DOE OSTI2017
Based on two issued US patent, US7947579B2 and US8187678B2 that were invented by our team member at the University of New Mexico and Sandia National Labs, using atomic layer deposition technique (ALD) and molecular self-assembly or templating technique, An ultra-thin CO 2 membrane with CO 2 permeance up to 1580 GPU and CO 2 /N 2 up to 62 has been achieved in phase I project, promising for cost-effective CO 2 separations. As the first step of the project, a robust membrane fabrication protocol was established. To address the defects in the membrane structure, ALD treatment on AO support, multiple dip-coating processes and slow calcination steps were adopted. Consequently, defects were no longer evident in the nanoporous support, which ensures a reliable procedure for lab-scale membrane fabrications as well as the high-fidelity experimental findings. The same procedures may also be helpful for future scaling-up protocols. The layer-by-layer membrane fabrication was started by using a bridged disilane BTEE as the ALD precursor, wherein the organic groups at the bridging site were expected to be aligned layer-by-layer to form Angstrom-channels for CO 2 transport. Using 2-level factorial experimental design, we have found that ALD reaction time and ALD temperature were important factors impacting the CO 2 permeance and selectivity. Longer reaction time and higher ALD temperature were found to be advantageous for better CO 2 permeance and selectivity, which is in agreement with our model that better-aligned CO 2 -channels formed by the layer-by-layer stacking of the organic ligands can be achieved by a slower “stacking” process (longer reaction time) and stronger self-alignment tendency (higher ALD temperature). This observation will be an important guidance for our future membrane designs. The CO 2 -channel templated by –(CH 2 ) 2 - in BTEE was found to be too tight to facilitate fast CO 2 transport, but larger channel diameter is in contradiction with higher CO 2 selectivity. Therefore, the size-exclusion mechanism by itself is not sufficient for effective CO 2 separation. For this reason, the chemistry of the CO 2 -channel has to be carefully tuned. An amine bridged-disilane BTMA was therefore used as ALD precursor for larger CO 2 channel and high CO 2 -affinity chemistry. For BMTA membrane with amine chemistry measured at 100C, the CO 2 permeance was 1030 GPU for 100-cycle-ALD sample with CO 2 /N 2 selectivity of 17, and 302 GPU for 200-cycle-ALD sample with CO 2 /N 2 selectivity of 44. The increase of CO 2 permeance and the CO 2 /N 2 selectivity at elevated temperature is an evidence showing that there is a preferential chemical interaction between CO 2 and –NH functional groups in the membrane, which facilitates selective transport of CO 2 through the membrane. Therefore, using larger organic ligands as the CO 2 -channel template and the introduction of –NH chemistry to the membrane structure were proved to be an effective strategy in achieving better CO 2 permeance and selectivity. Guided by the same strategy, BTMA membrane was irradiated with UV to remove the bridging ligands for larger CO-channel, followed by Zn 2+ and Ni 2+ chemistry modification. Results showed that Zn 2+ or Ni 2+ chemistries will both enhance the selective CO 2 transport thereby the CO 2 /N 2 selectivity, and increasing separation temperature will further improve the CO2 permeance and selectivity. For BTMA membrane with Ni 2+ chemistry measured at 100C, CO 2 permeance of 1580 GPU and CO 2 /N 2 of 62 were achieved, promising for scaled-up industrial applications. Further, same membrane was tested at simulated flue gas condition (with SO 2 at 60C), the CO 2 permeance was 920 GPU and the CO 2 /N 2 selectivity was 52, suggesting the membrane still worked well under simulated flue gas conditions, primarily due to the stable ceramic framework of this membrane.