SANDIA REPORT
SAND 2013-8256
Unlimited Release
September 2013
Corrosion of High Temperature Alloys in
Solar Salt at 400, 500, and 680°C
Alan M. Kruizenga, David D. Gill, Marianne LaFord, Geoffrey McConohy
Prepared by
Sandia National Laboratories
Albuquerque, New Mexico 87185 and Livermore, California 94550
Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation,
a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy's
National Nuclear Security Administration under contract DE-AC04-94AL85000.
Approved for public release; further dissemination unlimited.
Issued by Sandia National Laboratories, operated for the United States Department of Energy
by Sandia Corporation.
NOTICE: This report was prepared as an account of work sponsored by an agency of the
United States Government. Neither the United States Government, nor any agency thereof,
nor any of their employees, nor any of their contractors, subcontractors, or their employees,
make any warranty, express or implied, or assume any legal liability or responsibility for the
accuracy, completeness, or usefulness of any information, apparatus, product, or process
disclosed, or represent that its use would not infringe privately owned rights. Reference herein
to any specific commercial product, process, or service by trade name, trademark,
manufacturer, or otherwise, does not necessarily constitute or imply its endorsement,
recommendation, or favoring by the United States Government, any agency thereof, or any of
their contractors or subcontractors. The views and opinions expressed herein do not
necessarily state or reflect those of the United States Government, any agency thereof, or any
of their contractors.
Printed in the United States of America. This report has been reproduced directly from the best
available copy.
Available to DOE and DOE contractors from
U.S. Department of Energy
Office of Scientific and Technical Information
P.O. Box 62
Oak Ridge, TN 37831
Telephone: (865) 576-8401
Facsimile: (865) 576-5728
E-Mail: reports@[Link]
Online ordering: [Link]
Available to the public from
U.S. Department of Commerce
National Technical Information Service
5285 Port Royal Rd.
Springfield, VA 22161
Telephone: (800) 553-6847
Facsimile: (703) 605-6900
E-Mail: orders@[Link]
Online order: [Link]
2
SAND2013-8256
Unlimited Release
September 2013
Corrosion of High Temperature Alloys in Solar
Salt at 400, 500, and 680°C
Alan M. Kruizenga1, David D. Gill2, Marianne LaFord1
1. 8223: Materials Chemistry
2. 6123: Concentrating Solar Technologies
Sandia National Laboratories
P.O. Box 0969
Livermore, California 94551-MS9403
Abstract
Corrosion tests at 400, 500, and 680°C were performed using four high temperature alloys;
347SS, 321SS In625, and HA230. Molten salt chemistry was monitored over time through
analysis of nitrite, carbonate, and dissolved metals. Metallography was performed on alloys at
500 and 680°C, due to the relatively thin oxide scale observed at 400°C. At 500°C, corrosion of
iron based alloys took the form of chromium depletion and iron oxides, while nickel based alloys
also had chromium depletion and formation of NiO. Chromium was detected in relatively low
concentrations at this temperature. At 680°C, significant surface corrosion occurred with metal
losses greater than 450microns/year after 1025hours of exposure. Iron based alloys formed
complex iron, sodium, and chromium oxides. Some data suggests grain boundary chromium
depletion of 321SS. Nickel alloys formed NiO and metallic nickel corrosion morphologies, with
HA230 displaying significant internal oxidation in the form of chromia. Nickel alloys both
exhibited worse corrosion than iron based alloys likely due to preferential dissolution of
chromium, molybdenum, and tungsten.
3
4
ACKNOWLEDGMENTS
Authors would like to gratefully acknowledge the skillful metallography preparation and
methodical SEM work of Ryan Nishimoto and Andrew Gardea, with electron microprobe
analysis of Miles Clift.
5
CONTENTS
1. Introduction ................................................................................................................................ 9
2. Experimental methods ............................................................................................................. 11
3. Results ...................................................................................................................................... 15
A. Molten Salt Analysis ..................................................................................................... 15
i. Salt Chemistry Measurements ........................................................................ 15
ii. Changes in Molten Salt Physical Properties due to Corrosion and Impurities 17
B. Fe-Cr-Ni alloys (347SS and 321SS) ............................................................................. 24
i. 400-500°C corrosion ....................................................................................... 24
ii. 680°C corrosion .............................................................................................. 24
C. Ni-Cr alloys (HA230 and IN625) ................................................................................. 28
i. 400-500°C corrosion ....................................................................................... 28
ii. 680°C corrosion .............................................................................................. 29
4. Conclusions ............................................................................................................................... 37
5. References ................................................................................................................................ 38
Appendix:...................................................................................................................................... 41
A. Sample pull Schedule and Order................................................................................... 41
B. Salt Specifications ........................................................................................................ 42
Distribution ................................................................................................................................... 43
FIGURES
Figure 1: Visual appearance of salt operated at different temperatures........................................ 11
Figure 2: Sample tree and samples after exposure in 680°C salt.................................................. 12
Figure 3 : Example of heat flow data collected by DSC measurements ....................................... 13
Figure 4: Nitrite concentration with calculated equilibrium values [11] ...................................... 16
Figure 5: Dissolved metals detected in 500 and 680°C melt. Tungsten was only detected at 3000
hours in the 500°C melt. Nickel was detected in 680°C melt. Additional Cr at 680°C will arise
from the containment vessel, which is 316SS. ............................................................................. 16
Figure 6: Changes in Specific heat Curves for 680°C Operating Temperature ............................ 18
Figure 7: Changes in Specific Heat Curves for 600°C Salt for Various Operating Times ........... 18
Figure 8: Change in Melting Point for Salt at 680°C after Various Operating Times ................. 18
Figure 9: Change in Melting Point for Salt at 600°C ................................................................... 19
Figure 10: Comparison between a heat flow curve from Rogers and Janz (left) and this work
(right). ........................................................................................................................................... 20
Figure 11: Influence of Nitrite Content on Melting Point ............................................................ 21
Figure 12: Effect of nitrite and nitrate on melting temperature .................................................... 22
Figure 13: Changes in Melting Temperature after Heating at 300°C and 500°C ......................... 22
Figure 14: Graphs of heat of fusion (left column) and Specific Heats (right column) for 680°C
salt (top row) and 600°C salt (bottom row) .................................................................................. 23
6
Figure 15: Weight gain and descaled loss of iron based alloys for temperatures of 400°C and
500°C. ........................................................................................................................................... 25
Figure 16: Weight gain (left) and descaled weight loss (right, 1025 hour only) data for alloy
347SS and 321SS at 680°C. Note plot on right does not start from zero. .................................... 26
Figure 17: 347SS after 3064 hours at 500°C, iron oxide with no formation of NaFeO2. ............ 26
Figure 18: 321SS after 3064 hours exposure at 500°C. No detection of sodium, with XRD scans
indicating magnetite formation. .................................................................................................... 26
Figure 19: 347SS after 1025 hours exposure at 680°C. Phases of NaFeO2 were detected with
XRD after 520 hours (bottom left), with mixed oxide detected at 1025 hours (bottom right).
Corrosion proceeded uniformly with no obvious grain boundary depletion of chromium. ......... 27
Figure 20: 321SS corrosion at 1025 hours at 680°C. Slight Cr depletion was observed along
grain boundaries, as indicated by X-ray map (bottom left). ......................................................... 28
Figure 21: 500°C exposure for 3064 hours on HA230. Surface corrosion at lower temperatures
could not be adequately resolved by microscopy methods used, though XRD indicated formation
of NiO. .......................................................................................................................................... 30
Figure 22: In625 corrosion product evolution based on exposure temperature............................ 31
Figure 23: Weight gain and descaled information for HA230 and In625 at 400°C and 500°C. .. 32
Figure 24: Weight gain data for 680°C exposure of HA230 and In625 (left). Descaled data (right)
for 1025 hour exposure only. ........................................................................................................ 32
Figure 25: HA230 after 1025 hours at 680°C. Line scan and XRD data clearly indicates phases
of NiO and metallic nickel, with small peaks that appeared to correspond to iron oxide. ........... 33
Figure 26: XRD scan of HA230 after descaling process was performed on 1025 hour, 680°C
sample. No base alloy was detected, only phases of chromia and nickel, clearly indicating
internal oxidation present. ............................................................................................................. 33
Figure 27: In625 after 1025 hour exposure at 680°C. Nickel and NiO were present as primary
corrosion products. ........................................................................................................................ 34
Figure 28: 400°C removal schedule and sample ID ..................................................................... 41
Figure 29: 500°C removal schedule and sample ID ..................................................................... 41
Figure 30: 680°C removal schedule and sample ID ..................................................................... 42
TABLES
Table 1: Compositions of alloys investigated in 400, 500, and 680°C experiments .................... 12
Table 2: Values of Thermo-physical properties for Various Conditions of Solar Salt at 680°C
Operating Temperature ................................................................................................................. 17
Table 3: Values of Thermo-physical properties for Various Conditions of Solar Salt at 600°C
Operating Temperature ................................................................................................................. 17
Table 4: Optimelt Images of salt experiments for 680°C Salts .................................................... 19
Table 5: Optimelt Images of Salts Operated at 600°C For Various Lengths of Time .................. 19
Table 6: Corrosion Rates and projections for iron based alloys ................................................... 25
Table 7: Corrosion Rates and projections for nickel based alloys ................................................ 31
7
NOMENCLATURE
BSE Back Scatter Electron Image
CTE Coefficient of Thermal Expansion
EDS Energy dispersive spectroscopy
HA Haynes Alloy
In Inconel Alloy
M” Mass normalized by surface area (mg/cm2)
SEM Scanning Electron Microscopy
SEI Secondary Electron Image
SNL Sandia National Laboratories
SS Stainless Steel
X Oxide thickness
XRD X-ray diffraction
WDS Wavelength Dispersive Spectroscopy
8
1. INTRODUCTION
Concentrating solar power (CSP) efficiency is increased using high operating temperature where
bulk salt temperatures approach 600°C. Wall temperatures will be higher than this in order to
achieve adequate temperature differentials needed for turbulent convection heat transfer. Molten
salt corrosion tests in the past that have been performed above 600°C have attempted to control
salt chemistry by increasing the oxygen overpressure, thus altering equilibrium concentrations of
anions present [1]. However, practical considerations for large thermal energy storage tanks,
such as those used in Solar Two [2], with oxygen overpressures present non-trivial engineering
considerations and costs.
Past work looking at many alloys at 677°C were helpful in understanding qualitative rankings
between alloys [3]. 14 day tests (336 hour) were employed for ranking corrosion performance,
with little microscopy to ascertain if intergranular attack or internal oxidation was present. Four
alloys of interest, two iron based (347SS and 321SS) and two nickel based (HA230 and In625),
are investigated at 400, 500 and 680°C. These will provide a more complete data set to
complement work performed at 600°C [4].
9
10
2. EXPERIMENTAL METHODS
This investigation employed a series of static immersion tests in binary nitrate “solar salt”
(60/40wt% Na/K\\NO3) at 400, 500, and 680°C. Samples of 4 different alloys (see Table 1) were
removed periodically (500, 1000, 2000, and 3000 hours) for corrosion rate analysis, which used
ASTM methods as guidance for sample handing and analysis [5]. The corrosion apparatus [6]
used air sparging for both chemistry control and to avoid salt stratification within the salt
chamber. Samples were obtained from Metal Samples Co. (A Division of Alabama Specialty
Products, Inc.) with a surface finish of 120 grit.
Figure 1 provides a visual representation of salt coloring due to operational temperature. 680°C
salt turned a dark yellow to green color, likely due to a combination of nitrite and dissolved
metals present, which will be discussed further in the results section. Severe corrosion of the test
chamber caused a test halt around 2000 hours resulting in a salt leak. Corrosion of sample tree,
series 316SS - shown in Figure 2, exhibited thinning of the tree below the salt line. The stainless
steel tube, also series 316SS, thinned so significantly that the remaining tube became brittle and
broke upon contact.
Figure 1: Visual appearance of salt operated at different temperatures
11
Figure 2: Sample tree and samples after exposure in 680°C salt
Characterization of corrosion rates relied upon removal of corrosion products, which employed a
combination of chemical and mechanical techniques. After removal from molten salt immersion
vessels, samples were placed in a boiling sodium hydroxide and potassium permanganate
solution, rinsed, and placed in a diammonium citrate solution. 400 and 500°C samples, which
had minimal surface corrosion, were fully cleaned by repeating this chemical bathing process for
several cycles. 680°C samples required both chemical baths and mechanical abrasion using grit
blasting to complete the removal of corrosion products [4]. Further chemical baths were needed
after the grit blasting to remove scales still present on the HA230 and In625.
Table 1: Compositions of alloys investigated in 400, 500, and 680°C experiments
Cr Mo Ni Mn Si Ti Fe Co W Al Nb Cu
321SS† 17.28 0.34 9.10 1.80 0.63 0.16 70.37 - - - - 0.32
347SS† 17.45 0.32 9.43 1.57 0.63 - 69.72 - - - 0.62 0.26
HA230† 22.37 1.27 59.41 0.49 0.42 - 1.32 0.19 14.16 0.32 - 0.05
In625† 21.76 8.28 61.0 0.21 0.25 0.24 4.46 0.08 - 0.20 3.38 -
† Actual composition from heat
Metal loss rates calculated in this report are reported in microns/year using the following
equation:
( )
Equation 1
Where M” is descaled mass loss per unit area (mg/cm2), is alloy density (g/cm3) and T time
hours [7]. The assumption behind this equation is that corrosion proceeds in a linear fashion
over the course of a year, which is the conservative approach.
Standard metallography techniques were also used to assess materials corrosion by using
scanning electron microscopy, electron microprobe, and x-ray diffraction [4].
12
Three thermo-physical quantities of the aged salt were measured: melting point, heat of fusion,
and specific heat. Using a Mettler Toledo DSC (Model: DSC823, Serial Number: 5126041530),
the heat flow was measured from 100°C to 350°C using a heating rate of 10 K/ minute at
atmospheric pressures under argon cover gas. All samples were massed to an accuracy of ±0.1
mg. Samples were prepared for each condition of thermally aged salt and were tested on the DSC
in triplicate with a blank reference and a sapphire standard. All samples used a standard 40
microliter aluminum pan with a hermetically sealed lid for testing. The melting point and heat of
fusion were calculated using onset and integration functions on Mettler Toledo’s software
package. The specific heat for a given temperature was calculated using the following formula:
Equation 2
Where HF is the blank subtracted heat flow [mW], m is the mass in mg, and Cp is the specific
heat [J/g-K] for the object noted by the subscript. This equation is valid provided the heating rate
for both the sapphire standard and sample are the same. Before examining the unknown
properties of contaminated salts with the DSC, samples of potassium nitrate and sodium nitrate
were tested and compared with literature values in order to gain confidence in the methods
employed. Both the samples were measured within 5% of the literature values reported
elsewhere [8].
Below is a sample of DSC data obtained from Mettler Toledo’s software package:
Figure 3 : Example of heat flow data collected by DSC measurements
DSC results of the aged salts were found to show non-eutectic properties. Thus, it was necessary
to confirm the melting point data using a system called Optimelt (Stanford Research Systems,
Serial Number 97656) by the following procedure: First, samples were ground into a powder
and placed in 1.5 mm diameter capillary tubes. Second, samples were heated at 10 K/minute
from 120°C to 350°C in air. In this experiment, the liquidus point of the system can be most
accurately measured and is reported in Table 2. Error for this measurement is estimated to be ±3
°C for each sample.
13
14
3. RESULTS
A. Molten Salt Analysis
i. Salt Chemistry Measurements
Significant differences in temperatures studied over this corrosion test resulted in different salt
chemistry behavior. Analysis can be divided into two categories: results at or below 500°C and
results at 680°C. Salt samples were acquired during metal sample pulls, nominally at 500, 1000,
2000, and 3000 hours. Samples were stored in a closed vial after removal, within a desiccant
chamber, unless transported.
The nitrate salts used for the test were refined grade materials acquired from SQM America.
Though not specifically measured before testing, the salt is guaranteed to have component values
within the specifications shown in the Appendix B. Salt analysis of samples taken during
corrosion testing was performed using standard chemistry techniques, which tested for nitrite and
nitrate content, dissolved metals (Cr, Ni, Fe, Ti, W, Mo), chloride, and carbonate. Chloride was
constant throughout tests, as expected, present on an average of 0.058wt%.
The salts were conditioned after melting by maintaining the salt at 500°C for more than a week.
This insured that adequate time and temperature were used to release water from the salt and also
to react any magnesium into magnesium oxide which would precipitate to the bottom out of the
melt. This reaction is responsible for the majority of the NOx release with new salt and the week
timeframe with bubbling to stir has proven effective in significantly completing this reaction.
Carbonate ( ) concentration was less than 0.2wt% for the entire duration of tests performed
at temperatures of 400 and 500°C. concentration varied substantially at 680°C, ranging
from 0.2wt% at 520 hours to 1.7wt% at 2000 hours. This illustrated CO2, from the air sparge
gas, was behaving as an oxide scavenger. It is expected that a salt not subjected to a constant
source of CO2 would result in a higher concentration of oxide anions, which would further
participate in oxidation processes. High amounts of could lead to formation of insoluble
products over time that may lead to plugging of valves or piping [9].
Nitrite equilibrium behavior was also monitored over time. The relative concentration of nitrite
can be determined through the dominant reaction with oxygen:
⁄ Equation 3
( ) Equation 4
G is the Gibbs energy of formation as determined by [10, 11], R the ideal gas constant, K the
equilibrium coefficient. It was found that experimentally determined values of nitrite followed
qualitative trends in temperature, but were found to be consistently low as indicated in Figure 4.
It was unclear what caused this quantitative inconsistency in measured and calculated nitrite
values, but similar behavior was also documented in while performing 600°C tests within the
same system [4].
15
Dissolved metals varied based on temperature (Figure 5). At 400°C all dissolved metals were
less than 30ppm, where Cr was the highest concentration detected; 22ppm at 3000 hours. Cr
content increased slightly with the temperature of 500°C to 44ppm. Also, trace amounts of Mo
and W also detected (also Figure 5).
680°C yielded obvious increases in metal dissolution rate and content. Cr content over 7400 ppm
was detected at 2000 hours, with significant concentrations of Mo and W. Tungsten and
molybdenum are soluble in nitrate salts by formation of tungstates and molybdates [12-15]. The
detection of nickel in the melt, albeit in low concentrations, was not expected in light of previous
experiments. While this result continues to be somewhat surprising, the constant value
throughout 2000 hours of testing does indicate a low solubility. Nickel present in the melt,
whether soluble or insoluble, is from the corrosion of the nickel based alloys. Presence of soluble
nickel in the melt may be a result of a NiWO4, which was identified other work [15, 16] or
perhaps could be the result of a nickelate ion.
Figure 4: Nitrite concentration with calculated equilibrium values [11]
Figure 5: Dissolved metals detected in 500 and 680°C melt. Tungsten was only detected
at 3000 hours in the 500°C melt. Nickel was detected in 680°C melt. Additional Cr at 680°C
will arise from the containment vessel, which is 316SS.
16
ii. Changes in Molten Salt Physical Properties due to Corrosion and Impurities
Results and Discussion for 680°C and 600°C Salts
The data collected was used to compare the thermo-physical properties of pure solar salt to the
salt used in the corrosion studies. Below are tables and graphs summarizing different thermo-
physical quantities for various hours of use, including values from literature for comparison.
600°C salt was obtained from archival samples from a previous test [4].
Table 2: Values of Thermo-physical properties for Various Conditions of Solar Salt at
680°C Operating Temperature
Hours at Heat of Fusion Cp at 327°C Melting Point (°C) Liquidus (°C) Liquidus (°C)
680°C (J/g) ( J/g-K) (from DSC) (from Optimelt) (from DSC)
245 ± 1
Source [17] 112 ±1 1.59 ± 0.03 225 ± 1 -
(DSC measurement)
0 107± 1.5 1.77± 0.16 220± 0.29 242 ± 3 244 ± 3
520 85.3 ± 2.3 1.60 ± 0.22 161 ± 1.5 213 ± 3 209 ± 3
1025 74.6 ± 3.3 1.83± 0.59 151 ± 1.5 NA* 200 ± 3
2000 69.1 ± 2.2 1.63± 0.59 156 ± 0.7 202 ± 3 197 ± 2
* This value was unable to be determined because of large displacements of the molten salt during heating which
rendered changes in phase undetectable
Table 3: Values of Thermo-physical properties for Various Conditions of Solar Salt at
600°C Operating Temperature
Hours at Heat of Fusion Cp at 327°C Melting Point (°C) Liquidus (°C) Liquidus (°C)
600°C (J/g) ( J/g-K) (from DSC) (from Optimelt) (from DSC)
245 ± 1
Source[17] 112 ±1 1.59 ± 0.03 225 ± 1 -
(DSC measurement)
0 107± 1.5 1.77± 0.16 220± 0.29 242 ± 3 244 ± 3
1000 94.2 ± 6.2 1.53 ± 0.09 192 ± 0.6 231 ± 3 243 ± 3
3000 90.4 ± 0.65 1.95 ± 0.34 190 ± 0.5 227 ± 3 227 ± 3
4000 89.4 ±2.2 1.55 ± 0.12 192 ± 0.3 228 ± 3 227 ± 3
17
10
9
8
Specific Heat (J/ g K)
7 520
Hours
6
5 1025
Hours
4
2000
3
Hours
2
Pure Salt
1
0
Example Liquidus Points
110 160 210 260 for Pure and 2000 Hour
Temperature (°C) Salt
Figure 6: Changes in Specific heat Curves for 680°C Operating Temperature
10
9
8
Specific Heat (J/ g K)
7 3000
Hours
6
4000
5 Hours
4 Pure Salt
3
2 1000
Hours
1
0
150 170 190 210 230 250 270 290 310
Temperature (°C)
Figure 7: Changes in Specific Heat Curves for 600°C Salt for Various Operating Times
230
220
210
Melting point (°C)
200
This Work Literature Value
190
180
170
160
150
140
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Hours of Operation
Figure 8: Change in Melting Point for Salt at 680°C after Various Operating Times
18
Melting Point Results
Table 4: Optimelt Images of salt experiments for 680°C Salts
Pure Salt 520 h 2000
h
100°C 199°C 207°C 247°C
Optical images showing the melting process of nitrate salts. From left to right- Pure Solar Salt, 520 hours,
and 2000 hours of use at 680°C. Each photo starting at 199°C demonstrates the liquidus point of one of the
salts. Note the 40°C temperature difference between the liquidus point of 520 hour salt and pure solar salt.
Table 5: Optimelt Images of Salts Operated at 600°C For Various Lengths of Time
1000 h 3000 h 4000h
120°C 227°C 232°C
Optical images showing the melting process of nitrate salts. From left to right- 1000 hours, 3000
hours, and 4000 hours of use at 600°C. These three salts show a nearly identical liquidus point
which produces a discrepancy between the DSC data and optimelt data for the 1000 hour salt.
This is most likely an artifact of DSC measurements, as the melting point of all three salts is very
similar.
225
220 This Work
Melting point (°C)
215
210 Literature
205 Value
200
195
190
185
0 1000 2000 3000 4000 5000
Hours of Operation
Figure 9: Change in Melting Point for Salt at 600°C
19
Melting point Discussion
From the results above, it is clear that the melting range of the solar salt decreases as the salt is
exposed to high temperatures and impurities. The heat capacity curves for thermally aged salt
indicate that the melting range is broadened and shifted to lower temperatures. The broadening
of the melting point is due to these non-eutectic properties of the mixture. Each component of the
mixture melts at a different temperature and results in a larger melting range.
Optimelt measurements show very good agreement with data taken from the DSC, confirming
the decrease in melting point. The specific heat curves also indicate the liquidus point of the salt
(see Figure 6). This value was used in comparison with the values measured from Optimelt.
These effects have been documented previously by Rogers and Janz who showed that a small
concentration of nitrite ions and oxides in the melt can cause significant changes in thermo-
physical properties of solar salt [17]. Figure 10 shows the similarities between the heat curves
measured by Rogers and Janz and the heat curves measured in this work. Even though the salt
studied in this work is of composition 65/35 mol% sodium to potassium, as compared to 50/50%
for Rogers and Janz, the two heat curves bear striking resemblance to each other. These
similarities suggest that metal impurities produced by corrosion are not likely to be the cause of
the changes in melting point of the salt mixture. In addition, measurements taken from the Solar
Two project indicate similar results for a fully operational concentrating solar power plant [18].
Figure 11 shows the correlation between the nitrite concentration and the melting point.
Correlations between other impurities showed no significant relationship to melting point.
However, the oxide content of the salt remains unknown and could potentially have some effect.
Curve A- Before Corrosion Test
Curve B- After Corrosion Test at 600° for 1000 hours
Figure 10: Comparison between a heat flow curve from Rogers and Janz (left) and this
work (right).
20
250
200
Melting Point (°C) 150
y = -6.49x + 228.95
R² = 0.9434
100
50
680°C Operating Temperature 600°C Operating Temperature
0
0 2 4 6 8 10 12
Nitrite Content (wt%)
Figure 11: Influence of Nitrite Content on Melting Point
In addition to the previous literature, nitrite composition has been known to produce mixtures
with much lower melting points than solar salt. For example, the HITEC mixture of NaNO3 ,
KNO3, and NaNO2 has a melting point of 142°C [19]. This behavior also qualitatively matches
phase diagrams produced over 50 years ago by Berul and Bergman (See Figure 12). In order to
attempt and explain the changes in melting point of the nitrate salt mixture, samples of salts
operated at 680°C for 500 and 2000 hours were heated in a furnace at 300°C along with a sample
of pure solar salt. Alumina crucibles were used for heating and roughly 2-4 grams of each salt
were tested. Once every 24 or 48 hours, small amounts of the molten salt were collected and
tested using the Optimelt system to determine if any changes in nitrite concentration produced by
the 300°C furnace affected the melting point of the salt. After 168 hours at 300°C, the furnace
temperature was increased in order to increase the kinetics of the reaction between nitrite and
nitrate (Equation 2).
Different nitrate/nitrite ratios will change the melting point of the salt. If no change is found,
other impurities in the mixture are likely causing the change in melting point. Figure 13 shows
the changes in melting point of three salts as they are held at 300°C over many days. After 7
days with minimal changes in melting point, the temperature of the furnace was raised to 500°C
to attempt to change the nitrite concentration faster. At this temperature, the melting point was
observed to rise significantly, suggesting that the nitrite content is responsible for the decrease in
melting temperature. Furthermore, there are no external chemical sources or sinks that could
have changed the composition of the melt except for the change in nitrite concentration.
21
Figure 12: Effect of nitrite and nitrate on melting temperature
Figure 13: Changes in Melting Temperature after Heating at 300°C and 500°C
Specific Heat and Heat of Fusion Results and Discussion
Regarding the specific heat of used salt, measurements show no observable changes due to
impurities. Any change in value is within one standard deviation from the literature value of pure
22
salt. Rogers and Janz also noted that the specific heat of the mixture changes very little when
heated in air. This is most likely due to the relatively small amount of impurities added or
produced in the salt mixture. Since the specific heat can be calculated as a weighted average
based on the mole percent and specific heat of each component, very small additions of nitrite or
metals are not likely to have a significant effect on the specific heat. However, the heat of fusion
seems to show noticeable changes after all three of the salts operated at 680°C. This phenomenon
is more difficult to explain because heat of fusion cannot be calculated using simple analytical
methods.
120 3
Specific Heat at 327°C (J/ g K )
100
Heat of Fusion (J / g )
2.5
80 2
60
1.5
40
1
This Work Literature Value This Work
20
0.5
0 Literature Value
0 500 1000 1500 2000 2500 0
0 500 1000 1500 2000 2500
Hours of Operation Hours of Operation
120 2.5
Specific Heat at 327°C (J/ g K )
100
2
Heat of Fusion (J / g )
80
1.5
60
This Work 1
40
Literature Value 0.5
20 This Work Literature Value
0
0 0 2000 4000
0 1000 2000 3000 4000
Hours of Operation Hours of Operation
Figure 14: Graphs of heat of fusion (left column) and Specific Heats (right column) for
680°C salt (top row) and 600°C salt (bottom row)
This mixture, named solar salt, is a 40/60% by weight mixture of KNO3 and NaNO3. When
operating at high temperatures, the corrosion products are dissolved in the molten salt and the
high temperatures convert some of the nitrates in to nitrites. The previous experiments have
shown that the melting point undergoes the most significant change of the three properties
studied- specific heat, heat of fusion, and melting point. This change in melting point has been
shown to be caused by the change in equilibrium products at high temperatures and most likely
23
has little to do with the amount of corrosive products dissolved in the salt. The changes in
melting point have also been shown to be reversible when samples are heated to a slightly lower
operating temperature than the original corrosion experiments. Moreover, the heat of fusion
shows a slight decrease for an increase in operating time. However, the specific heat does not
show any significant change within the error of the measurements.
B. Fe-Cr-Ni alloys (347SS and 321SS)
i. 400-500°C corrosion
Iron based austenitic alloy 347SS and 321SS were investigated at 400, 500, and 680°C, with
behavior that can be separated into 400/500°C and 680°C corrosion. Corrosion rates, for
temperatures below 500°C, indicated minimal depth of affected alloy (Table 6). In fact, 400°C
corrosion was so slight that optical and SEM images yielded little information and are not
included here.
500°C exhibited surface corrosion with higher corrosion rates than 400°C. Performance of
alloys was excellent with rates that indicate corrosion to be insignificant under these conditions.
Formation of surface oxides showed no sodium oxides at 400/500°C with XRD results indicating
magnetite (Fe3O4); this is consistent with other results [20, 21]. Chromium depletion was present
in oxidation scales, as indicated by line scan data of 347SS (Figure 17) and 321SS (Figure 18),
where predominately iron oxide remained.
ii. 680°C corrosion
680°C exposure resulted in a significant increase in corrosion. Similarities are present between
oxidation at 600°C [4] and oxidation at 680°C, in that NaFeO2 forms on the outer surface and
mixed oxides beneath the sodium ferrite, there are obvious differences in the mixed oxides
present.
Oxides formed at 680°C are richer in nickel and chromium content than at 600°C (Figure 19,
Figure 20). This result is presumably driven by the accelerated rate of attack of the base alloy.
Simultaneous and competing processes are occurring; oxygen diffusing inwardly, metals
(chromium most obviously) diffusing outwardly, and corrosion product spallation. This resulted
in a quasi-steady state behavior where chromium is not fully depleted in corrosion scales.
Corrosion proceeded uniformly on 347SS with no intergranular attack (Figure 19). XRD and
line scan indicate typical corrosion products in the form of sodium ferrite and other iron oxides.
321SS had indication of preferential chromium depletion along grain boundaries, as observed in
x-ray maps in Figure 20. Line scan results yield significantly less sodium content than 347SS
corrosion products, which may be the result of higher spallation [4].
24
Table 6: Corrosion Rates and projections for iron based alloys
Alloy Descaled Metal
Exponent Fit k
Alloy Density Loss* Loss*
(n) [mg/ cm2-secn]
[g/cm3] [mg/cm2] [µm/yr]
321SS 7.94 - - - -
400°Ca - 0.29 (R2=1.0) 2300e-6 0.27 1
500°Ca - 0.85 (R2=0.96) 2e-6 1.98 7.1
600°C1 - - - - 15.9
680°Cb - - - 42.77 460
347SS 8.03 - - - -
400°Ca - 2
0.24 (R =0.99) 4100e-6 0.20 0.7
500°Ca - 2
0.59 (R =0.94) 95e-6 1.28 4.6
600°C1 - - - - 10.4
680°Cb - - - 42.05 447
a. Based on 3064 hour data (Equation 1).
b. Based on 1025 hour data
1. Reference [4]after 3000 hours
Figure 15: Weight gain and descaled loss of iron based alloys for temperatures of 400°C
and 500°C.
25
Figure 16: Weight gain (left) and descaled weight loss (right, 1025 hour only) data for
alloy 347SS and 321SS at 680°C.
Figure 17: 347SS after 3064 hours at 500°C, iron oxide with no formation of NaFeO2.
Figure 18: 321SS after 3064 hours exposure at 500°C. No detection of sodium, with XRD
scans indicating magnetite formation.
26
Figure 19: 347SS after 1025 hours exposure at 680°C. Phases of NaFeO2 were detected
with XRD after 520 hours (bottom left), with mixed oxide detected at 1025 hours (bottom
right). Corrosion proceeded uniformly with no obvious grain boundary depletion of
chromium.
27
Figure 20: 321SS corrosion at 1025 hours at 680°C. Slight Cr depletion was observed
along grain boundaries, as indicated by X-ray map (bottom left).
C. Ni-Cr alloys (HA230 and IN625)
i. 400-500°C corrosion
Nickel based alloys HA230 and In625 were investigated at 400, 500, and 680°C. Corrosion was
so insignificant at temperatures of 400°C and 500°C that it was difficult to obtain any
information regarding the oxidation behavior through typical SEM/XRD methods. Corrosion at
500°C did result in chromium depletion within the surface oxide (Figure 21). Micrographs of
alloy In625 are not included as the oxide scale was not adequately resolved by SEM.
Corrosion rates are effectively parabolic in nature, indicating the presence of a protective oxide.
HA230 was found to lose weight in a parabolic fashion, which may be due to chromium
dissolution into the melt. Table 7 has relevant fits and estimated annual metal loss values
determined in this study and previous work [1, 4]. It was determined that In625 performed
slightly better than HA230. Both alloys develop thin layers of NiO, shown by XRD and EDS
data in Figure 21, even at relatively low temperatures. Figure 22 shows the different progression
28
in corrosion products from base alloy through 500 and 680°C for In625. The observed
difference in performance between these alloys is likely due to relative concentrations of soluble
metals in the alloy, as HA230 has 14wt% tungsten.
ii. 680°C corrosion
Oxidation was severe for both alloys at 680°C. Table 7 shows a comparison of metal loss for
alloys at several temperatures and conditions, including experiments performed at Sandia and
elsewhere. Previous data investigating In625 in nitrate salts is available and an attempt was
made to compare, contrast, and understand if any conclusions could be made. The most similar
experiment using In625 to this test found performance that was 20-23% different [3], though this
is acceptable given the limited statistics available.
Corrosion of In625 increases dramatically above 650°C. Increase in corrosion rate is the result
of two major parameters: temperature difference and salt chemistry. Tests investigating
temperatures of 630 and 650°C employed an O2 overpressure to promote a difference in molten
salt equilibrium [1]. Direct comparison of nitrite equilibriums in both oxygen covered melts
were approximately equivalent to values predicted at 600°C in air, approximately 4-5wt%, in
addition to corrosion mechanisms being similar. These similarities motivated the use of
accelerated aging criteria, thus determining activation energies. Molten chemistry of nitrate salts
is quite complex [22, 23] with changes in ullage gas resulting in a variety of different melt
chemistries. Presence of an O2 cover gas is speculated to influence relative concentrations of
nitrite, nitrate, oxide, peroxide, superoxide, hydroxide [22], though to what extent this seems to
be controversial in the literature. Nitrite concentration, as one point of comparison, was nearly
more than a factor of two higher in air sparged 680°C melts. Significant changes in salt
chemistry do not lend toward a direct comparison of performance and more work is need for
similar chemistry conditions to understand the effect of corrosion at such high temperatures.
Alloy HA230 formed a complex oxide structure, consisting of several distinct layers (Figure 25).
Layers in direct contact with the salt tended to be rich in iron, sodium and oxygen. Largely this
was difficult to identify and trace peaks of magnetite were found in diffraction data. Formation
of NiO was present, confirmed by both XRD and EMP. Phases of metallic nickel separated a
comparatively thick layer of internal oxidation. This layer of oxide was determined to be
chromia (Cr2O3) after first removing surface oxide and utilizing XRD, shown in Figure 26. This
behavior can be understood by considering the Gibbs free energy of formation which favors
chromia formation in the presence of low partial pressures of oxygen[24]. The formation of
internal oxidation on HA230 was observed at 600°C [4], while no internal oxidation was found
present in other nickel alloys (HA242, In625, or HA214). One major difference in HA230 is
ternary nature of the alloy (Ni-22Cr-14W) and other work noticed susceptibility of Ni-W alloy to
internal oxidation at high temperatures[16].
Data in Figure 26 confirmed suspicions that the internal oxidation was not included in the
corrosion rate data for Figure 24 and Table 7 and will need to be considered from a design
standpoint, which will result in more metal loss than determined through descaling analysis.
29
No internal oxidation was observed in In625 (Figure 27). The corrosion layer in contact with the
melt had similar compositions of iron, sodium, and nickel when compared to HA230. Further
similarities in NiO and metallic nickel were observed.
Figure 21: 500°C exposure for 3064 hours on HA230. Surface corrosion at lower
temperatures could not be adequately resolved by microscopy methods used, though
XRD indicated formation of NiO.
30
Figure 22: In625 corrosion product evolution based on exposure temperature
Table 7: Corrosion Rates and projections for nickel based alloys
Alloy Descaled Metal
Exponent Fit k
Alloy Density Loss* Loss*
(n) [mg/ cm2-secn]
[g/cm3] [mg/cm2] [µm/yr]
HA230 8.97 - - - -
400°Ca - 0.19 (R2=0.87) 7300e-6 0.15 0.5
500°Ca - 0.48 (R2=0.92) 260e-6 0.55 1.8
600°C1 - - - - 19.81
680°Cb - - - 72.3 688
In625 8.44 - - - -
400°Ca - 0.12 (R2=0.48) 7900e-6 0.05 0.2
500°Ca - 0.19 (R2=0.83) 6900e-6 0.15 0.5
600°C1 - - - - 12.71
600°C4 6.74 21.74
630°C2 - - - 11.02 40.92
650°C2 - - - 19.1 2
71.02
677°C3 - - - - 7363
680°Cb - - 58.7 594
a. Based on 3064 hour data (Equation 1).
b. Based on 1025 hour data
1. Reference [4]
2. SAND2001-8758 (at 2794 hours) Ref. [1]; Oxygen cover gas used
3. Ref. [3] after 14 day exposure
4. Ref. [25] 3200 hour exposure (0.05% chloride content); thermally cycled
– no chemistry control of molten salt used.
31
Figure 23: Weight gain and descaled information for HA230 and In625 at 400°C and
500°C.
Figure 24: Weight gain data for 680°C exposure of HA230 and In625 (left). Descaled data
(right) for 1025 hour exposure only.
32
Figure 25: HA230 after 1025 hours at 680°C. Line scan and XRD data clearly indicates
phases of NiO and metallic nickel, with small peaks that appeared to correspond to iron
oxide.
Figure 26: XRD scan of HA230 after descaling process was performed on 1025 hour,
680°C sample. No base alloy was detected, only phases of chromia and nickel, clearly
indicating internal oxidation present.
33
Figure 27: In625 after 1025 hour exposure at 680°C. Nickel and NiO were present as
primary corrosion products.
34
35
36
4. CONCLUSIONS
Corrosion assessment of four steels (321SS, 347SS, HA230, and In625) was completed for three
temperatures; 400, 500, and 680°C. This test was a follow-on test to complement a previous
experiment at 600°C.
Salt chemistry measurements had nitrite concentrations slightly lower than expected by
equilibrium calculations. No significant amounts of metals were found in 400°C and 500°C salts.
Significant amounts of chromium, molybdenum, and tungsten were dissolved into the 680°C
melt. Low concentrations of nickel were detected in the melt at this temperature.
Corrosion at temperatures of 400°C and 500°C was low for all alloys investigated. Nickel based
steels resulted in slightly less corrosion. Performance at this temperature, best to worst, is as
follows: In625 > HA230 > 347SS > 321SS. The nickel alloys both form protective NiO, which
does well in protecting the base alloy. Iron based alloys form an iron oxide, which tend to be
less protective in practice than nickel oxide.
Corrosion at 680°C was significant as metal wastage occurred. Iron based alloys outperformed
nickel alloys by 30% to 50%. Soluble transition metals, specifically Cr, Mo, and W, were found
in appreciable concentrations in the melt and are thought to be the driving mechanism in
performance at 680°C. NiO was the primary corrosion product for both nickel alloys, though
phases of metallic nickel were also detected. HA230 exhibited internal oxidation in the form of
a nickel/chromium oxide, while In625 exhibited primarily NiO. This internal oxidation may not
be accounted for in corrosion rates and further microscopy will provide insight. Iron based
alloys out performed nickel alloys under these harsh conditions.
No attempt was made to determine activation energy for alloys tested here. Corrosion
mechanisms between low temperature conditions were different than mechanisms found at
600°C. Further data points in the range of 600-680°C are needed to complete an activation
energy study that could be used for accelerated aging predictions, where salt chemistry
conditions are appropriately similar.
37
5. REFERENCES
1. Bradshaw, R.W. and S. Good, Accelerated corrosion testing of a nickel-base alloy in a
molten salt, 2001, SAND2001-8758, Sandia National Laboratories, Livermore C. A.
2. Pacheco, J.E., et al., Final test and evaluation results from the Solar Two project, 2002,
SAND2002-0120, Sandia National Laboratories, Albuquerque N. M.
3. Slusser, J.W., et al., CORROSION IN MOLTEN NITRATE-NITRITE SALTS. Journal of
Metals, 1985. 37(7): p. 24-27.
4. Kruizenga, A.M., D.D. Gill, and M. LaFord, Materials Corrosion of High Temperature
Alloys Immersed in 600°C Binary Nitrate Salt, 2013, SAND2013-2526, Sandia National
Laboratories, Livermore C.A.
5. Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens, in
G1-032003, ASTM International.
6. Gill, D.D., et al., DESIGN, FABRICATION AND TESTING OF AN APPARATUS FOR
MATERIAL COMPATIBILITY TESTING IN NITRATE SALTS AT TEMPERATURES UP
TO 700°C, in 5th International Conference on Energy Sustainability2011, ASME:
Washington, DC, USA. p. 6.
7. Jones, D.A., Principles and prevention of corrosion1996, Upper Saddle River, NJ:
Prentice Hall.
8. Takahashi, Y., R. Sakamoto, and M. Kamimoto, Heat capacities and latent heats of
LiNO3, NaNO3, and KNO3. International Journal of Thermophysics, 1988. 9(6): p. 1081.
9. Bradshaw, R.W. and D.E. Meeker, High-temperature stability of ternary nitrate molten
salts for solar thermal energy systems. Solar energy materials, 1990. 21(1): p. 51-60.
10. Nissen, D.A., Thermophysical properties of the equimolar mixture sodium nitrate-
potassium nitrate from 300 to [Link].C. Journal of Chemical & Engineering Data,
1982. 27(3): p. 269-273.
11. Nissen, D.A. and D.E. Meeker, NITRATE NITRITE CHEMISTRY IN NANO3-KNO3
MELTS. INORGANIC CHEMISTRY, 1983. 22(5): p. 716-721.
12. Afanasiev, P. and D.H. Kerridge, Reactivity of V2O5, MoO3 and WO3 in molten KNO3,
studied by mass spectrometry. Journal of Alloys and Compounds, 2001. 322(1-2): p. 97-
102.
13. Oryshich, I.V. and O.S. Kostyrko, INFLUENCE OF MOLYBDENUM, TUNGSTEN,
AND COBALT ON THE CORROSION OF HIGH-TEMPERATURE STRENGTH
NICKEL ALLOYS IN MOLTEN SALTS. Metal Science and Heat Treatment (English
Translation of Metallovedenie i Termicheskaya Obrabotka, 1985. 27(9-10): p. 740-746.
14. Yurkinskii, V.P., et al., INTERACTION OF MOLYBDENUM AND TUNGSTEN WITH
MOLTEN NITRATES OF ALKALI METALS. Journal of applied chemistry of the USSR,
1984. 57(3 pt 1): p. 476-479.
15. Deloume, J.-P., et al., Study of the reaction of tungsten carbide in molten alkali metal
nitrates. Syntheses of divalent (s and d blocks) metal tungstates. Journal of Solid State
Chemistry, 2003. 174(1): p. 1-10.
16. Ganster, P., B. Pujilaksono, and K. Wolski. Kinetics and mechanism of high temperature
internal oxidation of Ni-14wt%W alloy at 1000C. in 8th International Conference on
Diffusion in Materials, DIMAT 2011, July 3, 2011 - July 8, 2011. 2012. Dijon, France:
Trans Tech Publications Ltd.
38
17. Rogers, D.J. and G.J. Janz, MELTING-CRYSTALLIZATION AND PREMELTING
PROPERTIES OF NANO3-KNO3. ENTHALPIES AND HEAT CAPACITIES. 1982. V
27(N 4): p. 424-428.
18. Speidel, P.J., et al., Performance of the solar two central receiver power plant. Journal
De Physique. IV : JP, 1999. 9(3): p. Pr3-181 - Pr3-188.
19. Kearney, D., et al., Assessment of a molten salt heat transfer fluid in a parabolic trough
solar field. Journal of Solar Energy Engineering, Transactions of the ASME, 2003.
125(2): p. 170-176.
20. Bradshaw, R.W. and S.H. Goods, Corrosion of alloys and metals by molten nitrates,
2001, SAND2000-8727, Sandia National Laboratories, Livermore C. A.
21. Goods, S.H., et al., Corrosion of stainless and carbon steels in molten mixtures of
industrial nitrates, 1994, SAND94-8211, Sandia National Laboratories, Livermore C. A.
22. Lovering, D.G., Molten salt technology1982, New York: Plenum Press.
23. Picard, G.S., T.A. Flament, and B.L. Tremillon, ACIDITY AND THERMOCHEMICAL
STABILITY OF MOLTEN SODIUM NITRATE AND NITRITE MIXTURES. Journal of
The Electrochemical Society, 1985. 132(4): p. 863-868.
24. Barin, I. and O. Knacke, Thermochemical properties of inorganic substances1973,
Berlin; New York: Springer-Verlag.
25. Bradshaw, R.W. and S.H. Goods, Corrosion resistance of nickel-base alloys in molten
alkali nitrates, 2000, SAND2000-8240, Sandia National Laboratories, Livermore C. A.
39
40
APPENDIX:
A. Sample pull Schedule and Order
Samples were removed using a sample interval method. Four additional alloys were tested using
this method, 625SQ, P91, Cr2-1/4 and HR120, but were not analyzed due to budgetary
constraints.
Figure 28: 400°C removal schedule and sample ID
Figure 29: 500°C removal schedule and sample ID
41
Figure 30: 680°C removal schedule and sample ID
B. SALT SPECIFICATIONS
42
DISTRIBUTION
External (electronic)
1 Joseph Stekli
US DOE Office of the Solar Energy Technology Program
Route Symbol: EE-2A
Building: L'ENF950
Washington, DC 20585
1 Bruce Kelly
Abengoa Solar
11500 West 13th Avenue
Lakewood, CO
1 David Wasyluk
Babcock & Wilcox Power Generation Group
20 S. Van Buren Avenue
Barberton, Ohio 44203
1 Greg Kolb
SolarReserve
1205 Lawrence Ct. NE
Albuquerque, NM 87112
1 Lee Flowers
Haynes International, Inc.
1020 West Park Avenue
Kokomk, IN 46904-9013
1 Matthew Bender
ATI Allegheny Ludlum
1300 Pacific Avenue
Natrona Heights, PA 15065, USA
1 Craig Tyner
eSolar
3355 W. Empire Avenue
Suite 200
Burbank, CA 91504
1 Trent Maxwell
Alstom Power
1201 Riverfront Parkway
Chattanooga TN 37402
43
Internal (electronic)
1 MS1127 David Gill 6123
1 MS1127 Subhash Shinde 6123
1 MS1127 William Kolb 6123
1 MS9153 Timothy Shepodd 8200
1 MS9402 Ryan Nishimoto 8651
1 MS9403 Alan Kruizenga 8223
1 MS9403 Adam Rowen 8223
1 MS0899 Technical Library 9536
44
45