the transport and chemistry of copper in pbwer plants as determined by laboratorv...

THE TRANSPORT AND CHEMISTRY OF COPPER IN PbWER PLANTS AS DETERMINED BY LABORATORV E%PERlMENTS

Donald A. Palmer, Pascale BCnCzeth, and J. Michael Simonson Chemical and Analytical Sciences Division, Oak Ridge National Laboratory

Bldg. 45OOS, P.O. Box 2008, Oak Ridge TN 37831-6110, U.S:A.

Andrei Y. Petrov* Moscow Power Institute, 14 Krasnokazarmennaya, Moscow 111250, Russia

*visiting scientist:

ABSTRACT

Solubility measurements have been made involving cupric (CuO, tenorite) and cuprous (Cu,O, cuprite) oxide in contact with aqueous solutions of NaOH and NH, to 350°C. The solubility of CuO was also measured in steam containing NH, to 4OO’C using a flow technique. These results were reinforced by “static” measurements of the partitioning of copper hydroxide from a saturated aqueous phase to the steam phase. A quantitative thermodynamic model of the speciation of copper(I1) in the liquid phase was formulated as a function of pH and temperature, that established the dominant species to be Cu(OH),’ and Cu(OH),- under most boiler water conditions. The former, neutrally-charged specie provides the only significant contribution for vaporous carry over to the steam phase. The relative volatility of CUE’ is high and shows remarkably iittle dependence on temperature. The major release of copper into the steam exiting the boiler appears to occur at startup due to the combination of oxidizing conditions, the presence of ammonia, and a higher probability of liquid droplet entrainment. Limited measurements with copper metal exposed to basic (NaOH) solutions containing oxygen indicate that copper levels in excess of the solubility limits of CuO and Cu,O can be readily achieved. On-going experiments with Cu,O confirm that the soiubiIity is much more dependent on temperature and pH than was found for CuO, rising to very high concentrations at high pH and high temperature (e.g., ca. 1OOppb at the solubility minimum of Cu,O at 250°C compared with ca. 2pbb for CuO). The experimental techniques and methods for data analysis will be outlined, and the results compared with those from previous studies. The implications of these laboratory experiments to the chemistry of the water/steam cycle of fossil plants wiIl also be presented.

BACKGROUND

The goal of this research is to identify the chemistry that occurs when copper and its oxides are exposed to conditions in fossil-fired boiler from start-up to normal operation. A combination of solubility and volatility experiments were initiated to determine quantitatively the concentrations of copper in solution (including steam) under varying conditions of temperature and pH where the oxidation state is controlled either by trace oxygen levels or by the copper metal/cupric oxide oxidation/reduction couple. These measurements are carried out in “inert” apparatus, precluding interactions with the wet walls of the containment that would otherwise jeopardize the application of rigorous equilibrium models.

Previous Copper Oxide Solubility Studk

There have been numerous studies of the solubility of CuO to high temperatures [l-6], although the reported copper concentrations differ by as much as three orders of magnitude under basic conditions [7]. Moreover, the pH dependence of the solubility varies dramatically according to the various studies, so that the speciation of copper(U)

“The submitted manuscript has been authored by a contractor of the U.S. Government under contract No. DE-AC05-OOOR22725. Accordingly, the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or allow others to do so. for U.S. Government n~wnosnc I’

,

in solution is very different. This impacts the predicted stability of the Cu(OH),’ molecule, which is considered to be the only significant copper species that partitions to the steam phase [7], as borne out by the current research. Heam et al. [Z] reported solubilities of CuO into the supercritical region that are generally at least an order of magnitude higher than found previously by Pocock and Stewart [ 11. Heam et al. [2] attributed the lower values of the earlier work to possible lack of equilibration and/or adsorption of copper on the corroded surfaces of the containment vessel.

As indicated previously [7], the formation of complexes with anionic impurities (e.g., chloride, sulfate, carbonate, acetate) in the liquid phase is favored with increasing temperature. However, even ppm levels of these impurities are not expected to influence the solubility significantly at the relatively high pH conditions prevailing in a boiler. The possible exception being chloride complexation of Cu(1) that may increase the solubility of Cu,O significantly at boiler operating conditions. On the other hand, complexes of Cu(I1) with ammonia are stronger at low temperatures [S]; therefore during start-up in the presence of oxygen, ammonia could raise the concentration of copper in boiler water significantly (note that our work shows that beyond 200°C this effect can be virtually ignored). These general conclusions and observations have been supported recently by specific plant data and modeling efforts [9].

Finally, the effect of adsorption and surface precipitation of copper on metal oxide surfaces cannot be ignored as both mechanisms operate over wide pH ranges and the former is believed to be dominant even at high pH for hydrated iron oxide surfaces [lo]. Moreover, these experiments were carried out at 25°C and our recent investigation of surface charge and calcium adsorption on t-utile (TiO,) has shown that both increase dramatically with increasing temperature at least to 290°C [ Ill.

Solubility studies of Cu,O are very limited with apparently only one high temperature study being carried out in the liquid phase [ 121, although Pocock and Stewart [ I] performed a limited number of analyses of Cu(I) in superheated steam. Var’yash [ 121 measured solubilities in the liquid phase from 150 to 450°C over a wide range of pH, although the technique for determining the pH was not given [6]. These results predict that the solubility of Cu,O increases by more than six orders of magnitude from 25 to 350°C at the solubility minimum with respect to pH where Cu(OH)’ is the dominant Cu(I) specie in solution. The minima occur very near the pH of boiler water, viz., 9.2,at 25°C. If these results are valid, then not only would Cu,O be highly soluble under boiler operating conditions, but it should be also highly volatile. Cu(OH)’ should readily partition to the steam phase as this neutral molecule has some covalent character (cf. the case of HCl[13,14]).

Copper Dissolution and Transport in Water/Steam Cycles of Power Plants

Effective control of water-steam chemistry can eliminate the leading cause of reduced availability and high maintenance costs for fossil steam power plants, as well as improve plant heat rate and profitability. On the other hand, poor water-steam purity can result in corrosion damage and deposition on major components. It can also lead to boiler tube and turbine blade failures, flow-assisted corrosion (FAC) in feedwater piping systems, turbine efficiency losses, and other problems [ 15,161.

Uncontrolled copper transport represents a potentially significant source of performance and reliability loss to fossil plant units with mixed metallurgy feedwater systems. Recent utility experiences with severe copper turbine fouling and other related problems illustrate the need to improve the understanding in this area and to develop operating guidelines for controlling copper deposition around the cycle [16-191. To this end, Project Copper has been focused on: 1) establishing which of the two major copper oxides (CuO and Cu,O) grow in mixed-metallurgy feedwater systems at each typical operating condition of temperature, pressure, air in-leakage, oxygen scavenger, electrochemical potential, and pH; 2) how these copper.oxides are transported with feedwater into drum boilers; 3) how they become entrained in steam; and 4) determining their solubility limits in steam. In this context, the major aim of the research at ORNL is to quantify solubility limits for the major copper oxide forms in liquid water and steam, thereby supporting development of operating guidelines for controiling copper deposition [ 15,161.

.

‘4

EXPERIMENTAL METHODS

The experimental technique employed at ORNL for measuring the solute volatilities is described by Simonson and Palmer [ 131, although minor modifications were made subsequently [20]. Succinctly, an equilibrated two-phase aqueous solution is contained in a platinum alloy liner seated in a one-liter pressure vessel and samples of both phases are withdrawn under controlled conditions for analysis by ion chromatography, Zeeman atomic absorption spectroscopy, or chemiluminescence. However, in all volatility experiments conducted here, solid copper oxide was present in the bottom of the liner and therefore both the liquid and steam phases were saturated with respect to their copper content. The solutions contained dilute NaOH in order to control the pH and experiments were performed at 100 to 350°C.

Direct solubility measurements were also made in a “packed column” configuration, whereby dilute NH, or NaOH (or a combination of both solutions) was pumped via a preheater into a tubular, Zircalloy pressure vessel packed with CuO, Cu,O+Cu, or Cu powder and the outlet stream was sampled for analysis. For the CuO experiments, the feed solution was either purged with helium or saturated with CO,-free air, or saturated with pure oxygen, whereas for the Cu,O solubility runs, the feed was first purged with a hydrogen/argon mixture and kept under positive helium pressure during the sampling period. In this latter case, it was also found necessary to inject either 0.02 mola! HNO, or 0.03 molal NaOH solution into the outlet stream at the experimental temperature to prevent precipitation of copper oxide/hydroxide in the sampling line upon cooling. The temperature range of these experiments was 100 to 400°C, and the pressure was varied from 50 to 200 bars such that solubility in the liquid and dry steam phases could be measured. In relatively high base solutions and at very mildly acidic conditions, it was possible to conduct some sampling experiments with a Teflon-lined, stirred batch reactor. It was necessary to use dilute trifluoromethanesulfonic acid for the latter runs to both minimize complexation of the dissolved copper and corrosion of the Hastelloy-C pressure vessel. In all of the above experiments, the tip of the platinum liquid sampling line was enclosed in a platinum filter to prevent solid copper particles from entering the sampling lines.

Ultra pure commercial grade, CuO was used in these experiments without prior treatment. The commercial grade Cu,O on the other .hand, required hydrothermal treatment and ultrasonication in ethanol before it could be utilized. The crystallinity, morphology, and particle size of both oxides were confirmed by XRD, SEM and BET before and after the experiments.

APPROACH

A thermodynamic approach to solute partitioning between liquid and vapor phases has been adopted in this research program [ 14, I.51 in preference to the traditional ray diagram, which, among other limitations cannot account for the effect of pH on reactions of the solute in the liquid phase [2 13. For example, CuO dissolves in water at increasing pH according to the following equihbria (ignoring higher-order copper hydroxide species such as Cu,(OH),2’):

CuO + 2H’ * CL?* + H,O (1) CuO + H+ * Cu(OH) (2) CuO + H,O * Cu(OH),’ (3) CuO + 2H,O w Cu(OH),- H’ (4) CuO + 3H,O * Cu(OH),*- + 2H’ (5)

As supported by all the results obtained in this investigation, the dominant copper(I1) specie in the steam phase over this range of pH is the neutrally charged Cu(OH),‘. Therefore, the partitioning constant corresponds to the simple Henry’s law equation:

KD=m 0 WOW (vap) /m

2 Cu(OH):(liq)

.

w where m represents the measured molalities of Cu(OH),’ in each phase. To a first order approximation, the activity coefficient can be taken as unity due to the low concentrations observed in either phase. It is immediately apparent that although the total copper concentration measured in the steam phase over solutions whose chemistry is controlled by equations l-5, can be taken as being that due to Cu(OH),‘(vapor), the concentration measured in the liquid phase must be distributed between the species in these equations in order to determine that due to Cu(OH),“(liquid). Therefore, at each temperature investigated, the solubility in the liquid phase needs to by studied over a wide range of pH in order to determine the equilibrium constants for equations l-5. In fact, due to the relatively high solubilities of CuO in acidic solutions these measurements are difficult to interpret as the pH ofthe solution can only be obtained by calculation, which involves large uncertainty. Note that our attempts to measure pH directly with hydrogen-electrode concentration cells, which normally provide very accurate data to 300°C [22], were unsuccessful due to rapid reduction of Cu(I1) to copper metal. The values of solubility constants for equilibria 1 and 2 were taken from the most recent version of the compilation by Shock et al. [23]. Fortunately, low pH values where these species are dominant are not of direct interest to power generation, but they are required to model the other solubility constants effectively.

CuO SOLUBILITY IN LIQUID WATER

The solubility profile (log of the total Cu molality versus pH at temperature) of CuO at 100°C is shown in Figure 1 in order to illustrate a number of points. Our solubilit-y results in dilute NH, and NaOH are in reasonable agreement with those of Var’yash [3]. Increasing the NH, concentration above about 10 ppm begins to have dramatic effect on

-3 4 5 6 7 8 9 10 11 12 13 ,

Shock et al [23] 8

z Var’yash [3] - model excluding Cu(OH),*- - model excluding Cu(OH),-

0 Batch experiments (NaOH) A Flow-cell (0.02 molal NH,) 0 Flow-cell (NaOH)

Flow-cell (NH,+NaOH)

3 4 5 6 7 8 9 10 11 12 _

PH 1OOT

Figure 1. The solubility profile of CuO at 100°C and infinite dilution

4

*

Cu(I1) species are Cu’+ and Cu(OH)+ (using the solubility constants obtained from Shock ef al. [23]), Cu(OH),O, and including either Cu(OH), (eq. 4). or CU(OH),*~ (eq. 5). However, the fit did not converge when both of these anionic copper species were included. These results predict that under start up conditions in the presence of oxygen and moderate concentrations of NH3, the Cu(I1) concentration in the boiler is independent of pH above pH,,-c of 9 (2.5 ppb), but would increase significantly if the pH was lowered due to CO2 ingress or the use of higher NH, dosing.

Another example of the solubility of CuO taken at 300°C is ilIustrated in Figure 2. Here, the disparity between the various studies is greater, although reasonable agreement was found in our study between the flow and static volatility measurements independent of using NH, or NaOH to fix the pH. There is a very small pH window in which batch experiments could be conducted in acidic solutions (i.e., where there were sufficient excess of hydrogen ions over CL?+ ions and virtually no Cu(OH)’ ions so that the pH could be calculated unambiguously). These data have not yet been incorporated into the fit, but will be used in the hture when the entire data set is recalculated given the present knowledge of the s.olution speciation to allow appropriate activity coefficient corrections to be made. Finally, although no apriori decision can be made as to which model better applies to this system, our model that includes Cu(OH),’ is favored in light of the excellent agreement seen in Figure 2 with the SUPCRT92 [23 ] prediction (noting that similar agreement was observed at the other temperatures studied). Moreover, experience with other metals such as zinc shows that the Zn(OH),- specie is also dominant in the mildly basic solutions, with Zn(OH): only being observable at very high pH values [24].

PHWC 3 4 5 6 7 8 9 10 11 12 13

-.+- model excluding Cu(OH),- Flow-cell (NaOH)

A Flow-cell (NH,) Volatility-apparatus ( Batch vessel (CF,SO,

3 4 5 6 7 8 9 10 11

PHWC


. ., . l

cuo SOLUBILITY IN STEAM AND DIsTRwTI~N CONSTANT

By fitting the solubility constants obtained at each temperature (100 to 35O’C) as functions of temperature a full description of the liquid phase chemistry of copper(I1) in the presence of CuO is obtained. The partitioning constant can now be calculated based on the concentration of Cu(OH),’ in the liquid phase and the total copper concentration measured in the coexisting condensed vapor (eq. 6). The temperature dependence of the partitioning constant for this specie in shown in Figure 3 together with those of other neutral solutes. Note that in this case, K, is equivalent to the more familiar distribution constant D used in the ray diagram. Under normal operating conditions the boiler water pH approximates the condition where Cu(OH),’ is the dominant copper(I1) specie in the boiler water. Nevertheless, it must also be remembered that although CU(OH),~ is apparently quite volatile and K, is approximately unity at subcritical temperatures, its concentration in the liquid is limited to about 2 ppb, compared to acetic acid which has unlimited solubility.

2

0

-6

-8 1.6 1.8 2.0 2.2 2.4 2.6

1000 K/T

Figure 3. The logarithm of the partitioning constants for neutral solutes as a function of reciprocal temperature (Kelvin), where Tc in the critical temperature.

The solubility of CuO observed in this study is fully compatible with that reported by Pocock and Stewart [I] whose measurements in steam were in the supercritical region. At these higher temperatures it is also apparent that pressure has a much greater influence on solubility than temperature, which was not the case in the subcritical region (i.e., in our studies). Thus, increases in pressure?rom 100 to 300bars at temperatures >374”C, can give a IO fold increase in solubility due to the concomitant increase in steam density. Finally, note that the solubilities of CuO reported by Heam et al. [2] are more than an order of magnitude higher than those reported here.

THE TRANSPORT AND CHEMISTRY OF COPPER IN PdWER PLANTS AS DETERMINED BY LABORATORY EXPERIMENTS

Donald A. Palmer, Pascale BCnCzeth, and J. Michael Simonson Chemical and Anaiytical Sciences Division, Oak Ridge National Laboratory

Bldg. 45OOS, P.O. Box 2008, Oak Ridge TN 3783 l-61 10, U.S:A.

Andrei Y. Petrov* Moscow Power Institute, 14 Krasnokazarmennaya, Moscow 111250, Russia

*visiting scientist:

ABSTRACT

Solubility measurements have been made involvin g cupric (CuO, tenorite) and cuprous (Cu,O, cuprite) oxide in contact with aqueous solutions of NaOH and NH, to 350°C. The solubility of CuO was also measured in steam containing NH, to 4OO’C using a flow technique. These results were reinforced by “static” measurements of the partitioning of copper hydroxide from a saturated aqueous phase to the steam phase. A quantitative thermodynamic model of the speciation of copper(I1) in the liquid phase was formulated as a function of pH and temperature, that established the dominant species to be Cu(OH),’ and Cu(OH),- under most boiler water conditions. The former, neutrally-charged specie provides the only significant contribution for vaporous carry over to the steam phase. The relative volatility of Cu(OH),’ is high and shows remarkably little dependence on temperature. The major release of copper into the steam exiting the boiler appears to occur at startup due to the combination of oxidizing conditions, the presence of ammonia, and a higher probability of liquid droplet entrainment. Limited measurements with’ copper metal exposed to basic (NaOH) solutions containing oxygen indicate that copper levels in excess of the solubility limits of CuO and Cu,O can be readily achieved. On-going experiments with Cu,O confirm that the solubility is much more dependent on temperature and pH than was found for CuO, rising to very high concentrations at high pH and high temperature (e.g., ca. 1 OOppb at the solubility minimum of Cu,O at 250°C compared with ca. 2pbb for CuO). The experimental techniques and methods for data analysis will be outlined, and the results compared with those from previous studies. The. implications of these laboratory experiments to the chemistry of the water/steam cycle of fossil plants will also be presented.

BACKGROUND

The goal of this research is to identify the chemistry that occurs when copper and its oxides are exposed to conditions in fossil-fired boiler From start-up to normal operation. A combination of solubility and volatility experiments were initiated to determine quantitatively the concentrations of copper in solution (including steam) under varying conditions of temperature and pH where the oxidation state is controlled either by trace oxygen levels or by the copper metal/cupric oxide oxidation/reduction couple. These measurements are carried out in “inert” apparatus, precluding interactions with the wet walls of the containment that would otherwise jeopardize the application of rigorous equilibrium models.

Previous Copper Oxide Solubility Studies

There have been numerous studies of the solubility Of CuO to high temperatures [l-6], although the reported copper concentrations differ by as much as three orders of magnitude,under basic conditions [7]. Moreover, the pH dependence of the solubility varies dramatically according to the various studies, so that the speciation of copper(U)

“The submitted manuscript has been authored by a contractor of the U.S. Government under contract No. DE-AC05-OOOR2272.5. Accordingly, the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or allow others to do so. for U.S. Government n!!rn~+m= ‘I

in solution is very different. This impacts the predicted stability of the Cu(OH),’ molecule, which is considered to be the only significant copper species that partitions to the steam phase [7], as borne out by the current research. Heam et al. [2] reported solubilities of CuO into the supercritical region that are generally at least an order of magnitude higher than found previously by Pocock and Stewart [ 11. Heam et al. [2] attributed the lower values of the earlier work to possible lack of equilibration and/or adsorption of copper on the corroded surfaces of the containment vessel.

As indicated previously [7], the formation of complexes with anionic impurities (e.g., chloride, sulfate, carbonate, acetate) in the liquid phase is favored with increasing temperature. However, even ppm levels of these impurities are not expected to influence the solubility significantly at the relatively high pH conditions prevailing in a boiler. The possible exception being chloride complexation of Cu(I) that may increase the solubility of Cu,O significantly at boiler operating conditions. On the other hand, complexes of Cu(I1) with ammonia are stronger at low temperatures [S]; therefore during start-up in the presence of oxygen, ammonia could raise the concentration of copper in boiler water significantly (note that our work shows that beyond 200°C this effect can be virtually ignored). These general conclusions and observations have been supported recently by specific plant data and modeling efforts [9].

Finally, the effect of adsorption and surface precipitation of copper on metal oxide surfaces cannot be ignored as both mechanisms operate over wide pH ranges and the former is believed to be dominant even at high pH for hydrated iron oxide surfaces [IO]. Moreover, these experiments were carried out at 25°C and our recent investigation of surface charge and calcium adsorption on rutile (TiO,) has shown that both increase dramatically with increasing temperature at least to 290°C [ 1 I].

Solubility studies of Cu,O are very limited with apparently only one high temperature study being carried out in the liquid phase [ 121, although Pocock and Stewart [l] performed a limited number of analyses of Cu(I) in superheated steam. Var’yash [ 121 measured solubilities in the liquid phase from 150 to 450°C over a wide range of pH, although the technique for determining the pH was not given [6]. These results predict that the solubility of Cu,O increases by more than six orders of magnitude from 25 to 350°C at the solubility minimum with respect to pH where Cu(OH)’ is the dominant Cu(I) specie in solution. The minima occur very near the pH of boiler water, viz., 9.2, at 25°C. If these results are valid, then not only would Cu,O be highly soluble under boiler operating conditions, but it should be also highly volatile. Cu(OH)O should readily partition to the steam phase as this neutral molecule has some covalent character (cf. the case of HCl [13,14]).

Copper Dissolution and Transport in Water/Steam Cycles of Power Plants

Effective control of water-steam chemistry can eliminate the leading cause of reduced availability and high maintenance costs for fossil steam power plants, as well as improve plant heat rate and profitability. On the other hand, poor water-steam purity can result in corrosion damage and deposition on major components. It can also lead to boiler tube and turbine blade failures, flow-assisted corrosion (FAC) in feedwater piping systems, turbine efficiency losses, and other problems [ 15,161.

Uncontrolled copper transport represents a potentially significant source of performance and reliability loss to fossil plant units with mixed metallurgy feedwater systems. Recent utility experiences with severe copper turbine fouling and other related problems illustrate the need to improve the understanding in this area and to develop operating guidelines for controlling copper deposition around the cycle [ 16- 191. To this end, Project Copper has been focused on: 1) establishing which of the two major copper oxides (CuO and Cu,O) grow in mixed-metallurgy feedwater systems at each typical operating condition of temperature, pressure, air in-leakage, oxygen scavenger, electrochemical potential, and pH; 2) how these copper oxides are transported with feedwater into drum boilers; 3) how they become entrained in steam; and 4) determining their solubility limits in steam. In this context, the major aim of the research at ORNL is to quantify solubility limits for the major copper oxide forms in liquid water and steam, thereby supporting development of operating guidelines for controlling copper deposition [15,16].

3

T!

EXPERIMENTAL METHODS

The experimental technique employed at ORNL for measuring the solute volatilities is described by Simonson and Palmer [ 131, although minor modifications were made subsequently [20]. Succinctly, an equilibrated two-phase aqueous solution is contained in a platinum alloy liner seated in a one-liter pressure vessel and samples of both phases are withdrawn under controlled conditions for analysis by ion chromatography, Zeeman atomic absorption spectroscopy, or chemiluminescence. However, in all volatility experiments conducted here, solid copper oxide was present in the bottom of the liner and therefore both the liquid and steam phases were saturated with respect to their copper content. The solutions contained dilute NaOH in order to control the pH and experiments were performed at 100 to 350°c.

Direct solubility measurements were also made in a “packed column” configuration, whereby dilute NH, or NaOH (or a combination of both solutions) was pumped via a preheater into a tubular, Zircalloy pressure vessel packed with CuO, Cu,O+Cu, or Cu powder and the outlet stream was sampled for analysis. For the CuO experiments, the feed solution was either purged with helium or saturated with CO,-free air, or saturated with pure oxygen, whereas for the Cu,O solubility runs, the feed was first purged with a hydrogen/argon mixture and kept under positive helium pressure during the sampling period. In this latter case, it was also found necessary to inject either 0.02 molal HNO, or 0.03 molal NaOH solution into the outlet stream at the experimental temperature to prevent precipitation of copper oxide/hydroxide in the sampling line upon cooling. The temperature range of these experiments was 100 to 400°C, and the pressure was varied from 50 to 200 bars such that solubility in the liquid and dry steam phases could be measured. In relatively high base solutions and at very mildly acidic conditions, it was possible to conduct some sampling experiments with a Teflon-lined, stirred batch reactor. It was necessary to use dilute trifluoromethanesulfonic acid for the latter runs to both minimize complexation of the dissolved copper and corrosion of the Hastelloy-C pressure vessel. In all of the above experiments, the tip of the platinum liquid sampling line was enclosed in a platinum filter to prevent solid copper particles from entering the sampling lines.

Ultra pure commercial grade, CuO was used in these experiments without prior treatment. The commercial grade Cu,O on the other.hand, required hydrothermal treatment and ultrasonication in ethanol before it could be utilized. The crystallinity, morphology, and particle size of both oxides were confn-med by XRD, SEM and BET before and after the experiments.

APPROACH

A thermodynamic approach to solute partitioning between liquid and vapor phases has been adopted in this research program [ 14,151 in preference to the traditional ray diagram, which, among other limitations cannot account for the effect of pH on reactions of the solute in the liquid phase [21]. For example, CuO dissolves in water at increasing pH according to the following equilibria (ignoring higher-order copper hydroxide species such as Cu,(OH)p):

CuO + 2H’ * Cut+ + H,O (1) CuO + H’ * Cu(OH)+ (2) CuO + H,O - Cu(OH),’ (3) CuO + 2H,O * Cu(OH),- H’ (4) CuO + 3H,O - Cu(OH),*- + 2H’ (5)

As supported by all the results obtained in this investigation, the dominant copper(B) specie in the steam phase over this range of pH is the neutrally charged Cu(OH), ‘. Therefore, the partitioning constant corresponds to the simple . Henry’s law equation:

KD=m 0 CWW (v)

/m 2 Cu(OH)i(liq)

(6)

where m represents the measured molalities of Cu(OH)*’ in each phase. To a first order approximation, the activity coefficient can be taken as unity due to the low concentrations observed in either phase. It is immediately apparent that although the total copper concentration measured in the’steam phase over solutions whose chemistry is controlled by equations l-5, can be taken as being that due to Cu(OH),‘(vapor), the concentration measured in the liquid phase must be distributed between the species in these equations in order to determine that due to Cu(OH)20(liquid). Therefore, at each temperature investigated, the solubility in the liquid phase needs to by studied over a wide range of pH in order to determine the equilibrium constants for equations l-5. In fact, due to the relatively high solubilities of CuO in acidic solutions these measurements are difficult to interpret as the pH ofthe solution can only be obtained by calculation, which involves large uncertainty. Note that our attempts to measure pH directly with hydrogen-electrode concentration cells, which normally provide very accurate data to 300°C [22], were unsuccessful due to rapid reduction of Cu(I1) to copper metal. The values of solubility constants for equilibria 1 and 2 were taken from the most recent version of the compilation by Shock et al. [23]. Fortunately, low pH values where these species are dominant are not of direct interest to power generation, but they are required to model the other solubility constants effectively.

CuO SOLUBILIN IN LIQUID WATER

The solubility profile (log of the total Cu molality versus pH at temperature) of CuO at 100°C is shown in Figure 1 in order to illustrate a number of points. Our solubility results in dilute NH, and NaOH are in reasonable agreement with those of Var’yash [3]. Increasing the NH, concentration above about10 ppm begins to have dramatic effect on

PH 25T 7 8 9 10 11 12 13

Shock et al. b23] ’ I

z Var’yash [3] - model excluding Cu(OH),*- - model excluding Cu(OH),-

0 Batch experiments (NaOH) A Flow-cell (0.02 molal NH,)

0 Flow-cell (NaOH) 0 Flow-cell (NH,+NaOH)

-0- Ziemniak et al. [4]

-5 - 3

-6- 2

-7 - 1

-8 - 0

-9 I I I I I I I I -1

3 4 5 6 7 8 9 10 11 12 .

PH 100°C


Cu(I1) species are CuZf and Cu(OH)+ (using the solubility constants obtained from Shock er al. [23]), Cu(OH)t, and including either Cu(OH),* (eq. 4), or Cu(OH),*- (eq. 5). However, the fit did not converge when both of these anionic copper species were included. These results predict that under start up conditions in the presence of oxygen and moderate concentrations of NH,, the Cu(I1) concentration in the boiler is independent of pH above PH,YC of 9 (2.5 ppb), but would increase significantly if the pH was lowered due to CO, ingress or the use of higher NH, dosing.

Another example of the solubility of CuO taken at 300°C is illustrated in Figure 2. Here, the disparity between the various studies is greater, although reasonable agreement was found in our study between the flow and static volatility measurements independent of using NH, or NaOH to fix the pH. There is a very small pH window in which batch experiments could be conducted in acidic solutions (i.e., where there were sufficient excess of hydrogen ions over CL?” ions and virtually no Cu(OH)’ ions so that the pH could be calculated unambiguously). These data have not yet been incorporated into the fit, but will be used in the future when the entire data set is recalculated given the present knowledge of the Solution speciation to allow appropriate activity coefficient corrections to be made. Finally, although no apriori decision can be made as to which model better applies to this system, our model that includes Cu(OH); is favored in light of the excellent agreement seen in Figure 2 with the SUPCRT92 [23] prediction (noting that similar agreement was observed at the other temperatures studied). Moreover, experience with other metals such as zinc shows that the Zn(OH),- specie is also dominant in the mildly basic solutions, with Zn(OH): only being observable at very high pH values [24].

PH 25T 3 4 5 6 7 8 9 10 11 12 13

- model excluding Cu(OH),- 0 Flow-cell (NaOH) A Flow-cell (NH,)

Volatility-apparatus ( Batch vessel (CF,SO,

3 4 5 6 7 8 9 10 11

P%lilT


CuO SOLUBILITY IN STEAM AND DISTRIBUTION CONSTANT

By fitting the solubility constants obtained at each temperature (100 to 350°C) as functions of temperature a full description of the liquid phase chemistry of copper(I1) in the presence of CuO is obtained. The partitioning constant can now be calculated based on the concentration of Cu(OH),’ in the liquid phase and the total copper concentration measured in the coexisting condensed vapor (eq. 6). The temperature dependence of the partitioning constant for this specie in shown in Figure 3 together with those of other neutral solutes. Note that in this case, K, is equivalent to the more familiar distribution constant D used in the ray diagram. Under normal operating conditions the boiler water pH approximates the condition where Cu(OH),’ is the dominant copper(I1) specie in the boiler water. Nevertheless, it must also be remembered that although Cu(OH),O is apparently quite volatile and K, is approximately unity at subcritical temperatures, its concentration in the liquid is limited to about 2 ppb, compared to acetic acid which has unlimited solubility.

-8- HCOOH

-6 -

4!;1, ,I/:

1.6 1.8 2.0 2.2 2.4 2.6

1000 K/T

Figure 3. The logarithm of the partitioning constants for neutral solutes as a function of reciprocal temperature (Kelvin), where Tc in the critical temperature.

The solubility of CuO observed in this study is fully compatible with that reported by Pocock and Stewart [ 1] whose measurements in steam were in the supercritical region. At these higher temperatures it is also apparent that pressure has a much greater influence on solubility than temperature, which was not the case in the subcritical region (i.e., in our studies). Thus, increases in pressure’from 100 to 300bars at temperatures >374”C, can give a 10 fold increase in solubility due to the concomitant increase in steam density. Finally, note that the solubilities of CuO reported by Hearn et al. [2] are more than an order of magnitude higher than those reported here.

./ ‘. . . + .

4

Cu,O SOLUBILITY

As mentioned previously, only two experimental invest@ations of Cu,O solubility to high temperatures were found in the literature; one in liquid wafer [ 121 and the other in superheated steam [I]. In the former study nitric acid was used to adjust pH in acidic solutions and this should have resulted in oxidation to CuO, whereas on the basic side, NaOH was used such that these high pH results can be considered reliable until proven otherwise. Indeed in our initial experiments, which were conducted from 50 to 25O”C, reasonable agreenient was found with the results of Var’yash [ 121 in mildly basic NaOH solutions with apparently lower solubilities being observed at the solubilit-y minimum at each temperature. These initial findings indicate that the solubility of Cu,O increases more dramatically with temperature and pH than was found for CuO, such that Cu,O becomes more soluble above about 100°C in the pH range of interest to fossil-fired power plants. These preliminary results would also indicate that Cu(OH)’ should be more volatile than Cu(OH)P, although this depends largely on the stability of Cu(OH)’ in the liquid phase and this has yet to be established.

,As in the case of CuO, there appears to be a strong effect of ammonia on the solubility of Cu,O at lOO”C, but it is anticipated that this enhancement in solubility will decrease with temperature so as to become inconsequential above 200°C. Finally, the effect of chloride on Cu,O solubility is much stronger than observed for CuO and this effect increases with temperature [25]. However, if only ppb levels of chloride contamination are present in boiler water, then this effect can also be ignored.

ACKNOWLEDGMENTS

Principal financial support for this research was provided by EPRI, Inc. under the direction of R. Barry Dooley.

REFERENCES,

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

F.J. Pocock and J.F. Stewart, “The Solubility of Copper and Its Oxides in Supercritical Steam,” Journal of Engineeringfor Power. Vol. 85, No. 1, pp. 33-45 (1963). B. Heam, M.R. Hunt, and A. Hayward, “Solubility of Cupric Oxide in Pure Subcritical and Supercritical Water,” Journai of Chemical and Engineering Data. Vol. 14, No. 4, pp. 442-447 (1969). L.N. Var’yash, “Hydrolysis of Cu(I1) at 25-35O”C,” Geochemistry International. Vol. 23, No. 1, pp. 82-92 (1986). S.E. Ziemniak, M.E. Jones, and K.E.S. Combs, “Copper Oxide Solubility Behavior in Aqueous Sodium Phosphate Solutions at Elevated Temperatures,” Journal of Solution Chemistry. Vol. 21, No. 2, pp. 179- 200 (1992). K.M. Akhmetov, E.A. Buketov, and M.Z. Ugorets, “Solubility of Copper Oxide in NaOH Solutions at High Temperatures,” Trudy Khimiko-Metallurgicheskogo Instituta AN KazSSR. Vol. 3: “Heterogeneous Processes in Aqueous Environments”. Nauka, Alma-Ala, USSR, pp. 119-128 (1967). A.Y. Petrov, “Copper Solubility and Volatility. Former USSR Studies,” Report to the International Association for the Properties of Water and Steam. 1998. D.A. Palmer, J.M. Simonson, and D.B. Joyce, “Volatility of Copper”. Interaction of Non-bon Based Materials with Water and Steam, Piacenza, Italy (June 11-13, 1996). EPRI, Palo Alto, CA: 1997. TR- 108236, pp. 7.1-7.17. I.D. Isaev, S.V. Tverdokhlebov, V.M. Leont’ev, G.L. Pashkov, and V.E. Mironov, “The Influence of Temperature and Ionic Strength on the Stability Constants of Copper (II) Ammines in Aqueous Solutions,” Russian Journal of Inorganic Chemistry. Vol. 35, No. 8, pp. 1159- 116 1 (1990). J. Stodola, P. Tremaine, V. Binette, and L. Trevani, “Volatility of Copper Corrosion Products in a Power Generation Cycle,” Power Plant Chemistv. Vol. 2, NO. 1, pp. 9- 13 (2000). K.G. Karthikeyan and H.A. Elliott, “Surface Complexation Modeling of Copper Sorption by Hydrous Oxides of Iron and Aluminum,” JournaI of Colloid and Interface Science. Vol. 220, No. 1, pp. 88-95 (1999).

.

“r:’ 1 2

I

11.

12.

13.

14.

15. 16.

17. 18.

19.

20.

21. 22.

23.

24.

25.

M.K. Ridley, M.L. Machesky, D.J. Wesolowski, and D.A. Palmer, “Ca(I1) Adsorption at the Rutile-Water Interface: A Potentiometric Study in NaCl Media to 250°C” Geochimica et Cosmochimica Acta. Vol. 63, pp. 3087-3096 (1999). L.N. Var’yash, “Equilibria in the Cu-Cu,O-H,O System at 150-45O”C,” Geochemistry International. Vol. 26. Pp. 80-90 (1989). J.M. Simonson and D.A. Palmer, “Liquid-Vapor Partitioning of HCl(aq) to 35O”C,” Geochimica et Cosmochimica Acta. Vol. 57, NO. 1, pp. 1-7 (1993). D. A. Palmer and J. M. Simonson, Behavior ofAmmonium Salts in Steam Cycies. EPRI, Palo Alto, CA: 1993. TR-102377. Boiler and Turbine Steam and Cycle Chemistry Target. Available from www.euri.com. R.B. Dooley, A. Aschoff, K.J. Shields, B. Syrett, and T. McCloskey, Copper in the Fossil Plant Cycle. Sixth International Conference on Qcle Chemistry in Fossil Plants, Columbus, Ohio (June 27-29, 2000). State-of-Knowledge of Copper in Fossil Plant Cycles. EPRI, Palo Alto, CA: 1997. TR-108460. K.J. Shields, B. Dooley, T.H. McCloskey, B.C. Syrett, and J. Tsou, “ Copper Transport in Fossil Plant Units,” Power Plant Chemistry. Vol. 1, NO. 4, pp. 8-16 (1999). A.G. Howell, “Mitigation of Copper Deposition in High Pressure Turbines of Utility Drum Boilers,” Power Plant Chemistry. Vol. 1, NO. 4, pp. 18-25 (1999). D.A. Palmer, J.M. Simonson, R.W. Carter, J.P. Jensen, D.B. Joyce, and S.L. Marshall, Volatility of Aqueous Sodium Hydroxide, Bisulfar and Sulfate. EPRI, Palo Alto, CA: 1999, TR- 10580 1. D.A. Palmer and J.M. Simonson, Assessment of the Ray Diagram. EPRI, Palo Alto, CA: 1996, TR- 1060 17. D.A. Palmer, P. BCnCzeth, and D.J. Wesolowski, “Aqueous High Temperature Solubility Studies. I. The Solubility of Boehmite as Functions of Ionic Strength (to 5 Molal, NaCl), Temperature (loo-29O”C), and pH as Determined by In Situ Measurements, ” Geochimica et Cosmochimica Acta. Submitted. E.L. Shock, D.C. Sassani, M. Willis, and D.A. Sverjensky, “Inorganic Species in Geological Fluids: Correlations among Standard Molal Thermodynamic Properties of Aqueous Ions and Hydroxide Complexes,” Geochimica et Cosmochimica Acta. Vol. 6 1, No. 5, pp. 907-950 (1997). P. Btnezeth, D.A. Palmer, and D.J. Wesolowski, “The Solubility of Zinc Oxide in 0.03m Ionic Strength as a Function of Temperature with In Situ pH Measurement,” Geochimica et Cosmochimica Acta. Vol. 63, No. 10, pp. 1571-1586 (1999). 2. Xiao, C.H. Gammons, and A.E. Williams-Jones, “Experimental Study of Copper(I) Chloride Complexing in Hydrothermal Solutions at 40°C to 3OO’C and Saturated Water Vapor Pressure,” Geochimica et Cosmochimiia Acta. Vol. 62, NO. 17, pp. 2949-2964 (1998).

the transport and chemistry of copper in pbwer plants as determined by laboratorv...

Documents