Electrostatic Separation of Carbon Dioxide by Ionization in Bifurcation Flow

Carbon dioxide is one of the major green house gases as well as impurities in process gases used for various manufacturing industries. In the present work, our recently developed ionization separator (Ito et al., Ind. & Eng. Chem. Res., 42, 5617-5621, 2003) was applied to the separation of carbon dioxide from inert gases. As a result, it was found that carbon dioxide can be separated mostly in the form of anion although some fraction of carbon dioxide decomposes by the soft X-ray irradiation. The maximum efficiency of electrostatic separation of carbon dioxide was 14% when helium stream contains 2.4 ppm of carbon dioxide at the applied voltage of 600V and the separation efficiency was decreased with increase in the inlet concentration. The dependency of separation efficiency on the applied voltage was qualitatively explained by the separation model that accounted for the electrical migration, the generation and the neutralization of anions and cations formed from carbon dioxide.


Introduction
Reduction of carbon dioxide (CO 2 ) concentration in air and other gases has been of great concern in various fields; e.g., control of indoor air quality by ventilation (Nabinger et al., 1994, Persily, 1997), green house effect (Hansen et al., 1981), contamination control of process gases for semiconductor manufacturing (Briesacher et al., 1991).
Adsorption of CO 2 with molecular-sieves and other adsorbents has been a common method for reducing CO 2 concentration.However, it requires the replacement or regeneration of adsorbents for maintaining the adsorption performance and the high pressure drop across the packed bed is a great issue when considering the energy consumption for gas purification.Metallic getter alloys are also used for gas purification (Briesacher et al., 1991).Although this method removes trace impurities  Recently, behaviors of ions, radicals and molecular species have been studied for the purpose of gas cleaning and purification from gas stream by corona discharge (Ohkubo et al., 1994), surface discharge (Oda et al., 1997) and electron beam injection (Hirota et al., 1995), etc. However these methods using high discharge energy are not effective for lowering the concentration of CO 2 .
Ionization is the process where electrically neutral atoms or molecules acquire either positive or negative electrical charge.In aerosol researches, ions are most frequently used to charge aerosol particles for electrical mobility classification (Knutson and Whitby, 1975).Gaseous molecules in the atmosphere are ionized by irradiation with radioactive source, electrical discharge, and combustion, etc. Radioactive-ray irradiation or corona discharge initially generates primary positive ions and free electrons.The free electrons readily attach to electronegative species in air to form negative ions.
Since the primary positive and negative ions are unstable, secondary ionization due to ion-molecule reaction occurs by the collisions between ions and neutral species.The ion-molecule reaction is influenced by the external electric fields.
We have proposed a new ionization separation technique which utilizes selective ionization of contaminant species and electrostatic migration of ions in bifurcating flow.The contaminant species removed by this method are toluene (Ito et al., 2002a, Ito et al., 2003) and ethanol vapor (Ito et al., 2002b).By this method, the contaminant species that have lower ionization potential and higher proton affinity compared to carrier gas becomes positive ions by the irradiation with α-ray or soft X-ray, and are electrostatically separated from the carrier gas.Negatively charged impurities generated by electron attachment can also be separated from carrier gas by applying an electrical field (Tammon et al., 1995).However, most of the molecular species in the atmosphere are ionized into both polarities.For instance, CO 2 can take both forms of cation and anion, such as (CO 2 ) x + , (CO 2 ) x CO + , (CO 2 ) x O + , (CO 2 ) x C + , (CO 2 )O 2 + , (CO 2 ) x -, (CO 2 ) x O -(x=1,2,3,…) according to Alger and Rees (1976).Nevertheless, since the generation rate and concentration of each ion species formed from CO 2 are different (Stamatovic et al., 1985), it might be possible to separate CO 2 in a form of cation or anion by ionization followed by electrostatic separation.
In this paper, our ionization separation technique was applied to CO 2 .We studied the separation performance of CO 2 in various carrier gases and the influence of CO 2 concentration on the separation efficiency, and discussed the separation mechanisms via the ion-molecule reaction kinetics of CO 2 in bipolar ionization field.

Experimental apparatus and procedure
Fig. 1 shows the ionization separator, whose structure is basically the same as used previously (Ito et al., 2002a), except that the improvement was made to avoid contamination from structural   materials by exchanging metal (SUS316) and quartz glass for plastics.Flow containing CO 2 is split into two ones, while the flow is being irradiated with soft X-ray from photoionizer (Model L6941, Hamamatsu Photonics, Japan; energy: 3.0-9.5 keV) under an electric field.The soft X-ray emitter has a dose of 4.7 × 10 -6 Gy/s at a distance of 1 m away from the source.The production rate of primary ions in the ionization separator is estimated to be in the order of 10 16 m 3 /s when the generation of an ion pair requires the energy of 34 eV in air (Ito et al., 2003).In the separator, CO 2 molecules are positively and negatively ionized as a result of ion-molecule reactions, and migrate towards either cathode or anode by the applied electric field.The ionization potential, proton affinity and electron affinity of CO 2 compared to other gases are listed in Table 1.
Fig. 2 shows the experimental setup.High purity helium (>99.9999%) and nitrogen (>99.9999%) are used as a carrier gas.The carrier gas is mixed with standard gas of nitrogen containing CO 2 (0.103%) to obtain a given concentration, and introduced into the ionization separator at the flow rate of 1.0 L/min.Inlet flow is equally divided into two outlet ones.The outlet concentrations of CO 2 are measured by sampling the gas by auto-sampler (GL Sciences Inc.: GSS-5000AH) followed by the determination with a gas chromatograph with FID (Shimadzu, GC-17A), converting CO 2 into methane by a methanizer (GL Sciences Inc.: MT221).

Experimental Results
The decomposition of CO 2 may occur by the irradiation with soft X-ray in the ionization-separator. Hatherly and Codling (1995) reported the dissociative ionization of CO 2 by soft X-ray irradiation   In order to evaluate the separation performance of ionization-separator for CO 2 taking into account the decomposition during the separation, we defined the relative CO 2 concentration from the separator as C * = C/C V=0 , where C is the outlet concentration with soft X-ray irradiation and voltage application, and C V=0 is the outlet concentration with the irradiation without the application.Fig. 4 shows the change in CO 2 concentration at the grounded outlet of ionization-separator with the voltage application under the irradiation of soft X-ray.Helium flow containing 2.4 ppm CO 2 was introduced into the separator at the flow rate of 1.0 L/min.CO 2 concentration was measured at the outlet of grounded electrode while varying the voltage between +600V and -600V as shown in the figure .As seen in the figure, CO 2 concentration is slightly higher than the inlet one of 2.4 ppm when the ground electrode is anode.It is considerably lower than when it is cathode, and follows the change in applied voltage, indicating that the separation of CO 2 does occur even though the dissociation of CO 2 takes place.Furthermore, we can see from it that the migration of CO 2 in a form of anion is dominant in the separator whereas toluene vapor was separated in the form of cation (Ito et al., 2002a(Ito et al., , 2003)).
Fig. 5 shows the influence of applied voltage on the relative concentration of CO 2 at the inlet CO 2 concentrations of 2.4 and 4.4 ppm.The relative CO 2 concentration has a minimum and maximum at the voltage of 600V and -600V for both inlet concentrations.The voltage corresponds with that at which the maximum separation of toluene was attained in the previous work (Ito et al., 2003).Fig. 6 shows the relative CO 2 concentration at the anode and cathode as a function of inlet CO 2   concentration at the applied voltage of 600 V. Relative CO 2 concentration at the anode increases as the inlet one decreases, while that at the cathode decreases.The dependency of relative concentration on the inlet CO 2 one is similarly reported for toluene in our previous work (Ito et al., 2003).However, the dependency on the inlet concentration is less pronounced compared to toluene.
The influence of carrier gas on relative CO 2 concentration was studied by using helium and nitrogen.Fig. 7 compares the relative CO 2 concentrations helium and nitrogen at the inlet CO 2 one of 2.4 ppm.Relative CO 2 concentration in helium is higher than that in nitrogen because ion-molecule reaction of CO 2 was probably enhanced due to higher mobility of ions in helium than that in nitrogen.
The mobility of ions in helium is about three times the mobility in nitrogen (Bricard and Pradel, 1966).The ion currents in the ionization-separator for both carrier gases were measured with an electrometer, and are shown in Fig. 8 as a function of applied voltage.The ion current in nitrogen steeply increases with the voltage and reaches constant value at a voltage of 600V.On the contrary, the ion current in helium slightly changes with the applied voltage, and is only one-third of the current in nitrogen.Because helium has a smaller proton affinity and a higher ionization potential compared to nitrogen, helium molecules are difficult to become cations compared to nitrogen ones.
The concentration of generated ions, N, was calculated by I = e κ N 2 AL (Liu and Pui, 1974), where e is the elementary unit of charge, κ is the recombination coefficient of ions (1.5×10 -12 m 3 /s in helium, 2.0 × 10 -12 m 3 /s in nitrogen, Glushchenko et al., 1988), AL is the effective volume in separator.The concentrations of ions calculated from the saturation current are 3.13 × 10 13 molecules/m 3 in helium and 4.66 × 10 13 molecules/m 3 in nitrogen.Since 1 ppm of CO 2 concentration corresponds to 2.5 × 10 19 molecules/m 3 , the number of generated ions measured by the electrometer is six orders of magnitude lower than that of CO 2 molecules.The following two explanations are plausible to explain the discrepancy between the ion concentration and CO 2 concentration: 1) the ion concentrations determined from the saturation current are not correct because many ionic species and electrons are involved in the transfer of electrical charge; 2) the CO 2 depletion mechanism is more complicated than just ionization-separation.

Discussion
The separation mechanism of ionization separator was discussed in our previous work by using a simple transport model, which employed one-dimensional convective diffusion equations for positively charged species and neutral species.In this paper, we applied the same separation model while including the transport of anions of CO 2 as well as the transport of cations.It was assumed that the net generation rates of CO 2 cations and anion are proportional to the number of neutral CO 2 molecules and that the depletion rates of CO 2 anions and cations are proportional to the numbers of CO 2 cations and anions without the decomposition of CO 2 .The mass balance of CO 2 molecules and CO 2 ions, as shown in Fig. 9, yields the following one-dimensional convective diffusion equations.
where C is the concentration, u is the gas flow velocity, D is the diffusion coefficient, α is the net depletion rate constant of CO 2 ions, β is the net generation rate constant of CO 2 ions, and Z and E are the electrical mobility of CO 2 ions and the electrical field strength, respectively.ZE is equal to the ion drift velocity, ZE = v.
The boundary conditions for the ionization-separator imposed for solving Eqs. 1, 2 and 3 are where L is the distance between the anode and the cathode.The following constraint for the mass balance is also adopted.The rate constants of α and β and electrical mobility of CO 2 ions are varied in the order of 10 2 to 10 3 referring to the previous work (Iinuma, 1991).We assigned the diffusion coefficient of CO 2 in helium at 293 K to be 6.2 × 10 -5 m 2 /s according to Poling et al. (2000).
The dimensionless concentration of CO 2 is calculated by solving Eqs. 1, 2 and 3 to see how the difference of electrical mobility between CO 2 cation and CO 2 anion influences CO 2 separation.Fig. 10 shows the dimensionless CO 2 concentration at the outlet as a function of applied voltage for various electrical mobilities while keeping the other parameters unchanged.No separation occurs when the mobilities of CO 2 anions and cations are the same (Z c = Z a ).However, when the mobility of cations is different from that of anions, which is conceivable, ionic species generated from CO 2 are transported to the cathode and anode at a different velocity, causing the separation of CO 2 with a maximum at a given applied voltage in all cases.When the electrical mobility of CO 2 cations is larger than that of anions (Z c > Z a ), CO 2 is concentrated at the cathode, whereas CO 2 concentration is higher at anode when the inequality is reversed (Z c < Z a ).Therefore, when the species to be separated are ionized in both polarities, the separation occurs by the difference between electrical mobility of cations and anions originated from the species.The extent of separation is more pronounced when the difference is larger.Furthermore, it is known from Fig. 10 that the decreases in cation's and anion's moblities lead to the increase in separation efficiency together with the shift of peak voltage toward higher one.
The calculated curves of dimensionless concentration are symmetric with respect to the origin point of V=0 and C*=1 for all mobilities although it was observed in Fig. 5 that the experimental curves are asymmetric with respect to the point.The asymmetry in the experimental curve of dimensionless concentration may result from the decomposition of CO 2 ions after separation because the decomposition of neutral CO 2 does not lead to the asymmetry in the concentration curves.
We also investigated the influence of net generation rate and net depletion rate constants when the electrical mobilities of cations and anions are the same.The results are shown in Fig. 11.The calculated curves are symmetric with respect to the origin point for various combinations of α and β.
Therefore, the asymmetry observed in the experimental curve as a function of applied voltage may result from the change in α and β with the applied voltage other than the decomposition of CO 2 .The larger net generation rate constant of CO 2 anions (compare curve i with ii), or the smaller net generation rate of cations (compare curve i with iv) causes the higher dimensionless outlet concentration.We also see from Fig. 12 that the increase in net depletion rate constant of CO 2 anions (compare curve ii with iii) or the decrease in net depletion rate constant of cations (compare curve iv with v) leads to the reduction in dimensionless concentration of CO 2 .
What follows from the calculations is that the separation efficiency of CO 2 is a function of the electrical mobilities of anions and cations and the depletion and generation rate constants of CO 2 ions, and that there exists an optimal voltage to separate CO 2 .Furthermore, the asymmetry in the dimensionless outlet concentration plotted as a function of applied voltage results from the changes in net generation and depletion rate constants with the voltage as well as the decomposition of CO 2 .
The experimental separation efficiency is no better than 15% as shown in Figs. 4 to 7. Therefore, further improvement is required in order to use the separator as a practical tool for CO 2 contamination control.The CO 2 separation efficiency can be improved by increasing CO 2 anion generation rate as well as avoiding the ion depletion.This might be achieved by using more intensive ionization source and improving the structure of separator so as to make the flow in it well-defined bifurcating one without stagnant regions in the separator.

Conclusions
The following findings are obtained through the application of ionization separator to CO 2 separation.i) CO 2 molecules in a helium and nitrogen can be separated mostly in the form of anion with the ionization separator although some fraction of CO 2 is decomposed by the soft X-ray irradiation.
ii) The CO 2 concentration relative to that without applied voltage increases with decreasing the inlet concentration as found in the separation of toluene of our previous works.
iii) Maximum relative concentration of CO 2 is 1.14 in helium carrier gas containing 2.4 ppm CO 2 at an applied voltage of 600V.
iv) The dependency of relative concentration on the applied voltage is qualitatively explained by the separation model that accounts for the transports of CO 2 cations and anions as well as neutral molecules together with the generation and depletion of these species.

Figure 1 .
Figure 1.Schematic of ionization-separator with Soft X-ray generator and wire mesh electrodes.

(Figure 3 .
Figure 3. Decomposition of CO 2 by soft X-ray irradiation in the ionization-separator. (a) Change in the CO 2 concentration with soft X-ray irradiation.(b) Outlet concentration with soft X-ray irradiation vs. outlet concentration without soft X-ray irradiation.

Figure 4 .
Figure 4. Change in the CO 2 concentration with applied voltage of +600V and -600V.

Figure 5 .
Figure 5. Change in the relative concentration of CO 2 as a function of applied voltage.

Figure 6 .
Figure 6.Influence of inlet CO 2 concentration on relative one with appliedvoltage of 600V.

Figure 7 .
Figure 7. Influence of carrier gas on relative CO 2 concentration.

Figure 8 .
Figure 8. Change in the ion current in ionization-separator with applied voltage.

Figure 9 .
Figure 9. Change in the ion current in ionization-separator with applied voltage.

Figure 10 .
Figure 10.Change in the calculated dimensionless concentration of CO 2 as a function of applied voltage at various electrical mobility of CO 2 ion.β a = β c = α a = α c = 500 s -1 .

Figure 11 .
Figure 11.Influence of net generation and depletion rate constant on calculated concentration.Z a = Z c = 1.4 m 2 V -1 s -1 .

Table 1 .
Ionization potential, proton affinity and electron affinity of various gases