Pressurized Chemical Looping for Direct Reduced Iron Production: Carbon Neutral Process Configuration and Performance

Abstract

To achieve net-zero iron and steel production by 2050, many iron and steel producers are turning to direct reduced iron (DRI)—electric arc furnace (EAF) steel production as an opportunity to achieve significant CO₂ emissions reductions relative to current levels. However, additional innovations are required to close the gap between DRI and net-zero steel. Pressurized chemical looping-DRI (PCL-DRI) is a novel technology explored to meet this target, in which the reformer firebox and fired process gas heaters are replaced with PCL combustion units. Captured CO₂ is conditioned and compressed for pipeline transportation and storage/utilization. The performance of two different PCL-DRI configurations relative to traditional DRI processes was explored via process simulation: a Midrex-type process and an Energiron-type process. The PCL-DRI processes were shown to have equivalent or lesser total fuel consumption (8% reduction) compared to the base cases, and greater process water production (170–260% increase), with minimal or no loss in thermal efficiency. PCL-DRI is a strong competitor to alternative methods of reaching net-zero DRI due to lower energy penalties for carbon capture, no required changes to stream chemistry in or out of the EAF, and no requirement for hydrogen infrastructure.

1. Introduction

1.1. Background

Ironmaking is one of Canada’s top CO₂ emitting industries. Canada ranked eighth in global iron ore production, producing 52.4 Mt of iron ore and 11 Mt of crude steel in 2020 [1]. Globally, iron and steel production results in 4–8% of anthropogenic CO₂ emissions, with specific emissions ranging from 1.4–3.5 t CO₂/t liquid steel [2]. Canadian iron and steel producers have committed to produce net-zero iron and steel by 2050, as a part of the global target to limit Earth’s temperature rise to 1.5 °C [3,4].

There are two main processes dominating the steel production industry; integrated steel mills (ISM), which use a blast furnace (BF) and basic oxygen furnace (BOF), and a mini-mill which uses direct reduced iron (DRI) and steel scrap to charge an electric arc furnace (EAF). The BF-BOF route represents the majority of steel production, with nearly 14-fold more steel produced in this manner globally compared to the DRI-EAF route in 2017 [5]. Transitioning to DRI-EAF can reduce the CO₂ emission intensity by 61–68% relative to BF-BOF, but further innovation is required to achieve net zero [6,7].

The two largest market shares for current DRI production are the Midrex and Energiron processes [8]. Process flow diagrams for both of these processes are shown in Figure 1. In both cases, solid iron ore is charged into the top of the shaft furnace. A reducing gas containing primarily CO, H₂, and CH₄ is injected from the side and reacts with the ore to reduce it to solid metallic iron. The depleted reducing gas exits the top of the shaft furnace (referred to herein as “top gas”). A portion of the top gas is recycled back for re-use as reducing gas, while the remainder is combusted as fuel to provide heat for reforming and sensible heating of the reducing gas. There are a few key differences between the Midrex and Energiron processes: (1) the shaft furnace operates at 100–200 kPa for Midrex, while Energiron operates at 600–800 kPa; (2) Midrex operates a dry reformer, while Energiron uses steam methane reforming and/or gases from other parts of the steel mill; (3) Energiron includes a CO₂ absorber to purge CO₂ from the recycled top gas, though this stream is typically vented, and (4) Energiron includes a fired gas heater that creates a second stream of CO₂ emissions [9,10,11,12].

Chemical looping combustion (CLC) is a means of heat production from fossil fuels with inherent separation and capture of CO₂. Through the use of a solid oxygen carrier, the combustion of the fuel is isolated from air without requiring a costly air separation unit [13,14,15]. The typical set-up requires two fluidized bed reactors: (1) the air reactor, where air oxidizes the oxygen carrier and (2) the fuel reactor, where fuel reacts with the oxygen in the oxygen carrier, producing a pure stream of CO₂ and water vapor [16]. Pressurization of the process (pressurized chemical looping, or PCL), provides several additional advantages, including reduced equipment size and capital costs, increased reaction rates, and the potential for latent heat recovery with process water recovery [17,18]. This work proposes novel configurations for net-zero DRI production through the integration of PCL wherever fossil fuels are being consumed to produce heat. The performance and opportunities for PCL-DRI are compared and contrasted with the most common DRI facilities currently in operation.

1.2. Design Specification for DRI

Given the emergence of DRI as a promising option for the iron and steel industry to substantially reduce CO₂ emissions, the scale of future DRI facilities must be comparable to that of existing and future iron and steel facilities. Thus, this work considers a DRI facility with a capacity of producing 2 million tonnes per annum (MTPA) of DRI to be comparable to the Voestalpine Midrex plant constructed in Texas, USA in 2017 [19]. To arrive at hourly flow rates, an annual availability of 91.3% was applied. This is based on the performance guarantee for Midrex facilities currently on the market [11]. For ease of comparison, the availability of the PCL-DRI facilities was considered to be the same. Given that Midrex and Energiron are the two DRI technologies with the largest market share, both of these configurations were used as a basis for comparison to potential PCL-DRI configurations. The key units of a DRI facility are described in the following subsections.

1.2.1. Shaft Furnace

The shaft furnace is the heart of the DRI process. Iron ore (usually in pellet form) is fed to the top of the furnace, while reducing gas is fed from the side and flows counter-currently to the ore. The reducing gas enters the furnace at 850–900 °C for Midrex or 920–1050 °C for Energiron III [12]. The upper region of the furnace is the reducing zone, where iron oxide is reduced upon interaction with the gas, as shown in Equations (1)–(6) below. Methane reforming and water-gas shift reactions (Equations (7) and (8)) also occur in this region [12].

$${3\mathrm{Fe}}_2{\mathrm{O}}_3{+\mathrm{H}}_{2 }\to { 2\mathrm{Fe}}_3{\mathrm{O}}_4{+\mathrm{H}}_2{\mathrm{O}\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=0.7 \mathrm{kJ}/{\mathrm{mol} \mathrm{H}}_2$$

(1)

$${\mathrm{Fe}}_3{\mathrm{O}}_4{+\mathrm{H}}_{2 }\to { 3\mathrm{FeO}+\mathrm{H}}_2{\mathrm{O}\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=7.8 \mathrm{kJ}/{\mathrm{mol} \mathrm{H}}_2$$

(2)

$${\mathrm{FeO}+\mathrm{H}}_2 \to { \mathrm{Fe}+\mathrm{H}}_2{\mathrm{O}\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=24.0 \mathrm{kJ}/{\mathrm{mol} \mathrm{H}}_2$$

(3)

$${3\mathrm{Fe}}_2{\mathrm{O}}_3+\mathrm{CO} \to { 2\mathrm{Fe}}_3{\mathrm{O}}_4{+\mathrm{CO}}_2{\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-40.4 \mathrm{kJ}/\mathrm{mol} \mathrm{CO}$$

(4)

$${\mathrm{Fe}}_3{\mathrm{O}}_4+\mathrm{CO} \to { 3\mathrm{FeO}+\mathrm{CO}}_2{\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-36.9 \mathrm{kJ}/\mathrm{mol} \mathrm{CO}$$

(5)

$$\mathrm{FeO}+\mathrm{CO} \to { \mathrm{Fe}+\mathrm{CO}}_2{\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-17.2 \mathrm{kJ}/\mathrm{mol} \mathrm{CO}$$

(6)

$${\mathrm{CH}}_4{+\mathrm{H}}_2\mathrm{O} \rightleftharpoons { \mathrm{CO}+3\mathrm{H}}_{2\hspace{1em}}{\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-206.2 \mathrm{kJ}/{\mathrm{mol} \mathrm{CH}}_4$$

(7)

$${\mathrm{CO}+\mathrm{H}}_2\mathrm{O} \rightleftharpoons { \mathrm{CO}}_2+{ \mathrm{H}}_2{\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-41.2 \mathrm{kJ}/\mathrm{mol} \mathrm{CO}$$

(8)

The depleted reducing gas exits the top of the furnace as top gas. In the lower region, the iron cools. Depending on the type of DRI product from the facility (hot DRI, cold DRI, or hot briquette iron), there may be additional cooling gas supplied to this region. This is also where carbon deposition onto the iron occurs via methane decomposition and the reverse Boudouard reaction [12,20]. The carbon contained within the DRI product improves the efficiency of the downstream EAF [6]. DRI products from the shaft furnace typically meet the following specifications: 90–94% iron, 92–97% metallization, and 0.5–4% carbon [10,11].

1.2.2. Reformer

The reformer supplies fresh syngas makeup, which is sent to the shaft furnace in combination with recycled top gas for the reduction of iron ore. In the Midrex process (Figure 2a), dry reforming reactions dominate, and the reformer operates at 200–300 kPa(g) [9]. All steam required for reforming is contained within the top gas recycled to the reformer feed, while additional natural gas is supplied for conversion. In the Energiron III process (Figure 2b), the reformer functions as a more traditional steam-methane reformer (SMR), and the recycled top gas is combined with the produced syngas further downstream. In this work, syngas exits the reformer at 925 °C and 220 kPa(a) (Midrex-type processes) or 880 °C and 945 kPa(a) (Energiron-type processes).

1.2.3. Top Gas Scrubber

The top gas scrubber is a wet scrubber that cleans and removes moisture from the top gas as it exits the shaft furnace [21]. In this work, the scrubber was operated at an outlet temperature of 50–51°C to achieve adequate water knock-out from the top gas recycled as reducing gas.

1.2.4. Top Gas Compressor

The top gas compressor re-pressurizes the fraction of the top gas that is recycled for incorporation into the reducing gas. In this work, the discharge pressure from the top gas compressor is 300 kPa(a) (Midrex-type processes) or 1000 kPa(a) (Energiron-type processes). In the base DRI configurations, the portion of the top gas used as fuel in the reformer firebox do not pass through this compressor, as it is already at sufficient pressure for passage through the firebox and downstream heat exchangers.

1.2.5. CO₂ Removal Unit (Energiron III-Type Configurations)

The CO₂ removal unit in the Energiron process is modeled as an amine absorber using MEA as the solvent. It removes CO₂ from the recycled top gas stream to prevent build-up of CO₂ in the system and ensure that the bustle gas re-entering the shaft furnace has sufficient partial pressure of reducing species. The amine solvent is regenerated in a regeneration tower that consumes medium-pressure steam. The majority of the heat required for regeneration is collected from the top gas in the heat recuperator. The operating conditions and performance of the CO₂ removal unit is identical in both the Energiron III base case and PCL-DRI E case. In this work, the amine system was modeled using a solvent concentration of 35 wt % MEA and a solvent loading of 0.25 mol CO₂/mol MEA. Top gas enters the absorber at 1000 kPa(a) and 109 °C. The regenerator operates at 180 kPa(a), with separated CO₂ exiting at 45 °C.

1.2.6. Syngas Treatment

Syngas treatment varies depending on whether it is a Midrex-type process or an Energiron-type process. For Midrex-type processes, the syngas exiting the reformer requires further heating to reach the desired injection temperature to the shaft furnace. This is achieved by direct addition of oxygen to the syngas in a duct burner. Natural gas is also added as a final adjustment to the reducing gas composition as it enters the shaft furnace.

For Energiron-type processes, the syngas exiting the reformer is cooled and then shifted (via a high-temperature water-gas shift reactor) to enrich its hydrogen concentration. It is then further cooled to remove excess water, before being combined with the recycled top gas exiting the CO₂ absorber. The syngas must then be reheated to enter the shaft furnace. This is done first in a fired heater, and then finished in a duct burner with the addition of oxygen and natural gas.

1.3. PCL Configuration for DRI

Two novel configurations were explored in this work to provide a concept-level assessment of how PCL can be applied to DRI processes to capture CO₂ emissions. PCL reactors have previously been investigated for other applications in which industrial heat is needed to produce electricity, steam, hydrogen, and syngas [17]. In the case of DRI, the PCL reactors serve as a replacement for the reformer fireboxes and process gas heaters in the typical Midrex and Energiron process configurations, thereby capturing the flue gas emissions that would result from those fossil fuel-fired units. The high-pressure flue gas exiting the fuel reactor allows for latent heat recovery, while the O₂-depleted air stream from the air reactor (which shall henceforth be referred to as “vitiated air”) can be directed to a turbine to offset the power requirement for the main air compressor. Midrex-type and Energiron III-type PCL process configurations are shown in Figure 2a,b, respectively. New or modified subunits for PCL-DRI are described in the sections below.

1.3.1. Air and Fuel Reactors

The air and fuel reactors comprise the core of the PCL process, whereby the heat supplied to the reformer and to heat process gas streams is produced. Ilmenite ((Fe, Ti)₂O₃) was used as the oxygen carrier due to its low toxicity, availability, and low cost [22,23], though numerous other oxygen carriers can be considered [24]. The fuel reactor is fired with top gas from the DRI shaft furnace, containing CH₄, CO, and H₂ as combustible components. Natural gas is added as supplemental fuel, as needed. Equations (9)–(14) show the reactions that occur in the fuel reactor between ilmenite and the fuel, while Equations (15) and (16) show the reactions between ilmenite and air in the air reactor [17].

$${3\mathrm{Fe}}_2{\mathrm{O}}_{3 }+\mathrm{CO} \to { 2\mathrm{Fe}}_3{\mathrm{O}}_4{+\mathrm{CO}}_2{\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-40.4 \mathrm{kJ}/\mathrm{mol} \mathrm{CO}$$

(9)

$${3\mathrm{Fe}}_2{\mathrm{O}}_3{+\mathrm{H}}_{2 }\to { 2\mathrm{Fe}}_3{\mathrm{O}}_4{+\mathrm{H}}_2{\mathrm{O}\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=0.7 \mathrm{kJ}/{\mathrm{mol} \mathrm{H}}_2$$

(10)

$${\mathrm{Fe}}_2{\mathrm{TiO}}_5{+\mathrm{CO}+\mathrm{TiO}}_{2 }\to { 2\mathrm{FeTiO}}_{3 }{+\mathrm{CO}}_2{\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-65.9 \mathrm{kJ}/\mathrm{mol} \mathrm{CO}$$

(11)

$${\mathrm{Fe}}_2{\mathrm{TiO}}_5{+\mathrm{H}}_2{+\mathrm{TiO}}_2 \to { 2\mathrm{FeTiO}}_3{+\mathrm{H}}_2{\mathrm{O}\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-24.6 \mathrm{kJ}/{\mathrm{mol} \mathrm{H}}_2$$

(12)

$${12\mathrm{Fe}}_2{\mathrm{O}}_{3 }{+\mathrm{CH}}_4\to { 8\mathrm{Fe}}_3{\mathrm{O}}_4{+\mathrm{CO}}_2+{2\mathrm{H}}_2{\mathrm{O}\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=170.0 \mathrm{kJ}/{\mathrm{mol} \mathrm{CH}}_4$$

(13)

$$4{\mathrm{Fe}}_2{\mathrm{TiO}}_5{+\mathrm{CH}}_4{+4\mathrm{TiO}}_{2 }\to { 8\mathrm{FeTiO}}_{3 }{+\mathrm{CO}}_2{+2\mathrm{H}}_2{\mathrm{O}\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=66.0 \mathrm{kJ}/{\mathrm{mol} \mathrm{CH}}_4$$

(14)

$${ 4\mathrm{FeTiO}}_3{+\mathrm{O}}_2\to { 2\mathrm{Fe}}_2{\mathrm{TiO}}_5{+2\mathrm{TiO}}_2{\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-434.0 \mathrm{kJ}/{\mathrm{mol} \mathrm{O}}_2$$

(15)

$${4\mathrm{Fe}}_3{\mathrm{O}}_4{+\mathrm{O}}_2\to { 6\mathrm{Fe}}_2{\mathrm{O}}_3{\hspace{1em}\mathsf{\Delta }\mathrm{H}}_{\mathrm{r}}^{°}=-486.0 \mathrm{kJ}/{\mathrm{mol} \mathrm{O}}_2$$

(16)

Due to the nature of the reactants, the reactions in both the air and fuel reactor are net exothermic, though the majority of the heat is released in the air reactor and is removed both in-bed and in the freeboard. The air and fuel reactor bed temperatures were fixed at a maximum of 950 °C to avoid the use of exotic metals for internal components. The preheat temperature of the fuel reactor was controlled such that the small exothermic nature of the reactions was consumed in sensible heating of the reactants, thus requiring no net heat removal from the fuel reactor. The operating pressure of both reactors was set to 800 kPa(a) based upon the optimum determined in previous work [25].

1.3.2. Reformer

For both PCL-DRI cases, the air reactor of the PCL unit provides the heat of reforming to the catalyst-filled tubes, acting as a replacement to a traditional reformer firebox. This offers many advantages, including reduced thermal stresses on the reformer tubes due to a lower temperature difference between the reformed gas and air reactor, lower mechanical stress due to a smaller pressure differential between the inside and outside of the reformer tubes (for the Energiron-type configuration), and lower external tube temperature (due to improved heat transfer). These factors allow for the selection of less costly tube materials.

1.3.3. Air Compressors and Power Recovery

The compressor supplying the air reactor has two stages. The first stage is directly coupled to an expansion turbine, and is driven by the expansion of the vitiated air stream exiting the air reactor. The second stage is electrically driven. The air discharge pressure from the second stage is 971 kPa(a). Previous work has investigated options for power recovery in PCL, including simple and high-temperature turbo expanders, incorporation of a duct burner for additional power recovery, or omission of power recovery [17].

1.3.4. CO₂ Processing

The CO₂ processing requirements will be site- and application-specific. Depending on the location of the facility, the captured CO₂ may be compressed and transported as a gas by pipeline, or may be converted to a liquid for transportation by rail, ship, or truck [26,27]. Here, we consider supercritical CO₂ transported by pipeline. Prior to compression, heat is recovered from the flue gas using various heat recovery heat exchangers. The gas is then further cooled in a direct contact cooler, to condition it for entry to the CO₂ compression train. CO₂ compression occurs in five stages with intercooling. The first stage has a compression ratio of 4, and is followed by a dryer to remove moisture down to the pipeline specification. The remaining stages are equally sized with compression ratios of 1.44, to arrive at a final product pressure of 12,060 kPa(a), which is within the specified range for the Alberta Carbon Trunk Line [28].

The required purity of the CO₂ is dependent upon the transportation pipeline specifications and on the final destination of the CO₂ (storage or utilization). For this comparative concept-level study, no additional purification is assumed to be required. Options for removing combustibles and inerts have been reviewed in literature and include pure oxygen injection, cryogenic fractionation, use of solvents, and/or membranes [29,30].

1.3.5. Top Gas Compressor

For PCL-DRI, the top gas compressor is larger than the base case, since all top gas is directed through it, rather than just the fraction that is recycled as reducing gas. The discharge pressure is maintained the same as the base case.

1.3.6. Fuel Gas Compressor (Midrex-Type Configuration)

The Midrex reformer operates at a lower pressure than the required fuel gas supply pressure to the PCL units. Thus, for the PCL-DRI M process, the fuel portion of the top gas is sequentially pressurized to 971 kPa(a) in a second compressor following the top gas compressor (the fuel gas compressor) to meet the supply pressure requirements of the fuel reactor.

1.3.7. Syngas Treatment

All aspects of the syngas treatment are the same except for one substitution: the heating that would have been done in a fired heater in the Energiron base case is now done inside the air reactor of the PCL unit. The heating duty and outlet temperature is kept constant between the two cases.

2. Materials and Methods

2.1. Process Simulation Methodology

All four configurations were simulated in Aspen HYSYS V11 (HYSYS). Since this was a preliminary study, a detailed shaft furnace model was not developed. Instead, the top gas exiting the top of the shaft furnace was treated as the inlet boundary of the models, and the reducing gas exiting the duct burner was treated as the outlet boundary. The conditions and compositions of these two streams were informed by literature. For the Midrex base case (Base—M), data for the Contrecoeur plant presented by Hamadeh was used [12]. For the Energiron III base case (Base—E), the data was taken from Zugliano et al. [31]. The former produces a cold DRI product, while the latter produces hot DRI. PCL-DRI M was built up using the Midrex shaft furnace data, while PCL-DRI E used the Energiron III data.

Before using either set of data, the mass and energy balances around each shaft furnace were verified in independent models in HYSYS. These models consisted of simplified conversion reactors and heat exchangers, applying the reactions identified in Section 1.2.1. As is discussed further in Section 3.1.1, the mass balance on the Midrex data did not close and required minor adjustments to the top gas stream. This corrected top gas, scaled to the desired DRI production rate, was used as the input to the models for Base—M and PCL-DRI—M. There were no errors identified in the Energiron data, and so the literature top gas composition and flow, after scaling up to the target operating flow, was used directly as the input to Base—E and PCL-DRI—E. The operating conditions of the remaining process units (such as the top gas scrubber and the reformer) and required flows of raw materials were adjusted until the composition and mass flow of the reducing gas exiting the duct burner of the models were in agreement with the literature values for the reducing gas entering the shaft furnace.

2.2. Case Descriptions

There are a total of five process configurations that are considered in this work, as summarized in

Table 1. Summary of cases simulated in HYSYS.

Case	Base—M	Base—E	Base—E + Comp	PCL-DRI—M	PCL-DRI—E
DRI process type	Midrex	Energiron	Energiron	Midrex	Energiron
PCL employed	No	No	No	Yes	Yes
CO2 compression train	No	No	Yes	Yes	Yes
CO2 fate	Vented	Vented	Partial capture	Full capture	Full capture

. There are two base cases—one for a Midrex process (Base—M) and one for an Energiron III process (Base—E), as shown in Figure 1. These cases assume that all CO₂ is vented to the atmosphere. There is limited public operating data for DRI facilities; other more optimal operating points may exist. For the purpose of comparison to PCL-DRI, the available data are sufficient, as the performance of the shaft furnace is kept constant for all Midrex-type cases and for all Energiron-type cases.

There are two PCL-DRI configurations, as shown in Figure 1. PCL-DRI M is the Midrex-type case, while PCL-DRI E is the Energiron-type case. Both of these cases employ a CO₂ compression train to condition the captured CO₂ for pipeline transportation.

Energiron DRI facilities include a CO₂ absorber in their standard flowsheet. Case “Base—E + Comp” explores a partial CO₂ capture scenario in which the CO₂ stream from the absorber is directed to a CO₂ compression train for transportation to a sequestration or utilization site.

2.3. Fluid Package and Thermodynamic Data

The Peng–Robinson fluid package was used for all streams except for the modeling of the CO₂ absorber in Energiron-type configurations. For those streams, the Acid Gas—Chemical Solvents package was used. Peng–Robinson is an equation of state model and is the best optimized model in HYSYS for gases at high temperature. For the reactions with the solid iron ore and ilmenite ore species, hypothetical compounds were created in HYSYS using thermodynamic data from FactSage [17,32].

2.4. Battery Limits

Inputs to the process simulations are ilmenite ore (the oxygen carrier in the PCL reactors), oxygen, natural gas, air, and top gas from the shaft furnace. Properties of these streams are provided in

Table 2. Properties of reactants in base case and PCL-DRI process simulations.

	Natural Gas	Ilmenite a	Oxygen	Air	Top Gas—Midrex	Top Gas—Energiron
Temperature (°C)	25	25	25	25	285	412
Pressure (kPa(a))	400	101	300|750 b	101	142	661
Species (mol frac)
CH4	0.9594	0	0	0	0.0300	0.0586
H2O	0	0	0	0	0.2000	0.2847
CO	0	0	0	0	0.1900	0.1391
CO2	0.0073	0	0	0	0.1600	0.0728
O2	0	0	0.9000	0.2100	0	0
H2	0	0	0	0	0.4000	0.4332
N2	0.0169	0	0.1000	0.7900	0	0.0117
C2H6	0.0160	0	0	0	0	0
C3H8	0.0004	0	0	0	0	0
TiO2	0	0.4124	0	0	0	0
Fe2TiO5	0	0.2918	0	0	0	0
Fe2O3	0	0.2958	0	0	0	0

^a PCL-DRI cases only. ^b A|B are for A = Midrex-type process and B = Energiron-type process.

. Oxygen is assumed to be provided as a utility at the pressure required for the process. Iron ore is an input to the DRI shaft furnace; however, since the furnace is not modeled directly, the ore is not considered here.

Outputs at the simulation boundaries common to all cases are the reducing gas and produced condensate. From the base case simulations, additional product streams are the flue gas from the reformer/fired gas heater and purge CO₂ from the absorber (Base—E only). For the PCL-DRI cases, these streams are replaced by the compressed CO₂ product and the vitiated air stream from the air reactor. Spent ilmenite is the final product from the PCL-DRI process. The properties of these streams are summarized in

Table 3. Properties of products in base case and PCL-DRI process simulations.

	Reducing Gas	Condensate	Flue Gas a	CO2 from Absorber b	Ilmenite c	Vitiated Air c	Compressed CO2 Product d
Temperature (°C)	957–962	50–67	106–351	45	191–425	17–170	49
Pressure (kPa(a))	200–725	122–621	101–121	180	800	101	12 060
Species (mol frac)
CH4	0.083–0.091	0	0	0.001	0	0	0.015
H2O	0.030–0.043	0.999–1.000	0.162–0.216	0.054	0	0	0
CO	0.169–0.320	0	0	0	0	0	0
CO2	0.022–0.050	0–0.001	0.091–0.111	0.943	0	0	0.935
O2	0	0	0.009–0.070	0	0	0.013	0
H2	0.509–0.662	0	0	0.002	0	0	0
N2	0.013–0.016	0	0.658–0.718	0	0	0.987	0.050
C2H6	0	0	0	0	0	0	0
C3H8	0	0	0	0	0	0	0
TiO2	0	0	0	0	0.412	0	0
Fe2TiO5	0	0	0	0	0.292	0	0
Fe2O3	0	0	0	0	0.296	0	0

^a Base—M, Base—E, and Base—E + comp cases only. ^b Base—E + comp case only. ^c PCL-DRI cases only. ^d PCL-DRI and Base—E + comp cases only.

, while case-specific stream information can be found in the Supplementary Materials.

2.5. Calculation of Thermal Efficiency

The thermal efficiency of the process is calculated as the percent of useful heat input to the process compared to the total heat input to the process, as shown in Equation (17).

$$\begin{gathered} \mathrm{Thermal} \mathrm{efficiency} (\%)=\frac{\mathrm{useful} \mathrm{heat}}{\mathrm{total} \mathrm{heat} \mathrm{input}}\ast 100 \% \\ =\frac{\left(\mathrm{total} \mathrm{heat} \mathrm{input}\right)-\left(\mathrm{waste} \mathrm{heat}\right)}{\mathrm{total} \mathrm{heat} \mathrm{input}}\ast 100 \% \end{gathered}$$

(17)

The total heat input is calculated as the summation of the electric power for all rotating equipment, the higher heating value multiplied by flow rate of all reactant streams, any hot utility stream duties, and the sensible heat of all reactants above a reference condition of 25 °C and 101 kPa(a). Waste heat was calculated as the summation of all cooling utility duties, and the sensible heat of waste streams (e.g., condensate, flue gas that is vented, spent oxygen carrier) above a reference condition of 25 °C and 101 kPa(a).

3. Results

3.1. Base Case Agreement to Literature Values

3.1.1. Base—M

A mass balance on the literature data provided in [12] for the shaft furnace in the Contrecoeur facility showed that the reported flow of top gas exiting the shaft furnace was incorrect. To achieve a correct mass balance, the mass flow of top gas was increased by 7% relative to the literature data and the composition was slightly adjusted. This discrepancy is believed to be a result of errors in measurement of the high-temperature process gas stream. Note that this additional mass flow all ends up in the excess purge stream (stream 14) and thus has no impact on the idealized process performance. A comparison between the reducing gas composition generated by the model and the target literature value shows good agreement (

Table 4. Comparison of Base—M model boundaries at shaft furnace to literature.

	Top Gas from Shaft Furnace		Reducing Gas to Shaft Furnace
	Literature	Model	Literature	Model
Temperature (°C)	285	285	957	957
Pressure (kPa(a))	142	142	n.d.	200
Mass flow (t/h)	299.8 1	319.8	229.5 1	229.5
Species (mol frac)
CH4	0.0295	0.3000	0.0908	0.0909
H2O	0.1903	0.2000	0.0428	0.0426
CO	0.1958	0.1900	0.3271	0.3197
CO2	0.1709	0.1600	0.0240	0.0216
H2	0.4028	0.4000	0.4966	0.5094
N2	0.0102	0.0200	0.0176	0.0157

¹ Scaled to 2 MTPA DRI production rate.

).

While no operating data for the balance of the plant is provided in [12], the operating conditions of the reformer in Base—M can be compared to typical ranges that would be expected for a Midrex reformer, where dry reforming reactions dominate.

Table 5. Comparison of Base—M reformer operating conditions to expected ranges.

Operating Condition	Base—M	Typical Midrex Reformer 1
Reformer feed pre-heat temperature (°C)	500	400–500
Reformed gas outlet temperature (°C)	925	925
Reformed gas outlet pressure (kPa(a))	220	200–300
CO2/Carbon ratio	0.80	0.80
Steam/Carbon ratio	0.48	0.65

¹ Typical values as reported in [9].

shows that all model parameters fall within the expected range except for the steam to carbon ratio. In the model, this ratio is lower than the expected value, however, it falls within the same magnitude and results in dry reforming. It is not possible to increase the steam to carbon ratio in the model without exceeding the specified steam content for the reducing gas in Table 4. Overall, the model provides a good representation of the Midrex reformer, and given that the same conditions are used in PCL-DRI—M, will not impact differential performance observations between Base—M and PCL-DRI—M.

3.1.2. Base—E

Verification of the mass balance around the shaft furnace did not show any major discrepancies for the data provided in Zulgiano et al. [31], thus no adjustments to the model’s input top gas composition were required. The reducing gas composition calculated by the model for the Base—E case shows good agreement with the literature values, as shown in

Table 6. Comparison of Base—E model boundaries at shaft furnace to literature.

	Top Gas from Shaft Furnace		Reducing Gas to Shaft Furnace
	Literature	Model	Literature	Model
Temperature (°C)	412	412	962	962
Pressure (kPa(a))	661	661	725	725
Mass flow (t/h)	293.71	293.7	193.0 1	192.7
Species (mol frac)
CH4	0.0586	0.0586	0.0830	0.0831
H2O	0.2881	0.2881	0.0296	0.0296
CO	0.136	0.136	0.1620	0.1691
CO2	0.0728	0.0728	0.0498	0.0502
H2	0.4332	0.4332	0.6629	0.6623
N2	0.0113	0.0113	0.0127	0.0129

¹ Scaled to 2 MTPA DRI production rate.

.

Table 7. Comparison of Base—E reformer operating conditions to typical steam methane reforming.

Operating Condition	Base—E	Typical SMR
Reformed gas outlet temperature (°C)	880	800–900 1
Reformed gas outlet pressure (kPa(a))	945	1500–3000 1
Steam/Carbon ratio	1.6	1.5–5 2

¹ As described in [35]. ² As described in [34,35,36].

shows the reformer operating conditions for Base—E in comparison to a typical SMR. The reformed gas outlet temperature and the steam to carbon ratio in the feed are both within expected ranges. The steam to carbon ratio was set near the minimum of the range because, unlike for H₂ production, DRI requires a larger yield of CO to act as a reductant in the shaft furnace. SMRs are typically operated at high pressure to limit downstream compression requirements at the hydrogen production facility, however, the reactions are less thermodynamically favorable as the pressure increases [33,34]. Given the Energiron shaft furnace operates at 600–800 kPa [12], there is no need to increase the reformer pressure beyond what is required as allowance for pressure drop across downstream processing units between the reformer and the shaft furnace.

3.2. Comparison of Process Performance

The PCL-DRI cases are equivalent or superior to the base cases in many aspects of DRI process performance as shown in

Table 8. Comparison of key performance indicators between base cases and PCL-DRI cases.

Performance Indicator	Unit	Base—M	PCL-DRI—M	Base—E	Base—E + Comp	PCL-DRI—E
Total natural gas consumption	GJ/t DRI	11.9	11.9	10.7	10.7	9.80
Fuel consumption for heating (reforming + fired heaters/fuel reactor)	GJ/t DRI	4.64	4.26	4.54	4.54	3.65
Oxygen consumption	Nm3/t DRI	27.7	27.7	16.4	16.4	16.4
Water produced	t H2O/t DRI	0.135	0.347	0.231	0.236	0.397
Nitrogen production	Nm3/t DRI	0	714	0	0	692
Utility cooling duty	MWh/t DRI	0.243	0.258|0.542 1	0.686	0.721	0.660|0.723 1
Net power import	MWhe/t DRI	0.055	0.114|0.148 1	0.056	0.082	0.085|0.130
Thermal efficiency	%	91.7	91.9	92.0	91.6	90.9
CO2 produced	t CO2/t DRI	0.393	0.354	0.450	0.450	0.401
% CO2 captured	%	0	100	0	47	100
% CO2 emitted	%	100	0	100	53	0

¹ Values A|B are: A = excluding contribution of DCC + CO₂ compression; B = including contribution of DCC + CO₂ compression.

. First, comparing Base—M to PCL-DRI—M, the overall natural gas consumption of the two processes, which serves only to supplement the feed gas to the reformer tubes and provide direct contact heating of the reducing gas in the duct burner, is identical. This is because the chemistry and the conditions of the recycle gas loop are maintained constant between both cases. Likewise, the oxygen consumption is identical. The required fuel consumption (in the form of top gas) to the fuel reactor of the PCL unit compared to the reformer firebox is less in the PCL-DRI—M case due to the lower operating temperature of the PCL unit, thus requiring less sensible heating of the gases. Since heat losses in the recycled gas loop are not considered in these idealized cases, there is an excess top gas stream that is required neither as a fuel for heating, nor as a recycle feed to the reformer. For the purpose of these simulations, this excess gas is directed to a purge stream and it is excluded from emissions considerations or its impact on the thermal efficiency of the plant in this work. When factoring in practical operating considerations, this stream would not exist, because all of the top gas would be required to be combusted as fuel in the Midrex firebox to account for heat losses, with further supplementation of natural gas. To avoid introduction of additional uncertainty through the estimate of heat losses for the PCL-DRI and Base—E configurations, no heat losses in the recycle gas and reforming loop are considered for any of the cases. Thus, fuel consumption and CO₂ production rates represent the minimum possible for these configurations.

Comparing the fuel consumption for Base—E and PCL-DRI—E, again, there is a reduction in the fuel consumption for the PCL fuel reactor due to its lower operating temperature. In these configurations, all of the available top gas plus additional natural gas is required to fulfill the heating requirement of the PCL unit/firebox; thus, this translates directly to an 8% reduction in the battery limit natural gas requirement for the process.

The PCL-DRI cases have potential value-added by-products. Process water is produced at a rate of 2.6 times greater for the Midrex-type cases and 1.7 times greater for the Energiron-type cases. This water is recovered from the fuel reactor flue gas during heat recovery/cooling, compression, and drying. This additional water stream can be evaluated for use as boiler feed water, for cooling and scale removal in the hot rolling mill, or for other applications at the iron and steel facility [37]. Additionally, there is production of a nitrogen stream from the air reactor, which may also be considered for cooling in the hot rolling mill to offset some of the water requirement.

CO₂ capture does have additional costs, which become evident when examining the increased power import and cooling duty of the PCL-DRI cases relative to the base cases. The net power imports of the PCL-DRI processes are more than double the power requirements for their respective base cases, despite the offset of power requirement achieved through the use of an expansion turbine on the nitrogen stream produced from the air reactor. Power consumption of PCL-DRI—M is greater than PCL-DRI—E due to the greater compression requirement for the top gas used as fuel in the fuel reactor. The Midrex top gas is near atmospheric pressure when it exits the shaft furnace, whereas the Energiron top gas is already near the operating pressure of the PCL units.

The two major draws for power in the PCL-DRI process are (1) the air compressors supplying air to the air reactor and (2) the CO₂ compression train that brings the captured CO₂ to pipeline pressure. To understand the impact of these two main power consumers, the power consumption for PCL-DRI is presented both including and excluding CO₂ conditioning and compression. A partial capture case (Base—E + comp) also shows the impact of adding a smaller CO₂ compression unit to the CO₂ stream that is separated from the CO₂ absorber that is inherently a part of Energiron-type DRI processes. CO₂ compression represents 23 and 34% of the total power import for PCL-DRI—M and PCL-DRI—E respectively, while it adds 46% for partial CO₂ capture from Base—E. With the push toward electrification and toward reducing the carbon intensity of the power grid, the increased power import of the PCL-DRI processes does not have to be considered as a negative impact; it is merely a shift in the type of energy that the process consumes. A detailed techno-economic assessment (TEA) is required to assess how this shift may affect operating costs of the DRI facility, as well as how the cost of carbon capture compares to competing CO₂ capture technologies. Further optimization can be performed to minimize power import, if desired.

Considering the cooling duty of the PCL-DRI processes relative to their respective base cases, the cooling duty excluding CO₂ conditioning and compression is quite similar. The addition of conditioning in a DCC, drying, and intercooling between the CO₂ compressor stages results in a 123% increase in the total process cooling duty for PCL-DRI—M and a 5% increase for PCL-DRI—E. The relatively small change for PCL-DRI—E results from the much higher cooling duty already required for Energiron-type processes, both in the condenser of the CO₂ absorber and in the cooling of the syngas exiting the reformer for water knock-out, prior to re-heating the reducing gas for entry to the shaft furnace. Even after heat recovery from the flue gas, the stream is still relatively hot (528 °C for PCL-DRI—M and 130 °C for PCL-DRI—E) and contains a significant amount of waste heat that must be cooled before compression can occur.

Chemical looping has often been described as a carbon capture technology with little to no energy penalty [38]. This is observed in the comparison of thermal efficiencies of PCL-DRI—M to Base—M, in which there is no loss in efficiency with integration of PCL. For PCL-DRI—E and Base—E + Comp, there is a small drop in efficiency that can be attributed to the increase in power and cooling duty for the CO₂ compression units, though this effect is much smaller than would be the case if amine capture were applied to all flue gas streams.

4. Discussion

The Base—M and Base—E cases were developed to have a reasonable point of comparison for PCL-DRI to current industrial standards; however, each of these models was built based on a single operating point and may not be representative of all Midrex or Energiron-III processes. If more industrial data were available at alternative operating points, or if a detailed shaft furnace model could be developed and integrated with the current models, future work could focus on broadening the understanding of PCL-DRI process performance under a range of common operating conditions. To evaluate the accuracy of the developed models, they were compared to emissions data in literature. The Midrex process typically produces 0.65 t CO₂/t DRI, while Energiron III produces 0.56 t CO₂/t DRI [7]. In this work, the idealized Base—M configuration emits 0.39 t CO₂/t DRI, however, if one considers the combustion of the excess top gas as fuel in the reformer to make up for heat losses, as would be done in a non-idealized plant, this value rises to 0.52 t CO₂/t DRI. The Base—E configuration likewise undershoots the expected value, at 0.45 t CO₂/t DRI, but again, the difference would likely be made up if heat losses were considered.

The performance of the PCL-DRI processes are competitive with traditional DRI processes, with the added benefit of full CO₂ capture. There was also no optimization of the top gas recycle loop for PCL-DRI; it is possible that superior performance could be achieved if the overall process pressure and reformer feed conditions could be modified to better suit PCL. For example, the shaft furnace in PCL-DRI—E could be operated such that the top gas is at a pressure sufficient to supply the fuel reactor directly, without requiring an additional fuel gas compressor. For the PCL-DRI—M configuration, options could be explored to reduce the natural gas feed to the reformer through the use of the additional top gas that would be available due to the reduced fuel requirement in the fuel reactor. Optimization activities such as these would require a detailed kinetic model of the shaft furnace in order to predict performance with changing reducing gas compositions and pressures.

Despite its many advantages, the challenges and drawbacks associated with PCL-DRI must also be considered. As has been stated in Section 3.2, the duty of the main air compressor(s) significantly increase the power consumption of the process. Due to their size, previous work has shown that the rotating equipment represents the largest portion of the capital costs of the plant, equal to or more than the cost of all other equipment combined [25]. Scale-up may also be a challenge, given the very high duty and solids circulation rates that would be required to meet the needs of a typical iron and steel facility. In this work, heat recovery from the air reactor approaches 200 MW, depending on the case, and the required solids circulation rate is up to 1370 t/h (oxidized). It is anticipated that two or more PCL units would be required to meet this demand, allowing scale-up to take a modular approach. Choice of reactor type will likely be a key driver in the economics and feasibility of such a process scale. Future work will examine reactor sizing and economics for a novel design with no external solids transfer between the air and fuel reactors.

PCL-DRI is not the only way to eliminate CO₂ emissions from a DRI facility. Post combustion capture units, such as amine absorbers, can be employed, or alternative feedstocks such as low-carbon hydrogen or bio-methane can be used [39,40,41,42]. PCL has already been shown to be more efficient than conventional carbon capture technologies for other applications [17,24,43]. Midrex and Energiron, as well as several other green steelmaking projects, have configurations under development that eliminate the reformer from the DRI process and allow direct use of H₂ as the reducing gas in the shaft furnace (purchased over-the-fence or produced on site); however, this alters the heat balance in the shaft furnace and requires additional heat input to make up the difference [6,39]. There is the additional caveat of the alteration of the composition of the DRI product when only hydrogen is used as the reductant. To aid in the efficiency of the downstream EAF, it is preferred to have 1.5–3 wt % carbon in the DRI. This carbon is then combusted with oxygen in the EAF to help melt the iron; thus, those CO₂ emissions would have to be managed downstream [6]. In traditional DRI processes, the carbon in the DRI is incorporated in the shaft furnace via carbon deposition from the thermal decomposition of methane and the inverse Boudouard reaction [8]. Switching to DRI-H₂ means that carbon cannot be deposited onto the DRI, and will increase the energy consumption of the EAF. PCL-DRI does not have this drawback, and will allow the EAF to operate without modification or loss of efficiency. Furthermore, PLC-DRI does not depend on the existence of a local low-carbon hydrogen producer or hydrogen distribution network, allowing net-zero DRI to be produced before the infrastructure for a hydrogen economy becomes available, as well as in remote regions where low-carbon hydrogen might be inaccessible.

This work considers the use of ilmenite ore as the oxygen carrier for PCL. This could be an attractive option for steelmakers, as there would be the potential for iron and titanium recovery from the spent oxygen carrier, thus valorizing an otherwise waste-product at its point of use. Iron ore can also be used directly as an oxygen carrier [23], which would simplify the supply chain for the iron and steel facility, since oxygen carrier makeup could be supplied from the iron ore already brought in to feed the shaft furnace. Many other oxygen carriers bear consideration, both natural and synthetic, and typically contain some combination of Cu, Mn, Fe, Co, and Ni species [24,44,45]. The choice of oxygen carrier will impact the heat balance of the PCL units as well as the flue gas purity, based on the reactivity of the oxygen carrier with the fuel. In the current work with ilmenite ore, the fuel reactor is net adiabatic (incoming sensible heat and exothermic reactions with CO and H₂ counterbalance the endothermic reactions with CH₄); however, with some oxygen carriers, heat is released in both the air and fuel reactors. This would alter the location of some of the heat recovery units in the PCL units, though at this stage of the analysis, it would not have a significant impact on the overall process performance. Ilmenite ore has a relatively low reactivity with CH₄ compared to CO and H₂ [22,46], and for this reason, the flue gas is expected to contain a fraction of unconverted methane (in this work, we assume 98.5% fuel conversion). Other oxygen carriers may have differing reactivities with the fuel, differing selectivities for products (resulting in differing yields of CO₂) or have oxygen uncoupling properties, all of which may impact flue gas treatment requirements to achieve purity specifications for transportation and storage.

5. Conclusions and Future Work

Process simulation of two novel DRI configurations show that PCL-DRI is a promising option for net-zero DRI production. The PCL-DRI process has many added benefits over traditional DRI configurations in addition to inherent CO₂ capture, including reduced fuel consumption, production of potentially valuable by-products, and, little or no reduction in thermal efficiency. Unlike other methods of eliminating emissions from iron production, there is no required change in the chemistry of the reducing gas or DRI product, allowing for operational flexibility and simplicity of integration with downstream processes as steelmakers transition toward net-zero steel production.

Future work will consider a full TEA-LCA for the PCL-DRI process, as well as investigation of alternative oxygen carriers. Additional configurations can be assessed to optimize the operating pressure of the PCL units for DRI and reduce the net power import of the process.