STEP—A Solar Chemical Process to End Anthropogenic Global

May 19, 2011 - Experimental support is presented of our STEP theory, which predicted a path to recycle and remove CO2 at high (∼50%) solar efficienc...
3 downloads 10 Views 8MB Size
Download PDF
ARTICLE pubs.acs.org/JPCC

STEP—A Solar Chemical Process to End Anthropogenic Global Warming. II: Experimental Results Stuart Licht,* Baohui Wang,† and Hongjun Wu Department of Chemistry, George Washington University, Washington, DC 20052, United States ABSTRACT: Alternative chemical processes are needed to decrease the level of atmospheric greenhouse gases. Experimental support is presented of our STEP theory, which predicted a path to recycle and remove CO2 at high (∼50%) solar efficiency. In STEP (solar thermal electrochemical production) of energetic molecules, electrolysis occurs, heated by excess and renewable thermal energy at high-temperature potentials below that of the room-temperature energy stored in the products. STEP is demonstrated in the efficient formation of metals, fuels, chlorine, and carbon capture. As one example, CO2 is converted to solid carbon, or CO, by distinguishing sunlight that is energy sufficient to drive photovoltaic charge transfer and applying all excess energy to heat and decrease the free energy of the enodothermic CO2 splitting electrolysis. The energy efficiency, based on chemical flow out to solar flow in, is high, at least 34%, and may reach over 50% but will depend on the extent of the solar thermal energy captured for the molten electrolysis. As another example, in an alternative to carbothermal iron production, which is responsible for 25% of worldwide industrial CO2 emissions, iron is produced without CO2 in solar thermal heated molten carbonate. Fe(III) (14 m) (as in the iron ore, hematite) is soluble in molten Li2O and Li2CO3, providing a facile path for STEP iron electrolysis. Also, water is efficiently STEP electrolyzed in molten salts to hydrogen. Finally, of relevance to bleach production and desalinization, chlorine, sodium, and magnesium are formed from the chlorides. STEP provides a different pathway for solar energy conversion, at high efficiency, and is capable of proactively removing carbon dioxide and also generating a range of useful chemicals without greenhouse gas emission.

’ INTRODUCTION One-third of the global industrial sector’s annual emission of 1  1010 metric tons of the greenhouse gas CO2 is released in the production of metals and chlorine.13 This together with the additional CO2 emissions for electrical generation, heating, and transportation comprise the majority of anthropogenic CO2 emissions. In the first paper in this series, an original process was derived for the renewable energy generation of energetically rich chemicals without greenhouse gas emission.4 STEP (solar thermal electrochemical production) of energetic molecules applies renewable and process thermal energy to decrease the necessary energy to drive endothermic electrolyses including carbon dioxide splitting and useful chemical syntheses (metals, fuels, chlorine, etc.). STEP originates from our theory and experiments derived for photoelectrochemical-driven water splitting.5 Solar-driven water splitting had been hampered by the high potential (.1.23 V = E°H2O(25 °C) necessary to generate the hydrogen (and O2) product.68 However, with increasing temperature, both the thermodynamic water electrolysis potential decreases and the kinetics of charge transfer improves (a decrease in overpotential).912 Hence, water may be split and hydrogen can be produced by a single external Si photovoltaic (band gap 1.1 V, maximum solar photopotential 0.7 V) driving electrolysis in molten hydroxides.10,11 The STEP theory generalized those original studies, from H2, to all endothermic electrolyses.4 r 2011 American Chemical Society

In this study, our recent brief letters and communications describing experimental STEP results for the efficient capture and conversion of CO2 and STEP iron production1315 and analysis16 are combined with additional experimental results on magnesium and chloride formation to present a broad base of experimental support for STEP theory. The general solar thermal electrochemical production of energetic molecules can proactively convert anthropogenic CO2 and eliminate CO2 emissions associated with industry production of chemicals.

’ EXPERIMENTAL SECTION Chemicals and Materials. Lithium carbonate, Li2CO3 (Alfa, 99%), ferric oxide, Fe2O3 (99.6%, JT Baker), Fe3O4 (Pflatz & Bauer 99.9%), iron wire (Alfa Aeasar), Ni foil (pure Ni 200 McMaster 9707K35), magnetite, platinum (Surepure Chemetals 99.9599.999% purity), inconel (McMaster Ni alloy 600), monel (McMaster Ni alloy 400), titanium (pure, McMaster grade 2), carbon (McMaster high-temperature conductive graphite rod), high-purity alumina crucible (AdValue Technology AL-2100), MgCl2 (Amresco, high-purity grade) were used as received. Received: December 11, 2010 Revised: April 24, 2011 Published: May 19, 2011 11803

dx.doi.org/10.1021/jp111781a | J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C Molten Electrolysis Cell Configuration. In the cyclic vol-

tammetry measurements, molten Li2CO3 or MgCl2 was contained in an alumina crucible. Initial experiments utilized platinum, to avoid materials with a tendency to corrode. The carbon dioxide flow into the molten Li2CO3 cell rate is maintained by an Omega FMA5508 mass flow controller. As described below, non-noble electrodes and alternate cell bodies are also effective. A high-purity alumina crucible is effective as a cell for the molten carbonate, but thermal expansion/contraction can cause cracking; when containing molten carbonate, a reaction to lithium aluminate is thermodynamically favored but was not observed, except under conditions of extended operation at 950 °C for high concentrations (9 m) of Li2O dissolved in molten lithium carbonate. In lieu of alumina, a cell body of monel is effective in air at temperatures up to 750 °C, beyond which it corrodes. A zirconium crucible (Alfa Aesar) as an electrolysis cell body is stable up to 950 °C. An electrolysis cell body of pure nickel is highly stable up to 900 °C and is stable in an argon gas environment to 950 °C. A cell body of stainless steel coats with a thin dark-gray overlayer in air at 950 °C; the coated stainless steel is moderately stable and nonreactive in contact with 850 °C molten Li2CO3. In Figure 3, anode measurements are conducted using the full cell potential by restricting the majority of overpotential effects to an 0.078 cm2 anode by using an oversized cathode (12.5 cm2). The opposite size (small cathode, large anode) electrodes are utilized in measurements which restrict the majority of overpotential effects to the cathode. In the measurements of Figure 6, each electrolysis chamber consists of a pure nickel body bored with a 4.5 cm diameter well, 9 cm deep, and filled to 7 cm with Li2CO3. The nickel body inner area (115 cm2) exposed to Li2CO3 serves as the anode and is extended by electrical coupling to an inner cylindrical electrode formed from pure nickel from foil (5.7 cm  15 cm = 85 cm2). The cathode is a cylindrical nickel from 6.5 cm  15.4  2 sides cm = 200 cm2 sandwiched between the inner and the outer components of the anode, separated by 0.20.4 cm. At higher temperatures, when CO, rather than solid C, is the main product, an alumina plate with holes between the electrodes minimizes the gasproduct interaction between the cathode and the anode. Determination of Salt Solubility in the Molten Carbonate. Low melting points of carbonates are achieved through a eutectic mix of alkali carbonates (Tmp = Li2CO3, 723 °C; Na2CO3, 851 °C; K2CO3, 891 °C; Li1.07Na0.93CO3, 499 °C; Li0.85Na0.61K0.54CO3, 393 °C17,18). Solubility is determined from multiple measurements including compositions, both approaching saturation, and also in compositions containing excess ferric oxide salts. When these salts are in excess in the molten mix, it is observed to descend to the bottom of the molten mix and is not removed with the molten liquid. The molten liquid is removed from the molten mix and analyzed. The extracted liquid is cooled, ground, and mixed with water. The carbonate component is water soluble, while the alkali iron oxide component is not and is removed by filtration and dried at 100 °C to determine the mass fraction of the ferric oxide salt that dissolved in the molten carbonate. LiFeO2 can be formed from Fe2O3 and Li2CO3 at 850 °C, produces carbon monoxide. On the left side of Figure 9 are plotted with the circle symbols our calculated thermodynamic electrolysis potentials for conversion of Fe2O3 or Fe3O4 to iron and O2 via eqs 3A, 3B, 4A, and 4B. The processes are each endothermic; the required electrolysis potential decreases with increasing temperature. The thermodynamic electrolysis potentials for Fe2O3 are ∼0.06 V lower than for Fe3O4, and in the temperature domain of molten Li2CO3, each is considerably below their respective rest potentials at 25 °C of 1.28 and 1.32 V. As seen in the figure, the experimental electrolysis potentials at a constant 0.050 A are similar for Fe2O3 and Fe3O4 at 800 °C but at T > 800 °C are lower for the Fe2O3 containing electrolyte. At high temperature it is seen on the left side of Figure 9 that the measured electrolysis potential can fall below the thermodynamic potential. The latter is calculated at unit activity, and the 11812

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C

ARTICLE

Figure 10. (Left) STEP iron solar thermal/solar electronic production. (Right) Example of HY-STEP solar thermal/wind electronic iron production. (Left) STEP iron production in which two molten carbonate electrolyses in series are driven by a concentrator photovoltaic. The 2.7 V maximum power of the CPV can drive either two 1.35 V iron electrolyses at 800 °C (schematically represented) or three 0.9 V iron electrolyses at 950 °C. The energy stored in the reduction of Fe2O3 to Fe at 25 °C is 1.28 V, eq 11. Hence, when 0.9 V is used to form iron at 950 °C, there is a considerable energy savings achieved through application of external heat, including solar thermal, to the system. (Right) Solar thermal/wind production of carbon dioxide free iron. In this Hy-STEP process concentrated sunlight heats and wind energy drives electronic transfer into the electrolysis chamber. The thermodynamic and kinetic energy required to electrolyze iron oxide to iron and oxygen is lower at higher temperature. The wind-powered electrolysis energy required is further diminished due to the high solubility of iron oxide in specific carbonates, such as Li2CO3. (Bottom) Iron is produced at high current density and low energy in 14 m Fe2O3 þ 14 m Li2O dissolved in molten Li2CO3.

former may reflect a Nernst, nonunit activity variation of the potential. The sharp drop in electrolysis potential occurring at T > 850 °C parallels our previous observation in iron-free molten Li2CO3, in which the main electrolysis product switches from solid C to CO at these temperatures. The measured full cell electrolysis potential, E, of Fe2O3 or Fe3O4 in Li2CO3 for a range of constant current densities, J, is presented in Table 1. J is determined by the cathode surface area. The subscript SA in J or E denotes a small (4-fold excess) and LA a large, (60-fold excess) surface area of anode. The measured potentials are similar with smooth Pt or iron cathodes and with anodes of smooth Pt or Ni (nickel oxide, prepared as McMaster 200 pure Ni sheet). E is stable during continuous electrolysis at this range of J. Comparing large and small relative anodes in Table 1, it is evident that anode (O2, rather iron) overpotentials dominate. E is less and J higher for textured electrodes or with increased mass transport (electrolyte convection or electrode movement/rotation). The low-voltage iron carbonate electrolysis presented in this study can be driven by conventional or renewable electricity, respectively, with diminished or no CO2 emission. The decrease in electrolysis potential with increasing temperature presents a low free energy opportunity for the STEP process. In this process, as shown for carbon capture in the previous section, solar thermal provides heat to decrease the electrolysis potential, Figure 9, and solar visible generates electronic charge to drive the electrolysis. A low-energy route for the CO2-free formation of iron metal from iron ores is accomplished by the synergistic use

of both visible and infrared sunlight. This provides high solar energy conversion efficiencies, Figure 2, when applied to eqs 10 and 11 in molten carbonate electrolytes. As described in the previous section for carbon capture, we use a 37% solar energy conversion efficient concentrator photovoltaic (CPV) as a convenient power source to drive the low electrolysis energy iron deposition without CO2 formation as schematically represented in Figure 10 in a lithium carbonate electrolyte, which also contains dissolved iron oxide. The endothermic nature of the new synthesis route provides an effective vehicle for the solar efficient, CO2-free, production of iron as a STEP process. A mechanism is determined for the unexpected solubility and low-energy electrolysis of iron oxides in lithium carbonate electrolytes. The facile charge transfer coupled with a sharp decrease in iron electrolysis potentials also provides a platform to demonstrate a new route for the hybrid solar thermal electrochemical production of iron (Hy-STEP). This route to produce iron is exemplified as driven by solar thermal and wind electronic energy, without release of CO2 in a process compatible with the predominant naturally occurring iron oxide ore, hematite, Fe2O3. Lithium oxide as well as iron, iron oxides, such as hematite, remain solid to a very high temperature (each has melting points above 1500 °C). However, as will be shown, solid Li2O dissolves in 4001000 °C molten carbonates and reacts rapidly with CO2(gas), in accord with the equilibrium in eq 3B. Here, rather than the prior parts per million reported solubility or our initial 20% by mass ferric oxide solubility in Figure 9 left, we report that 11813

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C

ARTICLE

Figure 11. (Left) Solubility of ferric oxides in alkali molten carbonates. (Right) Calculated electrolysis potentials of LiFe5O8, Fe2O3, or Li2CO3 at unit activity from the thermochemical data.9,10 Vertical arrows indicate Nernstian shifts at high or low Fe(III) solubilities.

higher Fe(III) solubilities, on the order of 50% in molten carbonates, are achieved via the reaction of Li2O with Fe2O3, yielding an effective method for CO2-free iron production. We prepare and probe the solubility and electrolysis of MFe5O8 and MFeO2 salts from Fe2O3. Collongues and Chaudron concluded that LiFeO2 and LiFe5O8 are the limiting members of the solid solutions formed by replacement of 2Fe2þ by LiþFe3þ in 2FeO or 2Fe3O4, respectively.48 The distinct solid products of the solid reaction of Fe2O3 and Li2CO3 have been recently characterized.49 Starting with either a 1:1 or 5:1 equivalent ratio of Fe2O3 to Li2CO3, the reactions proceed according to T > 400°C : Fe2 O3 þ Li2 CO3 Π2LiFeO2 þ CO2

ð12Þ

T > 600°C : Fe2 O3 þ 1=5Li2 CO3 Π2=5LiFe5 O8 þ 1=5CO2 ð13Þ Following preparation of specific iron oxide salts, we add them to molten alkali carbonate. The resultant Fe(III) solubility is similar when either LiFe5O8 or (Fe2O3 þ Li2O) is added to the Li2CO3. As seen in the left side of Figure 11, the solubility of Fe2O3 þ Li2O is high and increases from 7.1 m at 750 °C to 14.0 m (moles of Fe(III)/kg of carbonate solvent) in Li2CO3 at 950 °C. The measured solubility of Li2O in molten Li2CO3 (without added iron oxide salts, not shown) at 750, 800, 850, 900, or 950 °C, increases, respectively, from 9.1, 10.8, 12.0, 12.6, to 13.4 m. Upon supersaturation of Li2O, the molten Li2CO3 becomes viscous and excess solid Li2O is evident. The phase equilibrium of the FeNaO system has not yet been studied in detail, but as with the lithium analogues, reaction of Fe2O3 and Na2CO3 has been reported to produce both NaFeO2 and NaFe5O8 products.50 As seen in the left side of Figure 11, the solubility of the analogous sodium system NaFe5O8 in Na2CO3 remains substantially lower (,1 wt %) than lithiated ferric oxide even at 950 °C. Fe2O3, NaFe5O8, NaFeO2, KFe5O8, or KFeO2 solubility is insignificant ((,1 wt %) in either molten Na2CO3 or K2CO3. However, as seen, the solubility of (Li2O þ Fe2O3) in the alkali carbonate eutectic, Li0.87Na0.63K0.50CO3, is high and at 750 °C approaches one-half the solubility of that in the pure lithium carbonate electrolyte. The eutectic has the

distinction of a greater molten temperature range (extending several hundred degrees lower than the pure lithium system). The solubility of this lithiated ferric oxide in the LixNayKzCO3 mixes provides an alternative molten media for iron production, which compared to pure lithium carbonate has the disadvantage of lower conductivity14 but the advantage of even greater availability and a wider operating temperature domain. Also, as evident in the figure, while the solubility of the lithiated ferric oxide in the eutectic carbonate is high, the solubility of a mixed alkali pentairon octa-oxide salt, synthesized as Li0.435Na0.315K0.250Fe5O8, is low, reaching a maximum concentration of only 0.5 m Fe(III) at 950 °C. LiFe5O8 dissolves rapidly in molten Li2CO3; however, it reacts with the molten carbonate as evident in the mass loss of the solution. Similarly, Fe2O3, reacts with molten Li2CO3, resulting in a mass loss due to the evolution of CO2 proportional to the quantity of Fe2O3 contained in the Li2CO3 mix. As seen on the left side of Figure 12, over a wide range of temperatures and mass fractions of Fe2O3 in Li2CO3 the system proceeds to release 1 equiv of CO2 for each equivalent of Fe2O3 to form a steady-state concentration of LiFeO2 in accord with the reaction of eq 12 (but occurring in molten carbonate). Alternatively, dissolution of 1 equiv of Li2O and 1 equiv of Fe2O3 in molten Li2CO3 can drive the direct dissolution of LiFeO2 without the reactive formation of CO2 in accord with Fe2 O3 þ Li2 O f 2LiFeO2

ð14Þ

Addition of a 1:1 equivalent ratio of Li2O to Fe2O3 dissolves in Li2CO3 without the reactive formation of CO2 (CO2 forms when either pure Fe2O3 or LiFe5O8 are dissolved), equivalent to direct dissolution of LiFeO2 without CO2 evolution, and is summarized in eq 1. As seen in Figure 4, a CO2 atmosphere increases the molten carbonate electrolyte stability (moving equilibrium eq 3B to the left), and then the observed CO2 loss is very low from molten Li2CO3, even at high temperatures such as 950 °C. As seen on the left side of Figure 11, a CO2 atmosphere will diminish ferric solubility, although high concentrations of Fe(III) still remain soluble. Consistent with this observation, a high CO2 environment decreases free Li2O (eq 3A) and drives eq 14 to 11814

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C

ARTICLE

Figure 12. (Left) Thermogravimetric analysis of a mix of lithium carbonate with ferric oxide. Measured mass loss in time of Li2CO3. As indicated, the mixture is composed of either 20, 40, or 60 wt % of Fe2O3 in Li2CO3. The mass loss over time is measured at the indicated constant temperature of 650, 750, or 950 °C, corrected for CO2 evolution measured from the 100% Li2CO3 melt, then converted to moles of CO2, and finally normalized by the moles of Fe2O3 in the lithium carbonate ferric oxide mix. (Right) Thermogravimetric analysis of the moles of CO2 gained or lost after 5 h from molten Li2CO3 containing 3.5 m Fe2CO3 and 7 m Li2O.

the left, decreasing the concentration of the soluble LiFeO2 component. Equation 14 is of particular significance to the electrolysis of Fe2O3 in molten carbonate. As LiFeO2 is reduced to iron metal, Li2O is released (LiFeO2 f Fe þ 1/2Li2O þ 1/2O2), facilitating the continued dissolution of Fe2O3 without CO2 release and without a change in the electrolyte, or more concisely iron production via hematite in lithium carbonate is given by I and II I; dissolution in molten carbonate : Fe2 O3 þ Li2 O f 2LiFeO2

ð15Þ II, electrolysis, Li2 O regeneration : 2LiFeO2 f 2Fe þ Li2 O þ 3=2O2

ð16Þ iron production, Li2 O unchangedðI þ IIÞ : Fe2 O3 f 2Fe þ 3=2O2

ð17Þ As indicated in Figure 5, a molar excess of greater than 1:1 of Li2O to Fe2O3 in molten Li2CO3 will further inhibit the eq 1 disproportionation of lithium carbonate. The right side of Figure 9 summarizes the thermochemical calculated potentials constraining iron production in molten carbonate. Thermodynamically it is seen that at higher potential steel (iron containing carbon) may be directly formed via the concurrent reduction of CO2, which we observe in the lithium carbonate electrolyte at higher electrolysis potential, a process that will be delineated in an expanded study, as Li2CO3 f C þ Li2O þ O2, followed by carbonate regeneration via eqs 3A and 3B, to yield by electrolysis in molten carbonate steel production : Fe2 O3 þ 2xCO2 Π2FeCx þ ð3=2 þ 2xÞO2 ð18Þ From the kinetic perspective, a higher concentration of dissolved iron oxide improves mass transport to the cathode to sustain a larger source of iron available for reduction. This decreases the electrolysis overpotential and permits higher steady-state current densities of iron production. An increase of oxidized iron in the molten electrolyte solution will also substantially decrease the

thermodynamic energy needed for the reduction to iron metal. In the electrolyte Fe(III) originates from dissolved ferric oxides, such as LiFeO2 or LiFe5O8. The potential for the three-electron reduction to iron varies in accord with the general Nerstian expression for a concentration [Fe(III)] at activity coefficient R EFeðIII=0Þ ¼ E°FeðIII=0Þ þ ðRT=nFÞlogðRFeðIIIÞ ½FeðIIIÞÞ1=3 ð19Þ This decrease in electrolysis potential with increasing Fe(III) varies as 6.6  105 V  T(electrolysis, K)/K (=RT/nF), which is accentuated by high temperature. Hence, from the thermodynamic perspective with a simple assumption of unit activity coefficient, a 14 m iron concentration will decrease the 1223 K (950 °C) electrolysis potential by a further 0.1 V in eq 19. A higher activity coefficient, RFe(III) > 1, would further decrease the thermodynamic potential to produce iron. The measured electrolysis potential is presented on the right side of Figure 9 for dissolved Fe(III) in molten lithium carbonate and is low. For example, 0.7 V sustains a current density of 500 mA cm2 in 14 m Fe(III) in Li2CO3 at 950 °C, which is considerably less than the 1.6 V needed to drive the electrolysis at this current in 3 m Fe(III) at 800 °C. Higher temperature lowers the energy needed for electrolysis, and the higher concentrations accessed in this study result in an energy savings. As expected, the measured potentials are considerably less than the room-temperature 1.3 V thermodynamic potential required to convert Fe2O3 to iron and oxygen. When an external source of heat, such as solar thermal, is available the energy savings over roomtemperature iron electrolysis are considerable. Iron is deposited at high Coulombic efficiencies. It is interesting to note that the observed electrolysis potential at low current density is considerably smaller than the expected thermodynamic potential of 0.8 V at 950 °C. This is, at least in part, explained by the high, nonunit activity of the dissolved iron (14 m), providing significant voltage and energy savings. While a nickel anode appears stable during extended electrolyses, it is thermodynamically unstable. Initial anode results are presented in the STEP carbon-capture section. Other anode effects will be the focus of a future study, and we note in the right side of Figure 9 that when the nickel anode is replaced by a platinum electrode there is an observed increased in 11815

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C

ARTICLE

Figure 13. (Photograph lower left) Coiled platinum before (left) and after (right) MgCl2 electrolysis forming Mg metal on the cathode (shown) and evolving chlorine gas on the anode. (Main figure) Cathode size-restriced cyclic voltammetry of Pt electrodes in molten MgCl2 (cathode, 0.05 cm diameter, 2 cm length Pt wire; anode, 2.5  4.0 cm Pt foil). The oversized foil anode limits the majority of overpotential effects to the cathode. (Inset) Measured full cell potential during constant current electrolysis at 750 °C in molten MgCl2 at the same electrodes. Lower right: Thermodynamic and measured electrolysis potentials in molten MgCl2 as a function of temperature. Electrolysis potentials are calculated from the thermodynamic free energy components of the reactants and products as E = ΔG(reaction)/2F. Measured electrolysis potentials are stable values at 0.200 A and are measured between a 0.05 cm diameter, 16 cm long coiled Pt wire cathode and a 2.5  4.0 cm Pt foil anode.

the measured electrolysis potentials, particularly at high current density. The stability in time of the molten Li2CO3 electrolyte is consistent with eq 3B and may be regulated through control of the CO2 partial pressure above the electrolyte and/or through dissolution of excess Li2O. The right side of Figure 12 presents the stability after 5 h for 3.5 m Fe2O3 and 7 m Li2O (equivalent to 7 m LiFeO2 and 3.5 m Li2O) dissolved in a molten Li2CO3 electrolyte. In the absence of CO2 (an argon atmosphere over the electrolyte), the Li2CO3 loses mass through CO2 evolution. The rate of CO2 loss greatly diminishes in the open air (0.03% CO2), and the electrolyte gains mass (forms Li2CO3 through reaction of dissolved Li2O with external CO2) under an atmosphere of pure CO2. A new solar/wind hybrid solar thermal electrochemical production (Hy-STEP) iron electrolysis process is demonstrated. As seen in Figures 8 and 9, high temperature substantially lowers the

energy needed to drive the electrolysis. Hence, solar thermal heated electrolysis will increase the STEP system efficiency (left side of Figure 10). In lieu of solar electric, electronic energy can be provided by available alternative renewables, such as wind. As shown on the right side of Figure 10, in this Hy-STEP example the electronic energy is driven by a wind turbine and concentrated sunlight is only used to provide heat to decrease the energy required for iron splitting. In this process, sunlight is concentrated to provide effective heating but is not split into separate spectral regions as in our previous STEP system.13,14,16 Components we use in the Hy-STEP iron production are described in the Experimental Section, and iron production measured with one or two in-series 14 m Fe(III) molten Li2CO3 electrolysis cells is included in the lower right of Figure 10. Iron metal is produced. Steel (iron containing carbon) may be directly formed via the concurrent reduction of CO2, as will be delineated in an expanded study. Optimal matching of the maximum point power 11816

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C

ARTICLE

of the solar heat, wind electronic, and electrolysis cell components of the system is in progress will also be reported in a later study. STEP Chlorine and Magnesium Production (Chloride Electrolysis). The predominant salts in seawater (global average 3.5 ( 0.4% dissolved salt by mass) are NaCl (0.5 M) and MgCl2 (0.05 M). As we previously reported, the electrolysis potential for the industrial chloralkali reaction exhibits little variation with temperature, and hence, the conventional generation of chlorine by electrolysis would not benefit from inclusion of solar heating4 2NaCl þ 2H2 O f Cl2 þ H2 þ 2NaOH; E°ð25°CÞ ¼ 2:5 V

ð20Þ However, when confined to anhydrous chloride splitting, as exemplified in the lower right portion of Figure 13 for a magnesum chloride salt, the calculated potential for the anhydrous electrolysis of chloride salts is endothermic for the electrolyses, which generate a chlorine and metal product.4 Application of excess heat, as through the STEP process, decreases the energy of electrolysis and can improve the kinetics of charge tranfer for a range of chloride splitting processes MCln f n=2Cl2 þ M; E°MCln splitð25 °CÞ ¼ 3:98 V  M ¼ Na, 4:24 V  K, 3:98 V  Li, 3:07 V  Mg ð21Þ The thermodynamic electrolysis potential for conversion of NaCl to sodium and chlorine decreases from 3.24 V at the 801 °C melting point to 2.99 V at 1027 °C.4 Experimentally, at 850 °C in molten NaCl we observe the expected sustained generation of yellow-green chlorine gas at a platinum anode and of liquid sodium (mp 98 °C) at the cathode. Electrolysis of a second chloride salt, MgCl2, is also of particular interest. The magnesium as well as the chlorine electrolysis products are significant societal commodities. Magnesium metal, the third most commonly used metal, is generally produced by the reduction of calcium magnesium carbonates by ferrosilicons at high temperature,51 which releases substantial levels of carbon dioxide contributing to the anthropogenic greenhouse effect. However, traditionally, magnesium has also been produced by electrolysis of magnesium chloride using steel cathodes and graphite anodes, and alternative materials have been invesitgated.52 Of significance, here to the STEP process is the highly endothermic nature of anhydrous chloride electrolysis, such as for MgCl2 electrolysis, in which solar heat will also decrease the energy (voltage) needed for electrolysis. The rest potential for electrolysis of magnesium chloride decreases from 3.1 V at room temperature to 2.5 V at the 714 °C melting point. As seen in Figure 13, the calculated thermodynamic potential for electrolysis of magnesium chloride continues to decrease with increasing temperature to ∼2.3 V at 1000 °C. The 3.1 V energy stored in the magnesium and chlorine room-temperature products, when formed at 2.3 V, provides an energy savings of 35% if sufficient heat applied to the process can sustain this lower formation potential. Figure 13 also includes the experimental decrease in the MgCl2 electrolysis potential with increasing temperature in the lower right portion. In the top portion of the figure, the concurrent shift in the cyclic voltammogram is evident, decreasing the potential peak of magnesium formation with increasing temperature from 750 to 950 °C. Sustained electrolysis and generation of chlorine at the anode and magnesium at the cathode (Figure 13, photo inset) is evident at platinum electrodes. The measured potential during constant current electrolysis

at 750 °C in molten MgCl2 at the same electrodes is included as a figure inset. Note, the potential oscillations which occur during electrolysis. These correlate to temperature oscillations (controlled to within 10 °C in the cell). Higher temperature correlates to lower observed potentials. In the magnesium chloride electrolysis cell, nickel electrodes yield similar results to platinum and can readily be used to form larger electrodes. The nickel anode sustains extended chlorine evolution without evident deterioration; the nickel cathode may slowly alloy with deposited magnesium. The magnesium product forms as both the solid and the liquid (Mg mp 649 °C). The liquid magnesium is less dense than the electrolyte, floats upward, and eventually needs to be separated and removed to prevent an interelectrode short or to prevent reaction with chlorine that is evolved at the anode. In a scaled up cell configuration (not shown in Figure 13), a larger Ni cathode (a cylindrical nickel sheet (McMaster 9707K35) (6.5 cm  15  2 sides cm = 200 cm2) was employed, sandwiched between two coupled cylindrical Ni sheet anodes (total 200 cm2 of area across from the cathode) in a 250 mL alumina (Adavalue) crucible, and sustains multiamp large ampere currents. The potential at constant current is initially stable, but this cell configuration leads to electrical shorts, unless liquid magnesium is removed.

’ DISCUSSION The STEP theory arose from our studies to use sunlight to split water to hydrogen more efficiently by applying heat to tune the energetics of this process (tuning the electrochemical potential rather than the semiconductor band gap) and use of electrochemical redox processes to store solar energy.4,8 A synergy of solar photovoltaics and solar thermal energy capture opened a pathway to this alternative, higher efficiency solar energy conversion process. The approach was based on our theory and experimental observation that even a semiconductor with a band gap smaller than the water splitting potential (E(H2O) = 1.23 V at 25 °C), such as Si (Eg = 1.1 V), can split water at elevated temperature.10 STEP broadens this process, from solar hydrogen production, to the general production of all useful, energetic molecules formed by endothermic electrolyses. Experimentally, we demonstrated a sharp decrease in the electrolysis energy needed to split water in an unusual molten sodium hydroxide medium.911 STEP occurs at both higher electrolysis and higher solar conversion efficiencies than conventional room-temperature photovoltaic (PV) generation of hydrogen. Recently, we considered the economic viability of solar hydrogen fuel production. That study provided evidence that the STEP system is an economically viable solution for production of hydrogen.16 STEP occurs at both higher electrolysis and higher solar conversion efficiencies than conventional room-temperature photovoltaic (PV) generation of hydrogen. The study probed the economic viability, comparing four different systems: (1) 10% or (2) 14% flat plate PV-driven aqueous alkaline electrolysis H2 production, (3) 25% CPV-driven molten electrolysis H2 production, and (4) 35% CPV-driven solid oxide electrolysis H2 production. The molten and solid oxide electrolyzers are hightemperature systems that can make use of sunlight, normally discarded, for heating. This significantly increases system efficiency. Using levelized cost analysis, the study showed significant cost reduction using the STEP system. The total price per kilogram of hydrogen is shown to decrease from $5.74 to 11817

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C

Figure 14. Calculated potential needed to electrolyze carbon dioxide or water. The indicated decrease in electrolysis energy, with an increase in temperature, provides energy savings in the STEP process. High temperature is accessible through excess solar heat. Energies of electrolysis are calculated using thermochemical data at unit activity from NIST gas- and condensed-phase Shomate equations.28 Axes are extended to high temperature where ΔG at unit activity is zero.

$4.96 to $3.01 to $2.61 with the four alternative systems. As delineated in ref 16, the advanced STEP plant, which uses concentrated thermal, high-temperature electrolysis, and concentrator PVs requires less than one-seventh of the land area of the 10% flat cell plant due to the combined effects of the lower free energy needed for the high-temperature electrolyses and the higher efficiency of concentrator, rather than flat panel, PVs. To generate the 216 million kg of H2/year required by 1 million fuel cell vehicles, a 35% CPV-driven solid oxide electrolysis requires a plant only 9.6 mi2 in area. While PV and electrolysis components dominate the cost of conventional PV-generated hydrogen, they do not dominate the cost of the STEP-generated hydrogen. The lower cost of STEP hydrogen is driven by residual distribution (pipeline and gate) costs.16 STEP can be used to remove and convert carbon dioxide. As with water splitting, the electrolysis potential required for CO2 splitting falls rapidly with increasing temperature (Figure 14), and we demonstrate here (Figure 6) that a photovoltaic, converting solar to electronic energy at 37% efficiency and 2.7 V, may be used to drive three CO2 splitting lithium carbonate electrolysis cells, each operating at 0.9 V and each generating a 2-electron CO product. The energy of the CO product is 1.3 V (eq 1), even though generated by STEP electrolysis at only 0.9 V due to synergistic use of solar thermal energy. As seen in Figure 3, at lower temperature (750 °C, rather than 950 °C), carbon, rather than CO, is the preferred product, and this 4-electron reduction approaches 100% Faradaic efficiency. The CO2 STEP process consists of solar-driven and solar thermal-assisted CO2 electrolysis. Industrial environments provide opportunities to further enhance efficiencies; for example, fossil-fueled burner exhaust provides a source of relatively concentrated, hot CO2. The STEP product carbon may be stored or used, and the higher temperature STEP product carbon monoxide can be used to form a myriad of industrially relevant products including conversion to hydrocarbon fuels with hydrogen (which is generated by STEP water splitting911), such as smaller alkanes, dimethyl ether, or the FischerTropsch-generated

ARTICLE

middle-distillate range fuels of C11C18 hydrocarbons including synthetic jet, kerosene, and diesel fuels.53 The ability to remove CO2 from exhaust stacks or atmospheric sources provides a constructive response to anthropogenic CO2 emission. The high temperature of emitted CO2 in smokestacks is conducive to STEP carbon capture as an additional heat source. It is of interest whether material resources are sufficient to expand the process to substantially impact (decrease) atmospheric levels of carbon dioxide. The buildup of atmospheric CO2 levels from 280 to 392 ppm occurring over the industrial revolution comprises an increase of 1.9  1016 mol (8.2  1011 metric tons) of CO254 and will take a comparable effort to remove. It would be preferable if this effort results in useable, rather than sequestered, resources. We calculate below a scaled up STEP-capture process can remove and convert all excess atmospheric CO2 to carbon. In STEP, 6 kWh m2 of sunlight per day, at 500 suns on 1 m2 of 38% efficient CPV, will generate 420 kAh at 2.7 V to drive three series connected molten carbonate electrolysis cells to CO or two series connected series connected molten carbonate electrolysis cells to form solid carbon. This will capture 7.8  103 mol of CO2 day1 to form solid carbon (based on 420 kAh  2 series cells/4 Faraday mol1 CO2). The CO2 consumed per day is 3-fold higher to form the STEP carbon monoxide product (based on 3 series cells and 2 F mol1 CO2) in lieu of solid carbon. The material resources to decrease atmospheric carbon dioxide concentrations with STEP carbon capture appear to be reasonable. From the daily conversion rate of 7.8  103 mol of CO2 per square meter of CPV, the STEP-capture process scaled to 700 km2 of CPV operating for 10 years can remove and convert all of the increase of 1.9  1016 mol of atmospheric CO2 to solid carbon. A larger current density at the electrolysis electrodes will increase the required voltage and would increase the required area of CPVs. Contemporary concentrators, such as those based on plastic Fresnel or flat mirror technologies, are relatively inexpensive but may become a growing fraction of cost as concentration increases.55,56 A greater degree of solar concentration, for example, 2000 suns, rather than 500 suns, will proportionally decrease the quantity of required CPV to 175 km2, while the concentrator area will remain the same at 350 000 km2, equivalent to 4% of the area of the Sahara desert (which averages ∼6 kWh m2 of sunlight per day), to remove anthropogenic carbon dioxide in 10 years. A related resource question is whether there is sufficient lithium carbonate, as an electrolyte of choice for the STEP carbon-capture process, to decrease atmospheric levels of carbon dioxide. A 700 km2 CPV plant will generate 5  1013 A of electrolysis current and require ∼2 million metric tonnes of lithium carbonate, as calculated from a 2 kg/L density of lithium carbonate and assuming that improved, rather than flat, morphology electrodes will operate at 5 A/cm2 (1000 km2) in a cell of 1 mm thick. Thicker, or lower current density, cells will require proportionally more lithium carbonate. Fifty, rather than 10, years to return the atmosphere to preindustrial carbon dioxide levels will require proportionally less lithium carbonate. These values are viable within the current production of lithium carbonate. Lithium carbonate availability as a global resource has been under recent scrutiny to meet the growing lithium battery market. It has been estimated that the current global annual production of 0.13 million tonnes of LCE (lithium carbonate equivalents) will increase to 0.24 million tonnes by 2015.57 Potassium carbonate is substantially more available but 11818

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C

ARTICLE

Figure 15. Comparison of the industrial production of iron (left) and the STEP CO2-free production of iron (right) presented in this study. The new process can utilize renewable or nuclear power to drive iron formation and then is CO2 free. The new process alternatively can also be driven by fossil fuel electrical power, which as described in the text generates less CO2 than the carbothermic process.

as noted in the main portion of the paper can require higher carbon-capture electrolysis potentials than lithium carbonate. Iron smelting, the reduction of iron oxide ores with carbon, started as early as 3000 BCE and the Iron age began in the 12th century BCE with the collapse of the Bronze Age as shortages of tin or copper arose. Commercial iron today continues to be produced by this millennia old carbothermal process. In the carbothermal process iron oxide is reduced by carbon, via carbon monoxide or hydrogen as intermediate reductants. The carbothermal process releases the greenhouse gas, CO2, in the 3-electron reduction of Fe(III) accompanied by the 4-electron oxidation of carbon 2Fe2 O3 þ 3C f 4Fe þ 3CO2 ΔH 298 ¼ 466:4 kJ=mol ð22Þ This Fe2O3 reduction reaction is endothermic. To sustain this reaction, heat, ΔH, is provided by the burning of over one additional carbon, with concurrent release of further carbon dioxide C þ O2 f CO2 ΔH 298 ¼  393:5 kJ=mol

ð23Þ

From a thermodynamic perspective the carbothermal process releases over one CO2 per Fe formed (eqs 22 and 23 combined). In practice, many more than one CO2 per iron is released in the industrial formation of iron. Globally, the iron and steel industry accounts for about one-quarter of direct CO2 emissions from the industry sector, which total 27  109 tonnes of CO2 per year.58 Hence, 6.8  109 tonnes of CO2 is released in the annual production of 1.2  109 tonnes of iron and steel, from ∼2  109 tonnes of iron ore, resulting in the emission of 7 CO2 per molecule or mole of Fe formed.59 As demonstrated in Figures 8 and 9, iron may be formed at an electrolysis potential of as little as 0.6 V from hematite in molten Li2CO3, or from magnetite via eqs 10 and 11 in molten Li2CO3. The carbothermic and electrolysis component of the new STEP iron production process in this study are compared in Figure 15. Both STEP and Hy-STEP represent new solar energy conversion processes to produce energetic molecules. Individual components used in STEP are rapidly maturing technologies including wind electric60 and solar thermal technologies. An array of flat mirrors reflecting to a central tower, as demonstrated by Brightsource, can achieve temperatures of 550 °C;55 an individual parabolic heliostat can achieve temperatures over 800 °C;61 short

Table 2 CO2 emitted per iron generation process conventional smelting

Fe generated 7

new Li2CO3 electrolysis powered by fossil fuel electricity

2.5

STEP Li2CO3 (solar) electrolysis

0

focal length, plastic Fresnel generating optical concentrations of 5001000 suns have been deployed for concentrator photovoltaics, as demonstrated by Amonix,56 while better mirrors and secondary optics can achieve temperatures over 1500 °C.37,62 No CO2 is released in this iron production process when the heat and electronic charge are generated by renewable energy (solar, wind, hydro, geotheromal) or nuclear energy. Alternatively, we can calculate the CO2 release when fossil fuels are used to form electricity. As demonstrated in this study, iron may be formed at an electrolysis potential of as little as 0.9 V in molten as Li2CO3. In eq 10, the room-temperature rest potential (calculated from the free energy of the reaction) is 1.28 V and the thermoneutral potential (calculated from the enthalpy of the reaction) is 1.43 V. The latter voltage is the energy required to prevent the system from cooling during electrolysis and unlike the endothermic rest potential is nearly constant with changing temperature, for example, Ethermoneutral(1200 °C) = 1.40 V, and from eq 10 requires 6 Faraday per mol of iron. This is equivalent to 0.225 kWh/mol of Fe (from F = 96 485 A s, and 1 kW = 1000 VA) and will be less if an alternate heat source is used to maintain the system electrolysis temperature. Currently, fossil fuels release ∼11 mol of CO2/ kWh; specifically, the natural gas, oil, and coal generation of electricity have respective stack emissions of 7.5, 12, and 15 mol of CO2/kWh. Hence, even if fossil fuel, rather than renewable energy, is used to generate the heat and electricity and heat for iron by electrolysis, it will only emit 0.225  11 = 2.5 CO2 per Fe generated. This is less than the 7 CO2 per Fe emitted by the existing iron smelting processes.

’ SUMMARY Table 2 gives a brief summary. One salt source for the STEP generation of magnesium and chlorine from MgCl2 is via chlorides extracted from salt water, 11819

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C

Figure 16. Schematic representation of a separate (i) solar thermal and (ii) photovoltaic field to drive water purification, hydrogen generation, and endothermic electrolysis of the separated salts to useful products.

with the added advantage of generation of less saline water as a secondary product. In the absence of an effective heat exchanger, concentrator photovoltaics heat up to over 100 °C, which decreases cell performance. Heat exchange with the (nonilluminated side of) concentrator photovoltaics can vaporize seawater for desalinization and simultaneously prevent overheating of the CPV. The simple concentrator STEP mode (coupling superband gap electronic charge with solar thermal heat) is applicable when sunlight is sufficient to both generate electronic current for electrolysis and sustain the electrolysis temperature. In cases requiring separation of salts from aqueous solution followed by molten electrolysis of the salts, a single source of concentrated sunlight can be insufficient to both drive water desalinization and heat and drive electrolysis of the molten salts. Figure 16 presents a schematic representation of a separate (i) solar thermal and (ii) photovoltaic field to drive both desalinization and the endothermic electrolysis of the separated salts or water splitting to useful products. In the hybrid-solar thermal electrochemical production (Hy-STEP) process, separate thermal and electronic sources are used, respectively, to heat and drive energy-saving endothermic CO2-free electrolyses at high temperature. As illustrated, the separate thermal and electronic sources may each be driven by insolation, or alternatively, can be (i) solar thermal and (ii) (not illustrated) wind, wate,r, nuclear-, or geothermal-driven electronic transfer. An experimental demonstration of this Hy-STEP process is in progress and will be presented in a future study.

’ CONCLUSIONS To ameliorate the consequences of rising atmospheric carbon dioxide levels and its effect on global climate change, there is a drive to replace conventional fossil fuel-driven electrical production by renewable energy-driven electrical production. In addition to replacement of the fossil fuel economy by a renewable electrical economy, we suggest that a renewable chemical economy is also warranted. Solar energy can be efficiently used, as demonstrated with the STEP process, to directly and efficiently form the chemicals needed by society without carbon dioxide emission. Iron, a basic commodity, currently accounts for the release of one-quarter of worldwide CO2 emissions by industry. Replacement of the conventional method of iron production by STEP iron production will by itself decrease total industry emissions of carbon dioxide by 25%. STEP can be used in the

ARTICLE

efficient solar generation of hydrogen and other fuels, and STEP can also capture carbon dioxide and convert it to useful, higher energy products. In addition to removal of CO2, the STEP process is shown to be consistent with the efficient solar generation of a variety of metals as well as chlorine in place of conventional industrial processes. The unexpected solubility of iron oxides in lithium carbonate electrolytes, coupled with facile charge transfer and a sharp decrease in iron electrolysis potentials with increasing temperature, provides a new route for iron production. Iron is formed without an extensive release of CO2 in a process compatible with the predominant naturally occurring iron oxide ores, hematite, Fe2O3, and magnetite, Fe3O4. The endothermic nature of the new synthesis route provides an effective vehicle for the solar efficient, CO2-free, production of iron as a STEP process. In total, commodity production and fuel consumption processes are responsible for the majority of anthropogenic CO2 release, and their replacement by STEP processes provides a path to end the root cause of anthropogenic global warming as a transition beyond the fossil fuel, electrical, or hydrogen economy to a renewable chemical economy based on the direct formulation of the materials needed by society.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Present Addresses †

Northeast Petroleum University, Daqing, P. R. China.

Author Contributions

S. Licht designed the STEP and Hy-STEP process and together with B. Wang and H. Wu conducted the STEP and Hy-STEP solubility, electrochemical, and measurement analyses.

’ ACKNOWLEDGMENT We are grateful for the contributions of S. Ghosh, H. Ayub, and J. Ganley, who as described in ref 12 participated in the initial phases of experiments with molten eutectic rather than pure Li2CO3, Z. Zhang and H. Ayub for contributions reported in ref 15, and Z. Zhang who contributed to the Li2O solubility measurements. ’ REFERENCES (1) Pellegrino, J. L. Energy & Environmental Profile of the U.S. Chemical Industry, 2000; http://www1.eere.energy.gov/industry/chemicals/tools_ profile.html (2) Netherlands Environmental Assessment Agency, Global CO2 emissions, 2008; http://www.mnp.nl/en/publications/2008/ GlobalCO2emissionsthrough2007.html. (3) International Energy Agency, Tracking Industrial En. Efficiency & CO2, 2007, at: http://www.iea.org/Textbase/npsum/tracking2007SUM.pdf (4) Licht, S. J. Phys. Chem., C 2009, 113, 16283–16292. (5) Licht, S. Electrochem. Commun. 2002, 4, 37–38. (6) Fujishima, A.; Honda, K. Nature 1972, 238, 37–38. (7) Licht, S.; Wang, B.; Mukerji, S.; Soga, T.; Umeno, M.; Tributsch, H. J. Phys. Chem. B 2000, 104, 8920–8924. (8) In Semiconductor Electrodes and Photoelectrochemistry; Licht, S., Ed.; Bard, A., Series Ed.; Wiley-VCH: Weinheim, Germany, 2002; Vol. 6 (Encyclopedia on Electrochemistry). (9) Licht, S. J. Phys. Chem. B 2003, 107, 4253–4260. 11820

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821

The Journal of Physical Chemistry C (10) Licht, S.; Halperin, L.; Kalina, M.; Zidman, M.; Halperin, N. Chem. Commun. 2003, 2003, 3006–3007. (11) Licht, S. Chem. Commun. 2005, 2005, 4623–4646. (12) Licht, S. Solar thermal & solar thermal hybrid generation of hydrogen. In Solar Hydrogen and Nanotechnology; Wiley: New York, 2010; Chapter 21, pp 641664. (13) Licht, S.; Wang, B.; Ghosh, S.; Ayub, H.; Ganley, J. J. Phys. Chem. Lett. 2010, 1, 2363–2368. (14) Licht, S.; Wang, B. Chem. Commun. 2010, 46, 7004–7006. (15) Licht, S.; Wu, H.; Zhang, Z., Ayub, H. Chem. Commun. 2011, DOI: 10.1039/c0cc05581f. (16) Licht, S.; Chityat, O.; Bergmann, H.; Dick, A.; Ghosh, S.; Ayub, H. Int. J. Hydrogen Energy 2010, 35, 10867–10882. (17) Cairns, E. J.; MacDonald, D. I. Nature 1962, 194, 441–442. (18) Rolin, M. Bull. Soc. Chim. Fr. 1964, 2104–210. (19) Berbenni, V.; Marini, A.; Capsoni, D. Z. Naturforsch., Sect. A: J. Phys. Sci. 1998, 12, 997–1003. (20) Hoffman, A. Naturewiss 1938, 26, 431. (21) Widatallah, H.; Johnson, C.; Berry, F. J. Mater. Sci. 2002, 21, 4621–4625. (22) Yang, H.; Wang, Z.; Song, L.; Zhhao, M.; Wang, J.; Luo, H. J. Phys. D. Appl. Phys. 1996, 29, 2575–2578. (23) Ahniyaz, A.; Fujiwara, T.; Song, S.; Yoshimura, M. Solid State Ionics 2002, 151, 419–423. (24) Vanetsev, V.; Ivanov, V.; Tret’yakov, Y. Dokl. Chem. 2002, 387, 332–334. (25) Zakharchenko, N. Russ. J. Appl. Chem. 2001, 74, 229–234. (26) Braun, P. B. A. Nature 1952, 170, 1123. (27) Ohla, G.; Surya; P., Licht, S.; Jackson, N. Reversing Global Warming: Chemical Recycling and Utilization of CO 2. Report of the National Science Foundation sponsored 7-2008 Workshop; National Science Foundation: 2009; 17 pages; full report available at http://www.usc.edu/dept/chemistry/loker/ReversingGlobalWarming. pdf (28) Chase, M. W. J. Phys. Chem. Ref. Data 1998, 9, 1; data available at http://webbook.nist.gov/chemistry/form-ser.html (29) In Principles and Applications of Molten Salt Electrochemistry; Zhang, Z., Wang, Z., Eds.; Chemical Industry Press, Beijing, 2006; p 191. (30) Kojima, T.; Miyazaki, Y.; Nomura, K.; Tanimoto, K. Density. J. Electrochem. Soc. 2008, 155, F150–F156. (31) Sunmacher, K.; Kienle, A; Pexch, H. J.; Berndt, J. F.; Huppmman, G. Molten Carbonate Fuel Cells; Wiley-VCH: Weinheim, 2007. (32) Lehman, R. L.; Gentry, J. S.; Glumac, N. G. Thermochim. Acta 1998, 316, 1–9. (33) King, R. R.; Law, D. C.; Edmonson, K. M.; Fetzer, C. M.; Kinsey, G. S.; Yoon, H.; Sherif, R. A.; Karam, N. H. Appl. Phys. Lett. 2007, 90, 183516–183518. (34) Miller, J. E.; Allendorf, M. D.; Diver, R. B.; Evans, L. R.; Siegel, N. P.; Stuecker, J. N. J. Mater. Sci. 2008, 43, 4714–4728. (35) Woolerton, Y., W.; Sheard, S.; Reisner, E.; Pierce, E.; Ragsdale, S. W.; Armstrong, F. A. J. Am. Chem. Soc. 2010, 132, 2132–2133. (36) Benson, E.; Kubiak, C. P.; Sathrum, A. J.; Smieja, J. M. N. Chem. Soc. Rev. 2009, 38, 89–99. (37) Chueh, W. C.; Falter, C.; Abbott, N.; Scipio, D.; Furler, P.; Haile, S. M.; Stenfield, A. Science 2011, 330, 1797–1801. (38) Barton, E. E.; Rampulla, D. M.; Bocarsly, A. B. J. Am. Chem. Soc. 2008, 130, 6342–6344. (39) Andrieux, L.; Weiss, G. Compt. Rend. 1944, 217, 615. (40) Haarberg, G. M.; Kvalheim, E.; Rolseth, S.; Murakami, T.; Pietrzyk, S.; Wang, S. ECS Trans. 2007, 3, 341–345. (41) Wang, S.; Haarberg, G. M.; Kvalheim, E. J. Iron Res. Int. 2008, 15, 48–51. (42) Li, G. M.; Wang, D. H.; Chen, Z. J. Mater. Sci. Technol. 2009, 25, 767–771. (43) Yuan, B. Y.; Kongstein, O. E.; Haarberg, G. M. J. Electrochem. Soc. 2009, 156, D64–D69. (44) Palmaer, W.; Brinell, J. A. Chem. Metall. Eng. 1913, 11, 197. (45) Eustis, F. A. Chem. Metall. Eng. 1922, 27, 684.

ARTICLE

(46) Mostad, E.; Rolseth, S.; Thonstad, J. Hydrometallurgy 2008, 90, 213–220. (47) Qingeng, L.; Borum, F.; Petrushina, I.; Bjerrum, N. J. J. Electrochem, Soc. 1999, 146, 2449–2454. (48) Collongues, R.; Chaudron, G. Compt. Rend. 1950, 124, 143–145. (49) Wijayasinghe, A.; Bergman, B.; Lagergren, C. J. Electrochem. Soc. 2003, 150, A558. (50) Lykasov, A.; Pavlovskaya, M. Inorg. Mater. 2003, 39, 1088. (51) Li, H. Q.; Xie, S. S. J. Rare Earths 2005, 23, 606–610. (52) Demirci, G.; Karakaya., I. J. Alloys Compd. 2008, 465, 255–260. (53) Andrews, A.; Logan, J. FischerTropsch Fuels from Coal, Natural Gas, and Biomass: Background and Policy. Congressional Research Service Report for Congress, 2008, RL34133, March 27, 2008; available at http://assets.opencrs.com/rpts/RL34133_20080327.pdf (54) Tans, P. An accounting of the observed increase in oceanic and atmospheric CO2 and an outlook for the future. Oceanography 2009, 22, 26–35. (55) Parabolic solar concentrator technologies are described at stirlingenergy.com. (56) Fresnel solar concentrator technologies are described at amonix. com, energy innovations.com/sunflower. Parabolic solar concentrator technologies are described at stirlingenergy.com. (57) Tahil, W. The Trouble with Lithium 2; Under the Microscope; Meridan International Research, Martainsville, France, 2008; 54 pages. (58) Intl En. Agency, Tracking Industrial En. Efficiency & CO2, 2007, at http://www.iea.org/Textbase/npsum/tracking2007SUM.pdf. (59) http://www.issb.co.uk/global.html. (60) Barbier, E. How is the global green deal going?. Nature 2010, 464, 832–833. (61) Power tower solar technologies are described at brightsourceenergy.com, ausra.com, esolar.com, bengoasolar.com/corp/web/en/ our_projects/solana/. (62) Pitz-Paal, R. High Temperature Solar Concentrators. In Solar Energy Conversion and Photoenergy Systems Galvez, J. B., Rodriguez, S. M., Eds.; EOLSS Publishers: Oxford, UK, 2007.

11821

dx.doi.org/10.1021/jp111781a |J. Phys. Chem. C 2011, 115, 11803–11821