Report on metrological infrastructure to support CCUS processes

21GRD06 MetCCUS Report on the metrological infrastructure comprising the instruments adopted to determine the thermophysical properties of mixtures of interest for CCUS processes. The report will include at least 2 simplified formulations for the prediction of the density of typical CO2 mixtures (in liquid and gas phase), suitable for use by flow computers to support the CCUS process design. D8 - A4.3.6 Organisation name of the lead participant for the deliverable: Istituto Nazionale di Ricerca Metrologica S. Lago (INRiM), P. Alberto Giuliano Albo (INRiM), F. Gugole (VSL), S. Nadiri (PTB), S. Bell (NPL), P. Carroll (NPL), S. Kairy (NPL), S. Zhou (NPL), M. Thol (RUB), R. Span (RUB), A. Ortega (UVa), X. Mendez (UVa), A. Fateev (DTU), L. Chaubet (Air Liquide FuE) Due date of the deliverable: 30.09.2025 Actual submission date of the deliverable: 07.10.2025 Confidentiality Status: PU - Public, fully open (remember to deposit public deliverables in a trusted repository) Deliverable Cover Sheet Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or EURAMET. Neither the European Union nor the granting authority can be held responsible for them. The project has received funding from the European Partnership on Metrology, co-financed from the European Union’s Horizon Europe Research and Innovation Programme and by the Participating States. 1 of 16 TABLE OF CONTENTS 1 Summary 3 2 The meteorological infrastructure for properties of CO2 thermophysical systems 4 4. A dedicated equation for designing CO2 transport infrastructures 5 5. CO2 gas mixtures: a comparison among EOS-CG-2021, Soave-Redlich-Kwong (SRK) and Peng-Robinson (PR) equations of state 7 5.1. Numerical comparison on CO2 mixtures relevant for the CCUS industry 8 6. A dedicated formulation for designing carbon capture systems 10 7. Conclusion 14 8. Formulations available on Zedono 14 2 of 16 21GRD06 MetCCUS 1 Summary With the aim to validate models needed to support the design of CCUS processes and to provide the necessary support to flow metering operations involving CO2 mixtures in the liquid phase, vapour phase or in supercritical conditions, significant progress was made on the study of amines, their mixtures with CO2 and H2O, and for transportation of CO2 mixtures in pipelines. This report describes the characteristics of the metrological infrastructure implemented during the development of the MetCCUS project with the aim to obtain the thermophysical properties of CO2 mixtures of interest for industrial applications. To provide a simple access to the measured properties, two simplified equations of state were implemented: one for designing CO2 transport infrastructure and one for modeling carbon capture processes using aqueous amines. 3 of 16 21GRD06 MetCCUS 2 The meteorological infrastructure for properties of CO2 thermophysical systems With the aim to provide procedures, measurements and the specific instrumentation needed to build up confidence in the safe and efficient operation of pipelines transporting CO2 mixtures, innovative methods for the online monitoring of eventual phase transitions in the CO2 through the measurements in-site were developed. In particular, two spectroscopy-based systems were developed for online monitoring of phase transitions in CO2 such as gas to super-critical fluid flows. The first system is based on far-UV method and the second one on NIR method. Both systems were successfully tested at laboratory conditions with use of a commercial CO2 gas cylinder with deep tube and validated at ExtrateX company (Nancy, France). The measurements at ExtrateX have been done on a proprietary dedicated setup, where it was possible to utilize gas-, super-critical or liquidphase CO2 flows. The far-UV and NIR systems have been coupled to the ExtrateX’s setup. The far-UV and NIR measurements of CO2 gas-super-critical phase transitions at ExtrateX agree with the laboratory measurements. DTU is investigating the possibilities for a patent application for the NIR method. Moreover, considering that even a small amount of water can chemically react with the CO2 to form corrosive compounds, part of the contribution was dedicated to the development of a calibration method for the online humidity sensors that are used in CCUS processes. NPL have developed a new facility for corrosion testing of pipeline steel in impurity-containing dense phase CO2 environment. Methods for controlling and measuring the concentration of impurities (H2O and O2) in dense phase CO2 during corrosion testing were successfully optimised and validated. A robust test method was successfully developed to assess the corrosion rate of X65 pipeline steel in dense phase CO2 containing H2O and O2 at 80 bar and 25 °C. Key steps of the method include: 1. Deaeration: Adequate deaeration (i.e., O2 below 1 ppmv) at ambient pressure requires purging the autoclave with CO2 with a volume at least 15 times greater than the autoclave volume. This protocol prevents exposure of test specimens to residual atmospheric O2 during testing, thereby avoiding artefacts in corrosion rate measurements. 2. Test Environment Establishment: Gradual pressurisation (~ 40 minutes) combined with continuous stirring at 700 rpm enables rapid and uniform distribution of impurities within the autoclave, achieving target concentrations near test specimens within 1 hour. This procedure minimises uncertainty in exposure time and test environment composition, which is critical for short-duration corrosion tests. 3. Depressurisation: Diluting the test environment with pure CO2 to reduce H2O concentration below 500 ppmv, followed by controlled depressurisation (1 L/min or 3 L/min), effectively prevents artefacts in corrosion rate measurements arising from unintended H2O condensation on test specimens. Corrosion tests on X65 steel in H2O-saturated dense phase CO2 conducted using the developed robust test method demonstrated good repeatability. The corrosion products were uniform and semi-protective, with no significant difference between horizontally and vertically orientated specimens. This result contributed to the development of an international best practice guide: 4 of 16 21GRD06 MetCCUS AMPP Guide 21577 - Laboratory Corrosion Testing for CO2 Transport and Injection, supporting standardisation efforts in corrosion testing of CO2 infrastructure. A further activity was dedicated to the implementation of the standard test procedures necessary to ensure the safe and cost-effective selection of materials for CO2 pipelines. A robust and reproducible test method was developed that is more representative of conditions in service than conventional test methods that have been employed to date in the literature. NPL has adapted the use of an existing commercially made humidity generator to operate with CO2. A Thunder 3900 Low Humidity generator was adapted so that it can dynamically generate humidity values in CO2 in the frost-point temperature range -60°C to 0 °C (equivalent to water vapour amount fraction range 10 µmol mol-1 to 0.5 %). This involved considering materials compatibility, operating range relevant to CCUS, gas-specific flow measurement and control, the CO2 phase diagram, safe exhaust of the gas, and establishing that operation in the ice regime minimises any issue of dissolved CO2. For CO2, there is little published literature for water vapour enhancement factor, but NPL has identified available data and used this to enable conversion from frost point to amount fraction. The primary traceability to dew-point temperature units (°C) is via a reference platinum resistance thermometer (PRT) calibrated against NPL Temperature Standards. The adapted standard was used to calibrate a chilled-mirror hygrometer in terms of frost point in both CO2 and air. Finally, an experimental apparatus comprising an acoustic/microwave resonant cavity and a simplified saturator was designed and realized in order to be able to measure speed of sound and refractive index of gaseous pure CO2 and CO2 + H2O mixtures. Test measurements were carried out in Ar in order to evaluate the performances of it. Using the novel experimental system, comprising an acoustic/microwave resonant cavity and a simplified saturator, measurements of speed of sound and refractive index of gaseous pure CO2 and humid CO2 at two isotherms (323 and 330) K, spanning the overall pressure range between 0.5 and 1.2 MPa with water mole fractions 0.5% and 1%, respectively, were carried out; besides along an isochoric line at 0.7 MPa at 11 temperatures between 323 K and 357 K, with a water mole fraction x = 1.0 % (by mass). With the aim to estimate the interaction carbon dioxide – water virial coefficient, measurements will be continued spanning beyond the lifetime of the project. 4. A dedicated equation for designing CO2 transport infrastructures Multiparameters equations of state have proven to be able to provide predictions of the thermodynamic properties of mixtures with an accuracy that is comparable to the one of the experimental data when their parameters are adjusted using available experimental measurement results. However, the update of such equations requests time, especially when they are needed to describe multicomponents mixtures. For this reason, while updated formulations will be available, a simplified model is proposed to support the design of CO2 compression, liquefaction and pumping systems. More than the accuracy of the formulation, its simplicity was privileged so that it can be easily implemented in openand closed-source designing software platforms. 5 of 16 21GRD06 MetCCUS Considering that CO2 applications can range from low to high temperature cycles involving necessary or sometimes undesired phase transitions, it has been chosen to adopt cubic equations of state optimized in the range of CCUS processes. Historically, cubic equations are utilized for designing hydrocarbon distillation and their coefficients are calculated specifically to reproduce the saturation curve and the saturated liquid density. However this is not the main scope of the CO2 transport applications where the estimation of thermophysical properties of CO2 mixtures are necessary to define dimensions and range of operativity of compressors and pumping systems. So that, the coefficients of a Patel-Teja equation of state were recalculated for pure CO2, O2, SO2, Ar, CO and N2 to optimize the predictions in the range of temperature between 263 K and 423 K, and for pressure up to 12 MPa. Available experimental measurements of mixtures of the selected components were used to determine the binary interaction coefficients and then they have been used to predict the density of a quaternary mixture composed by CO2 + N2 + O2 + Ar. Comparisons of the implemented model with Soave-Redlich-Kwong (SRK), Peng-Robinson (PR) cubic equation and with EOS-CG-2021 fundamental equation is provided. Details of the adopted Patel-Teja (PT) equation can be found in [Patel et al.] while only necessary information to implement this model are hereafter reported. PT equation is a cubic equation explicit in pressure p as a function of the temperature T and molar volume v: 𝑝(𝑇,𝑣)= 𝑅 𝑇 𝑣−𝑏 −𝑎(𝑇) 𝑣(𝑣+𝑏)−𝑐(𝑣−𝑏) (1) where R=8.31446 J/(mol K) is the molar gas constant, b and c are the co-volumes and a(T) is the thermal function. The parameters of the equation are the critical temperature Tc, and the critical pressure pc while the coefficients to be fitted are ζc and one, or more, coefficients used to define the function a(T) not reported in the equation (1). From those quantities a, b and c can be obtained. However, for a sake of simplicity, final expressions for a, b and c are explicitly given in table 1. Equation (1) has the property to become an SRK equation when c=0 and a PR equation when c=b. Thanks to this flexibility, PT equations perform, usually, a bit better than both SRK and PR but its quality remains in the range of expected capability of cubic equations thus non comparable with performances of multiparameters equations of state like BWR and equations in Helmholtz free energy. To calculate the density expressed in kilogram per cubic meter, the pressure must be expressed in pascal and the temperature in kelvin. Solving equation (1), the molar volume v, expressed in cubic meters per mole, is obtained while the corresponding density can be calculated as ρ=m/v, when the molar mass m is expressed in kilograms per mole. To predict the properties of a multicomponent mixture, at least the binary interaction parameters are needed when the mixing rules described in Patel et al. are adopted. Thanks to the stability of cubic equations, the determination of interaction coefficients can be usually performed using very few experimental points since they are approximated by constants and thus are not dependent on the temperature, as reported in table 1. In the proposed approximation, the coefficients not reported in the table have been removed from the model since this simplified approximation considers only mixtures with a molar fractions of CO2 higher than 90 %. No theoretical formulations for evaluating the binary interaction coefficients were adopted. Values reported in table 1 seem unreasonable since expected values should be positive and in a boundary of 0.1. This is probably 6 of 16 21GRD06 MetCCUS a side of having removed some interactions that cannot be considered negligible. This problem will be corrected in future contributions. Table 1. Binary interaction coefficients k12 to be used in the term (1-k12) of Patel-Teja mixing rules. Mixture k12 Mixture k12 Mixture k12 CO2-CO –0.2231 CO2-O2 0.1860 CO2-SO2 -0.0894 CO2-N2 0.8833 CO2-Ar -4.9065 5. CO2 gas mixtures: a comparison among EOS-CG-2021, Soave-Redlich-Kwong (SRK) and Peng-Robinson (PR) equations of state Correctly estimating the thermodynamic properties of a real gas is a challenge that scientists have been trying to solve for more than a century. A milestone in that direction was achieved by van der Waals and his proposed equation of state (EOS) for real gases and liquids (van der Waals, 1873). This equation was the first EOS relating the pressure, volume, number of molecules and temperature in a fluid. The Van der Waals EOS formulation is cubic in volume and was the first EOS that could represent a two-phase system (i.e., liquid and gas). Since then, many modifications to this equation have been proposed to, e.g., enhance the phase behaviour predictions of the model. Currently, the equation proposed by Soave-Redlich-Kwong (SRK) (Soave, 1971) and the one presented by Peng-Robinson (PR) (Peng & Robinson, 1976) are the two cubic EOS that are used the most both in the industry and in the scientific community. These EOS find ample application thanks to their ease of use and low computational requirements although do not always reach the required level of accuracy. An alternative to cubic EOS is to use Helmholtz-energy based EOS. One of the most notable examples of this class of EOS is given by the GERG-2008 EOS for natural gas mixtures (Kunz & Wagner, 2012). EOSs based on the Helmholtz energy include a multicomponent mixture model based on the combination of binary mixture models of the constituents. Each binary-specific model is built on highly accurate Helmholtz energy equations of state for pure fluids, which are then combined with interaction parameters fitted to experimental data. The level of tuning for each binary system thus depends on the availability of experimental data. However, as more and more data become available, EOS formulations based on the Helmholtz energy become more and more accurate as it was the case for the GERG-2008 EOS. The CCUS (carbon capture, utilisation and storage) industry has been widely using cubic EOS however CCUS mixtures often require the modelling of chemical association and polar components, which is not properly captured by cubic EOS. It is possible to include such modelling in cubic EOS, at the expense of increasing the complexity of the EOS and limiting the ability of the EOS to extrapolate beyond experimental data, thus losing some of the attraction of cubic EOS. In recent years, the scientific community has developed a Helmholtz energy EOS specific for CCUS-mixtures, EOS-CG (Gernert & Span, 2016), which is being updated and constantly improved as more and more experimental data on CO2 mixtures with relevant impurities are being 7 of 16 21GRD06 MetCCUS collected. It is important to mention that although GERG-2008 includes many CCUS-relevant compounds, these are treated as minor components in the formulation of the GERG-2008. Therefore, when using GERG-2008 for the calculation of thermodynamic properties of a CCUS-relevant mixture the uncertainties associated with the EOS result can exceed those of the experimental measurements. In MetCCUS, it has been performed a numerical comparison of the predictions provided by SRK, PR and EOS-CG-2021 against experimental measurement results available in the literature on CO2 mixtures (with CO2 as the most abundant component). For the comparison, we use SRK, PR and EOS-CG as implemented in TREND 6.0 (Span, et al., 2025). 5.1. Numerical comparison on CO2 mixtures relevant for the CCUS industry As a first comparison, we analyse the predictions of EOS-CG, SRK, PR and PT (as newly implemented in this project) on binary mixtures of carbon dioxide and sulfur dioxide, oxygen or carbon monoxide. The presence of any of these three impurities may have fatal consequences for the CO2 transportation system causing, e.g., corrosion and running ductile fraction (Simonsen, Hansen, & Pedersen, 2025). It is therefore of vital importance to monitor the presence of these impurities in CCUS processes. The experimental data used for the comparison are given in (Gimeno, Artal, Velasco, Fernández, & Blanco, 2018) for the mixture with sulfur dioxide, (Mantovani, Chiesa, Valenti, Gatti, & Consonni, 2012) for the mixture with oxygen, and (Souza, Herrig, Span, & Trusler, 2019) for the mixture with carbon monoxide. The comparison is performed in terms of the average of absolute deviations (AAD), which is calculated as 𝐴𝐴𝐷= 100 𝑁𝑖=1 𝑁 ∑|ρ𝑒𝑥𝑝, 𝑖−ρ𝐸𝑂𝑆,𝑖| ρ𝑒𝑥𝑝,𝑖 (2) where denotes the experimental value of the density of the CO2 mixture, and the density ρ𝑒𝑥𝑝 ρ𝐸𝑂𝑆 value predicted using an equation of state. We calculated the AAD per phase, where we estimated the critical point based on the characteristics of the pure components and on the proportion of the components. The results are summarized in Table 2. In most cases of the mixtures with sulfur dioxide and oxygen, the EOS-CG-2021 (Neumann, Herrig, Bell, Beckmüller, Lemmon, Thol & Span, R., 2023). provides predictions with a lower AAD than the cubic EOS. However, in the case of the mixtures with carbon monoxide EOS-CG-2021 returns considerably larger AAD values, while the fitted PT equation returns AAD values similar as for the other mixtures.These results emphasize that some of the binary models (e.g., for carbon monoxide) in EOS-CG-2021 need more data to better represent the interactions between the different components (or at least in the temperature and pressure ranges here considered). 8 of 16 21GRD06 MetCCUS Table 2: AAD calculated per phase and per EOS on binary mixtures of CO2 with CCUS relevant impurities. For all mixtures, the amount fraction of CO2 is calculated as cmol/mol where XX is either 𝑥𝐶𝑂2=100−𝑥𝑋𝑋 SO2, O2 or CO. Mixture Phase EOS-CG-2021 SRK PR PT 𝑥𝑆𝑂2=.69 cmol/mol Gas 0.98 1.59 1.63 1.35 Liquid 0.36 11.29 2.59 2.28 𝑥𝑆𝑂2=4.68 cmol/mol Gas 3.19 3.72 2.36 2.22 Liquid 2.84 12.69 3.16 2.04 𝑥𝑂2=6.07 cmol/mol Gas 1.44 1.79 2.17 2.30 Supercritical 1.70 8.81 2.16 2.31 𝑥𝑂2=12.91 cmol/mol Gas 1.72 1.90 2.75 2.25 Supercritical 2.54 7.78 1.46 2.25 𝑥𝐶𝑂=5.031 cmol/mol Gas 7.31 8.00 6.22 1.25 Liquid 3.49 12.53 4.40 2.35 Supercritical 13.52 20.22 14.34 2.46 𝑥𝐶𝑂=10.107 cmol/mol Gas 4.13 5.06 3.12 1.78 Liquid 6.27 14.14 5.75 2.13 Supercritical 8.10 15.02 8.94 1.95 Alongside the EOS predictions for binary mixtures, we analysed also the predicted density for a more complex mixture of carbon dioxide ( cmol/mol), nitrogen ( cmol/mol), 𝑥𝐶𝑂2=89.83 𝑥𝑁2=5.05 oxygen ( cmol/mol) and argon ( cmol/mol) for which experimental results are 𝑥𝑂2=3.07 𝑥𝐴𝑟=2.05 given in (Nazeri, Chapoy, Burgass, & Tohidi, 2017). These impurities can seriously compromise the integrity of the pipelines and thus increase the risk of failure of CO2 transportation systems. Results are reported in Table 3. In this case, the fitted PT EOS has the lowest AAD for all the three phases considered, with EOS-CG-2021 returning similar AAD values in two out of three phases. However, the AAD values are in general quite high, as seen also for the binary mixture CO2 + CO and might not meet the accuracy requirements needed in the CCUS industry. Thus more work should be done to improve the binary-specific models in EOS-CG-2021 or the binary interaction parameters in PR or in the proposed PT EOS. 9 of 16 21GRD06 MetCCUS Scholz, C. W. & Span R., Measurement of the (p, ρ, T) Behavior of Liquid MEA and DEA at Temperatures from (293.15 to 423.15) K and Pressures up to 90 MPa, Int. J. Thermophys. 42 (2021). DOI: 10.1007/s10765-021-02808-x. Simonsen, K. R., Hansen, D. S., & Pedersen, S. (2025). Framework for CO2 impurity monitoring in CCUS infrastructure. Carbon Capture Science & Technology, 100453. Soave, G. (1971). Equilibrium constants from a modified Redlich-Kwong equation of state. Chemical Engineering Science, 1197-1203. Souza, L. F., Herrig, S., Span, R., & Trusler, J. P. (2019). Experimental density and an improved Helmholtz-energy-explicit mixture model for (CO2+CO). Applied Energy, 113398. Span, R., Beckmüller, R., Buchenfeld, J., Fiedler, F., Jäger, A., Mickoleit, E., . . . Thol, M. (2025). TREND. Thermodynamic Reference and Engineering Data 6.0. Bochum: Lehrstuhl für Thermodynamik, Ruhr-Universität Bochum. Téllez, P., Medeiros, M., (2013). Modeling CO2 and H2S solubilities in aqueous alkalonamine solutions via an extension of Cubic-Two_State equation of state, Fluid Phase Equilibria 344, pp. 45-58. van der Waals, J. D. (1873). On the Continuity of the Gaseous and Liquid States - Ph. D. Thesis. Leiden: University of Leiden. 16 of 16