI/C weight ratio = / 3.82 = 78.8 / From Table S3, the N/C stoichiometric ratio is reproducible among all samples, about 0.. The amount of H decreases with time, most likely because of desorption of captured H 2 O since the samples were run in chronological order. For the same reason the difference between weighted and calculated weight (Diff%) also decreases due to loss of captured gases. Note that these differences are due to the presence of oxygen. The amount of oxygen can be estimated as the difference between amount of sample taken and total amount of CHN measured. Table S3. Results from the elemental analysis of PEI-C conditioned in wet CO 2. PEI-C wet CO 2 Sample mg C E H E N E O = E mg Diff% % Sample mg C E H E N E O = E mg Diff% % % Sample mg C E H E N E O = E mg Diff% % The sample most representative of PEI-C conditioned in wet CO 2 is the first less exposed to the air. The formula obtained for C -PEI conditioned in wet CO 2 is then C 1 H 2. N 0.31 O From this, it is possible to find the absorbed amounts of CO 2 and H 2 O as follows: PEI = (aziridine) X = (C 2 H 5 N) X PEI-C + CO 2 + H 2 O = (C 2 H 5 N) X (C ) Y (CO 2 ) Z (H 2 O) W C 1 H 2. N 0.31 O 0.48 = (C 2 H 5 N) X (C ) Y (CO 2 ) Z (H 2 O) W = C 2X+Y+Z H 5X+2W N X O 2Z+W 1 = 2X + Y + Z ; 2. = 5X +2W ; 0.31 = X ; 0.48 = 2Z + W X = 0.31, W = (2.-5*0.31)/2 = 0.32, Z = ( )/2 = 8, Y = (1-2*0.31-8)/ = 0 X = 0.31, Y = 05, Z = 8, W = 0.32

7 MW(C 2 H 5 N) = g/mol PEI = MW(C 2 H 5 N)*0.31 = g MW(C ) = 7.64 g/mol C = MW(C )*0 = 3. g MW(CO 2 ) = g/mol CO 2 = MW(CO 2 )*8 = 3.52 g MW(H 2 O) = g/mol H 2 O = MW(H 2 O)*0.32 = 5.76 g PEI/C weight ratio = / 3. = 78.8 / The PEI/C weight ratio is exactly as that previously calculated for PEI-C conditioned in dry Ar. The CO 2 and H 2 O uptakes can be calculated as follows: Captured CO 2 = 3.52 / ( ) = 0.21 g/g Captured H 2 O = 5.76 / ( ) = 0.34 g/g The total captured mass is 0.55 g/g, the same as the amount measured with the TGA (Figure S8). The overall CO 2 :H 2 O stoichiometric ratio is 8:0.32 = 1:4 (as that for the CO 2 :N ratio). The N:H 2 O (aziridine:h 2 O) stoichiometric ratio is 0.31:0.32 = 1:1. The total CO 2 capture capacity of C -PEI is then 0.21 g/g corresponding to 4.8 mmol/g. It is interesting to note that PEI-C absorbs mainly H 2 O during stage II (Figure S6), but this does not affect the total amount of CO 2 captured in the presence of moisture in the gas feed. Also, the overall CO 2 :H 2 O molar ratio of 1:4 and a N:H 2 O one of 1:1. It seems that PEI-C absorbs H 2 O until each amine group in the absorbent is saturated with one molecule of H 2 O. Effect of the molecular weight of PEI We have previously shown with PEI-SWNT conjugates that the higher the PEI molecular weight the greater the uptake of CO 2 (19). Figure S10 shows the thermogravimetric curves for the absorption of wet CO 2 of PEI-C prepared with different PEIs, but using the same synthetic procedure and PEI/C weight ratio. It is clear that the overall absorption of CO 2 +H 2 O improves with increasing molecular weight. It is important to note that in the case of PEI(0)-C the adduct melts at the C. The best absorption performance is obtained with branched PEI 25,000 Da reaching 0.15 g/g uptake in min. at C and a total of CO 2 +H 2 O of 0.45 g/g going to 25 C in other 1 min. T ( o C) wet CO 2 Ar PEI0 PEI1,0 PEI10,000 PEI25,000 Gas uptake ( g/g) Fig. S10. Thermogravimetric characterization of the CO 2 absorption/desorption performance of PEI-C at C prepared from PEIs of different molecular weights. Identification of the chemical species involved in the absorption of CO 2 Nuclear magnetic resonance: The solid state 13 C and 15 N NMR spectra of the PEI-C conditioned in wet CO 2, dry CO 2, and wet N 2 are shown in Figure S11 (note that this figure is the same as Figure 4 of the main manuscript and it is reported here again for convenience). The 15 N spectra demonstrate the presence of carbamate. The 13 C spectra do not allow a secure differentiation of the carbamate carbonyl signal from the bicarbonate carbonyl signal that may also be present, as will now be discussed.

8 Fig. S11. Solid state (a) 1 H- 13 C and (b) 1 H- 15 N CP-MAS NMR characterization of PEI-C after conditioning in wet CO 2, dry CO 2 or N 2. Two distinctive bands are present in all 13 C CP-MAS NMR spectra, Figure S11 (a): a band with a peak maximum at about ppm from the sp 3 carbons of PEI and a weaker band with a peak maximum at about 1 ppm from the sp 2 carbons of functionalized C. A third sharper signal at 164 ppm is also evident in spectra for samples exposed to wet or dry CO 2 (but not in the spectrum exposed to wet N 2 ). Since this signal can be attributed to carbonate and/or carbamate species, we cannot readily determine the relative contributions of these two species in a MAS 13 C NMR spectrum of PEI-C conditioned in CO 2 ; this is not surprising in light of prior work. Ammonium bicarbonate gives a single, very sharp MAS 13 C NMR signal at ppm in agreement with a previous report [1]; the carbonyl peak position of bicarbonate is commonly observed at about 162 ppm in CO 2 -amine-h 2 O solutions [2]. In contrast, ammonium carbamate (H 2 NCO 2 NH 4 ) gives a pair of 13 C signals differing in peak height and linewidth in an MAS NMR spectrum obtained at a low magnetic field strength [1] because the 14 N quadrupole prevents complete averaging of the 13 C- 14 N dipolar coupling [3-5]. The effect becomes greater as the field strength decreases, and the separation of the two signals is field dependent, as demonstrated by the carbonyl lineshape of ammonium carbamate at 25 MHz 13 C [1] and at MHz 13 C (our work). These aspects significantly complicate analyzing the carbonyl region in an MAS spectrum if both bicarbonate and carbamate are present, as the bicarbonate signal may overlap part of the complex carbamate signal [1] since their chemical shifts are very similar (an aqueous solution of ammonium carbamate gives a signal at ppm [6-7], while solutions of anionic N-alkyl carbamates give a signal at ppm [8-12]). Two other factors cause additional difficulties in interpreting the signal at 164 ppm in the MAS 13 C NMR spectrum of PEI-C conditioned with CO 2. First, the distinctive, very sharp signal of NH 4 HCO 3 is significantly broadened when mixed with H 2 NCO 2 NH 4 [1] and thus is no longer as easily recognizable; the same effect would be expected if bicarbonate is present in our sample. Second, the carbamate carbonyl signal may be just a single peak, even at MHz 13 C, as shown by the MAS 13 C NMR spectrum of CO 2 adsorbed on PEI immobilized on a mesoporous silica support [13]. The half-height linewidth of this signal [13] is similar to that of the carbonyl signals in Figure S11 (a) with the same field strength and same line broadening used. Fortunately, CP-MAS 15 N NMR presents a much more secure way to determine the presence of carbamate in the presence of bicarbonate, and, so, the 15 N CP-MAS NMR spectra of the PEI-C conditioned in N 2 and dry CO 2 were also collected, Figure S11 (b). In the 15 N spectrum recorded after conditioning in dry CO 2, the band at about -347 ppm can reasonably be assigned to PEI amine nitrogen environments, while the signal at about -297 ppm can reasonably be assigned to PEI-NH-COO - carbamate species. In particular, the carbamate assignment is consistent with the 15 N chemical shifts reported for a CO 2 -treated, amine-modified, nanoporous clay [14] and with the 15 N chemical shifts of ammonium carbamate and other simple carbamates [15-18]. In the sample conditioned in N 2, the only appreciable signal, after more than,000 scans, was that of the PEI amine nitrogens. X-ray photoelectron spectroscopy: The XPS characterization of PEI-C conditioned in wet CO 2 (Figure S12) also supports the formation of bicarbonate and/or carbamate species. The C1s high resolution spectrum shows a peak at ev consistent with the presence of the sp 2 C=C bonds of C, while the two peaks at and ev are related to the sp 3 C-C and C-N bonds in PEI. A fourth smaller peak is also present at ev in agreement with the formation of bicarbonate and/or carbamate species both found in this binding energy region [19-]. The intensity of this peak cannot be related to the effective amount of CO 2 captured by PEI-C because most of the CO 2 is likely desorbed in the UHV chamber. The contributions at and ev in the O1s spectrum are consistent with the presence of two different CO 2 absorption species such as those of bicarbonate and carbamate. The other peak at ev is due to a residual trace of H 2 O present in the absorbent.

9 Fig. S12. High resolution (a) C1s and (b) O1s XPS characterization of PEI-C after conditioning in wet CO 2. References 1. X. Li, E. Hagaman, C. Tsouris, J. W. Lee, Energy Fuels 17, (03). 2. J. Y. Park, S. J. Yoon, H. Lee, Environ. Sci. Technol. 37, (03). 3. J. G. Hexem, M. H. Frey, S. J. Opella, J. Am. Chem. Soc. 103, (1981). 4. A. Naito, S. Ganapathy, C. A. McDowell, J. Chem. Phys. 74, (1981). 5. E. M. Menger, W. S. Veeman, J. Magn. Reson. 46, (1982). 6. N. Wen, M. H. Brooker, J. Phys. Chem. 99, (1995). 7. F. Mani, M. Peruzzini,; P. Stoppioni, Green Chem. 8, (06). 8. R. J. Hook, Ind. Eng. Chem. Res. 36, (1997). 9. A. Dibenedetto, M. Aresta, C. Fragale, M. Narracci, Green. Chem. 4, (02). 10. W. Böttinger, M. Maiwald, H. Hasse, Fluid Phase Equilib. 263, (08). 11. F. Barzagli, F. Mani, M. Peruzzini, Energy Environ. Sci. 2, (09). 12. G-j. Fan, A. G. H. Wee, R. Idem, P. Tontiwachwuthikul, Ind. Eng. Chem. Res. 48, (09). 13. T. C. Drage, A. Arenillas, K. M. Smith, C. E. Snape, Microporous Mesoporous Mater. 116, (08). 14. M. L. Pinto, L. Mafra, J. M. Guil, J. Pires, J. Rocha, Chem. Mater. 23, (11). 15. J. P. Warren, J. D. Roberts, J. Phys. Chem. 78, (1974). 16. J. M. Burns, M. E. Ashley, G. C. Crockett, T. H. Koch, J. Am. Chem. Soc. 99, (1977). 17. H. R. Kricheldorf, Org. Magn. Reson. 14, (19). 18. W. McGhee, D. Riley, K. Christ, Y. Pan, B. Parnas, J. Org. Chem., (1995). 19. I. George, P. Viel, C. Bureau, J. Suski, G. Lecayon, Surf. Interf. Anal. 24, (1996).. B. Luo, J. E. Rossini, W. L. Gladfelter, Langmuir 25, (09).