Some researchers have proposed the use of alkaline and alkaline-earth ce­ramic oxides to prepare membranes that are able to separate CO2 at high temperatures via a different transport mechanism than those observed on porous membranes. Li2ZrO3 and Li4SiO4 based membranes are examples of the aforesaid. Permselectivity of CO2 through these membranes takes place not only due to the selective CO2 adsorption properties of ceramic phases but also via a mechanism of gas separation that involves the trans­port of CO32- and O2- ionic species through the electrolytes (carbonate-
metal oxide) phases formed by the reaction of the membrane with the CO2 [87-89].

Kawamura et al. [87] fabricated and characterized a membrane for CO2 separation at high temperatures. The membrane was made of lithi­um zirconate (Li2ZrO3), an alkaline ceramic oxide that reacts with CO2 to produced Li2CO3 and ZrO2. These two reaction products are electro­lyte materials produced in-situ when the membrane is exposed to the rich carbon dioxide atmosphere. The electrolytes formed thus are capable to transport both CO2 and O2 across the membrane via a dual ion conduc­tion mechanism. The prepared membrane exhibited a separation factor of

4.9 between CO2 and CH4 gas molecules at a temperature of 600 °C. The obtained separation factor is higher than the Knudsen diffusion limit, 0.6. Therefore, the results clearly suggest the potential use of this kind of mem­brane system for CO2 separation such as the case of CO2 removal from natural gas.

Yamaguchi et al. [88] investigated the concept of the dual-ion conduc­tion facilitated mechanism previously observed for the case of Li2ZrO3 membranes by focusing their efforts on the preparation of a CO2 permse­lective membrane based on lithium orthosilicate (Li4SiO4). The supported membrane was prepared via a dip coating technique by using Li4SiO4 sus­pensions. The coating process was repeated several times before impreg­nation of the membrane with a Li2CO3/K2CO3 carbonate mixture and final sintering at 750 °C. In this membrane system, Li4SiO4 reacts in-situ with CO2 to form Li2CO3 and Li2SiO3.

Gas separation studies were performed by using CO2/N2 mixtures as feed gas. The observed CO2 permeance values were about 1 x 10-8 mol m-2s-1Pa-1 in the temperature range of 525-625 °C. The CO2/N2 separation factor was estimated between four and six. Figure 3 shows a scheme of the dual-ion conduction mechanism explained as follows. In the feed side, carbon dioxide dissolves in the material and diffuses as carbonate ions through the molten carbonate electrolyte due to a concentration gradient. Then, in the downstream side of the membrane, the formation of gaseous CO2 implies the formation of oxygen ions which must diffuse back to the feed side across the membrane and apparently through the formed Li2SiO3 skeleton to obtain the charge balance.

FIGURE 3: Schematic representation of a membrane system for the CO2 separation via a dual-ion conduction mechanism.

Separation Factor (C02/N2)

o

00

P’co2> P”co2
Ceramic Oxide Phase
Vo	Vo
ch4 N2l
CO32
2
CO
cP
Molten Carbonate Phase

FIGURE 5: Schematic representation of a membrane system for the CO2 separation and SEM image of a ceramic oxide-carbonate dual-phase membrane

The proposed transport mechanisms supports the higher selectivity val­ues observed in the permeation test for both systems, Li2ZrO3 and Li4SiO4. Figure 4 shows the separation factor values (CO2/N2) obtained for differ­ent ceramic membranes described in the present report. The pure Knudsen value is written as baseline and separation factor of nonporous Li4SiO4 for comparison purposes. However, it is important to mention that the oxygen ion diffusion process is not totally clear. Indeed, there is no experimental study regarding the oxygen ionic conductivity properties of Li2SiO3 phase. On the other hand, pure ZrO2 exhibits poor bulk oxygen ion conductivity. In fact, good conduction properties are observed only in acceptor-doped ZrO2 based materials with oxygen vacancies being the predominant charge carriers [90]. Therefore, oxygen ion conduction through the membrane must be related to different transport paths, such as grain boundaries and interfacial regions formed between the ceramic and molten carbonate on the membrane.

More recently, the promising concept of ceramic oxide-carbonate dual­phase membranes has been proposed for carbon dioxide selective separa­tion at intermediate and high temperatures (450-900 °C) [91-97].

This concept involves the fabrication of nonporous membranes capable of selectively separating CO2 via its transport, as carbonate ions. Dual phase membranes are made of an oxygen ion conductive porous ceramic phase that hosts a molten carbonate phase. Rui et al. [98] proposed the CO2 separa­tion by the electrochemical conversion of CO2 molecules to carbonate ions (CO32-), which are subsequently transported across the membrane. Carbon­ate ionic species (CO32-) are formed by the surface reaction between CO2 and oxygen that comes from the ceramic oxide phase (feed side, Eq.(10)) and then transport of CO32- takes place through the molten carbonate.

CO2 + OOx ~ CO32- + VO" (10)

Once carbonate ions have reached the permeate side, molecular CO2 is released to the gas phase, delivering OOx species back to the ceramic ox­ide solid phase. This process takes place due to a chemical gradient of CO2 in the system (Figure 5). Here, it is important to emphasize that dual-phase membranes are nonporous and therefore exhibit high separation selectivi­ty as a result of the transport mechanism. Figure 5 also shows the SEM im­age of the cross section of a ceramic oxide-carbonate membrane prepared by pressing La06Sr04Co08Fe02O3-s powders and subsequent infiltration of the obtained porous ceramic (bright phase) with carbonate (dark phase).

Table 3 summarizes the different studies reported and certain advances that have been achieved so far regarding the dual-phase membrane con­cept. This table also includes the Li2ZrO3 and Li4SiO4 nonporous mem­branes previously described. Although the original reports do not clearly explain the operational mechanism [26-27], the dual-phase membrane concept gives a much better idea of the possible phenomenology involved [30,33,36].