A Real Options Model for CCUS Investment: CO₂ Hydrogenation to Methanol in a Chinese Integrated Refining–Chemical Plant

Abstract

The scaling up of carbon capture, utilization, and storage (CCUS) deployment is constrained by multiple factors, including technological immaturity, high capital expenditures, and extended investment return periods. The existing research on CCUS investment decisions predominantly centers on coal-fired power plants, with the utilization pathways placing a primary emphasis on storage or enhanced oil recovery (EOR). There is limited research available regarding the chemical utilization of carbon dioxide (CO₂). This study develops an options-based analytical model, employing geometric Brownian motion to characterize carbon and oil price uncertainties while incorporating the learning curve effect in carbon capture infrastructure costs. Additionally, revenues from chemical utilization and EOR are integrated into the return model. A case study is conducted on a process producing 100,000 tons of methanol annually via CO₂ hydrogenation. Based on numerical simulations, we determine the optimal investment conditions for the “CO₂-to-methanol + EOR” collaborative scheme. Parameter sensitivity analyses further evaluate how key variables—carbon pricing, oil market dynamics, targeted subsidies, and the cost of renewable electricity—influence investment timing and feasibility. The results reveal that the following: (1) Carbon pricing plays a pivotal role in influencing investment decisions related to CCUS. A stable and sufficiently high carbon price improves the economic feasibility of CCUS projects. When the initial carbon price reaches 125 CNY/t or higher, refining–chemical integrated plants are incentivized to make immediate investments. (2) Increases in oil prices also encourage CCUS investment decisions by refining–chemical integrated plants, but the effect is weaker than that of carbon prices. The model reveals that when oil prices exceed USD 134 per barrel, the investment trigger is activated, leading to earlier project implementation. (3) EOR subsidy and the initial equipment investment subsidy can promote investment and bring forward the expected exercise time of the option. Immediate investment conditions will be triggered when EOR subsidy reaches CNY 75 per barrel or more, or the subsidy coefficient reaches 0.2 or higher. (4) The levelized cost of electricity (LCOE) from photovoltaic sources is identified as a key determinant of hydrogen production economics. A sustained decline in LCOE—from CNY 0.30/kWh to 0.22/kWh, and further to 0.12/kWh or below—significantly advances the optimal investment window. When LCOE reaches CNY 0.12/kWh, the project achieves economic viability, enabling investment potentially as early as 2025. This study provides guidance and reference cases for CCUS investment decisions integrating EOR and chemical utilization in China’s refining–chemical integrated plants.

1. Introduction

In 2023, global carbon emissions reached 37.4 billion tons, with China contributing a significant 12.604 billion tons, or 31.2% of the total. CCUS is a crucial method for significantly reducing carbon dioxide CO₂ emissions. EOR using CO₂ has been an innovative practice since the 1960s [1]. In 2005, the Intergovernmental Panel on Climate Change (IPCC) officially defined carbon capture and storage (CCS) as a vital process where CO₂ is extracted from energy production and industrial activities, then transported to a designated site for permanent storage, keeping it safely out of the atmosphere [2]. With advancements in carbon dioxide utilization technologies, CCS has been integrated into energy processes in China, leading to the emergence of CCUS [3]. Utilization is an essential complement to CCS, enriching our approach to environmental sustainability. As of March 2024, there are impressive 844 CCS/CCUS projects worldwide. The United States is at the forefront with an incredible 293 projects, followed by the United Kingdom (92), Canada (74), and Australia (36). China ranks sixth with 30 projects [4]. CCUS is developing an integrated industrial chain that includes capture, transportation, storage, enhanced oil recovery, and resource utilization. The global CO₂ mitigation capacity of CCUS is projected to reach 1.6 billion tons per year (tCO₂/a) by 2030, approximately 4.0 billion tCO₂/a by 2035, and 7.6 billion tCO₂/a by 2050. By that time, it is expected that around 50% of fossil fuels will be used in combination with CCUS, enabling the capture of about 3.5 billion tCO₂/a from fossil fuel sources [5]. By 2060, China is anticipated to achieve an annual CO₂ reduction of roughly 1.0 billion tCO₂/a through the deployment of CCUS [6].

The refining industry faces substantial pressure to decrease its emissions. Refineries are major sources of CO₂, with processes like crude oil fractionation and cracking producing substantial greenhouse gas emissions [7]. From 2000 to 2021, emissions from China’s refineries surged from 102.2 million tons to 313.3 million tons, highlighting the urgent need for effective climate policies [8]. Industries with high emissions, such as coal power, cement, steel, and refining, are facing increasing pressure to reduce their carbon output. Achieving net-zero production emissions will be nearly impossible without advanced technologies like CCUS. The current methods for utilizing carbon dioxide include geological, chemical, physical, and biological approaches. EOR and Enhanced Gas Recovery (EGR) technologies significantly increase hydrocarbon recovery rates and provide substantial economic benefits while helping to reduce carbon emissions [9]. As an important C1 feedstock, CO₂ can be widely used in urea synthesis, methanol synthesis, and other processes [10]. Converting CO₂ and hydrogen into methanol is especially promising within the CCUS sector [11,12]. CO₂-to-methanol conversion offers synergistic benefits: effective utilization of surplus renewable electricity, efficient transformation of electrical energy into storable chemical energy, concurrent carbon emission reduction, and enhanced liquid fuel security.

CCUS exhibits inherent compatibility with the petrochemical industry. The potential for utilization in the petrochemical industry is much greater than in the coal-fired power sector. In the coal-fired power industry, CCUS options are relatively limited, primarily relying on geological storage and EOR. The economic feasibility of these options is largely influenced by policy support and the revenue generated from EOR. EOR can significantly boost profits when oil prices are high. However, during periods of low oil prices or when low oil prices accompany high carbon prices, the potential returns from EOR projects decrease significantly. This decline in returns can discourage investment in such initiatives. In contrast, chemical utilization offers more stable economic growth opportunities. CCUS in the petrochemical industry offers a broader range of pathways, including methanol synthesis, urea production, fuel generation, cyclic carbonate and polycarbonate synthesis, and salicylic acid production. In addition to relying on policy incentives and EOR revenues, these industries can also earn extra income through product sales. This facilitates the development of market-driven carbon utilization chains and enhances their economic self-sufficiency. Despite its great potential, CCUS technologies face significant challenges, including low technology maturity, high costs, and long payback periods, which result in cautious investment. CCUS investment in the petrochemical industry is influenced by various uncertainties, so it is necessary to conduct a scientific analysis of CCUS investment decisions and put forward policies on this basis. The incentive strategy is critical for advancing the commercial deployment of CCUS. As shown in

Table 1. Comparison of costs across different methanol production technologies [13].

Technical Route	Current Cost (CNY/ton)	Future Cost (CNY/ton)	Cost-Driving Factors	Development Potential
Coal-based methanol	1800–2700	2700–3600	Coal price, carbon price	Low, constrained by carbon prices
Gas-based methanol	1600–3000	2800–3500	Gas price, carbon price	Low, constrained by limited domestic resources and volatile global gas prices
Biomass-based methanol	3800–5000	1900–2500	Biomass pellet prices	Moderate, dependent on biomass pellet price reduction
Electricity-based methanol	4400–4600	2100–2200	Green electricity price, carbon capture cost	High, may become mainstream

, among the various methanol production pathways, coal-based methanol currently exhibits the lowest production cost, typically ranging from CNY 1800 to 3300 per ton. However, this route is highly sensitive to coal price fluctuations and is associated with significant carbon emissions. With the increasing stringency of carbon pricing mechanisms, the cost advantage of coal-based methanol is expected to diminish. In the future, as coal prices continue to rise, its production cost is projected to exceed CNY 3000 per ton. Methanol derived from natural gas demonstrates a degree of economic competitiveness, yet it remains constrained by limited resource availability and pronounced price volatility. Electricity-based methanol, produced from CO₂ and green hydrogen, is currently more costly, with estimated production costs ranging from CNY 4000 to 4600 per ton. Nonetheless, ongoing reductions in the cost of renewable electricity and carbon capture technologies, coupled with continuous process optimization, are expected to drive significant cost declines, potentially lowering production costs to CNY 2100–2200 per ton in the long term. This pathway is well aligned with carbon neutrality objectives and is projected to become a mainstream production option in the medium to long term. Biomass-based methanol currently incurs production costs of approximately CNY 3800 per ton. Despite this, it offers considerable potential for cost optimization and serves as a valuable supplementary option in regions with abundant biomass resources. In the long term, its production cost is projected to decline to around CNY 2300 per ton, or even as low as CNY 1900 per ton under favorable conditions. Overall, methanol produced from renewable sources is anticipated to reach cost parity with fossil-based methanol in the foreseeable future, thereby supporting the transition toward a low-carbon energy system.

The key contributions of this study are as follows: First, most prior research concentrates on the deployment of CCUS in the coal-fired power sector, primarily assessing the economic feasibility of retrofitting or constructing new units. In contrast, limited attention has been paid to its applications in the refining and petrochemical industries. This paper extends the scope of CCUS analysis by focusing on an integrated refining–chemical enterprise. Second, in contrast to conventional approaches that emphasize CO₂ utilization via EOR alone, this study proposes a hybrid utilization pathway that integrates EOR with methanol synthesis. By embedding this dual-route within a CCUS business model, the analysis captures synergistic benefits arising from diversification and enhanced resource efficiency, thereby offering a more resilient and commercially viable CCUS strategy. Third, a real options framework is developed to model investment decisions under multiple interacting uncertainties, including carbon and oil price fluctuations, dynamic capital costs driven by technological learning curves, and variable revenues from both EOR and methanol production. The model is solved through dynamic programming and least squares Monte Carlo simulation. Sensitivity analyses are performed on oil prices, carbon prices, photovoltaic electricity costs, EOR subsidies, and initial equipment investment subsidies to quantify their impacts on investment thresholds and optimal timing. These findings provide a rigorous decision-support tool for CCUS deployment in the petrochemical sector and deliver policy-relevant insights to inform the design of carbon pricing mechanisms and investment incentives.

The remainder of this paper is organized as follows: Section 2 presents a literature review on the application of real options theory in the CCUS field. Section 3 describes the problem and outlines the theoretical assumptions. Section 4 focuses on solving the model, while Section 5 contains a case study analysis. Finally, Section 6 concludes the study.

2. Literature Review

Research on CCS investments has increasingly shifted towards CCUS, with analytical methods moving from Net Present Value (NPV) to the more flexible real options method. Scholars worldwide have extensively examined the economic viability of CCS investments, particularly through techno-economic analyses. NPV models have evaluated the influence of carbon prices on CCS investments in power plants. In China, the critical carbon prices for CCS investments were estimated at USD 61 per ton for coal-fired plants, and USD 72 per ton for integrated gasification combined cycle (IGCC) plants [14]. Research showed that pulverized coal (PC) technology was the most cost-effective without carbon constraints, while IGCC is more advantageous under stringent regulations [15]. Further analysis indicates that pre-investments had limited impact on IGCC costs and that IGCC could achieve significant emissions reductions under moderate carbon prices [16]. Rubin et al. (2007) [17] highlighted the rising costs of CCS due to increased capital and operational expenses and stressed the need to assess environmental impacts. Zhai et al. (2015) [18] found that CCUS retrofits in U.S. coal-fired units could reduce CO₂ emissions by 30%. Liang et al. (2009) [19] identified factors such as investment type and carbon prices that influence the feasibility of CO₂ Capture Ready (CCR) investments in Guangdong, China. Dahowski et al. (2012) [20] showed that over 80% of major emission sources in China could be sequestered at a cost below CNY 70 per ton. It is important to note that NPV was a static tool that cannot adequately capture the variability and uncertainties associated with CCUS projects [21]. Consequently, it had significant limitations in investment evaluation, was ineffective in optimizing investment timing, and could not account for optimal timing or project uncertainties [22].

In contrast, the real options approach employed dynamic stochastic modeling to quantitatively incorporate managerial flexibility, significantly improving both asset valuation precision and risk mitigation capacity. Myers (1977) [23] introduced real options as a way to integrate financial options theory into investment decision-making, highlighting project uncertainty and managerial flexibility. Fuss et al. (2010) [24] modeled technological progress in CCS as a Poisson process and analyzed how uncertainties in technology and fuel prices affect CCS investment strategies. Compared to traditional discounted cash flow (DCF) methods, the real options approach was more effective in evaluating uncertain investments [25]. Real options pricing included discrete-time models, like tree-based methods, and continuous-time models, such as Monte Carlo simulations [26]. For instance, Zhang et al. (2021) [27] used a binomial tree to assess policy incentives for CCUS investment in China while accounting for uncertainties like carbon prices and technological progress. Trinomial trees and quadrinomial trees further expanded these models, allowing for additional scenarios, such as stable carbon prices. Fan et al. (2023) [28] combined a CCUS source-sink matching model with a trinomial tree to determine the critical carbon price for CCUS investment in China. Li et al. (2024) [29] applied a real options trinomial tree model for coal-to-hydrogen investments, identifying a critical carbon price of CNY 353.3 per ton and highlighting optimal regions for “blue hydrogen” production. Wang and Du (2015) [30] created a quadrinomial tree model for CCUS investments in coal-fired power plants, focusing on carbon trading and fuel prices. While multi-tree models faced challenges in capturing continuous carbon price fluctuations, Monte Carlo simulations generated multiple scenarios to more accurately assess investment values and timing. Research by Geske et al. (2010) [31] found that government carbon regulations push firms to monitor carbon prices, while technological innovation reduced profitability risks. Zhu and Fan (2011) [32] integrated real options and Monte Carlo simulations to evaluate CCS investment in China and found that climate policies greatly influenced CCS adoption. In further studies, they and others analyzed external and internal uncertainties affecting the feasibility of CCS retrofitting in power plants [33]. Chen et al. (2016) [34] created a model to analyze investment decisions under uncertainty, using Monte Carlo simulation to study how power generation subsidies affect CCS retrofitting in Chinese coal-fired power plants. Chu et al. (2016) [35] examined the influence of carbon prices, cost reductions, and subsidies on power plant investments, finding that carbon price fluctuations significantly affect investment timing. Yao et al. (2019) [36] concluded that although Coal-to-Liquids with Carbon Capture and Storage (CTL-CCS) projects are currently unviable, delayed investments may hold value under high carbon prices or supportive policies.

In addition, recent research on CCUS has increasingly focused on its business models and investment decision-making. CCUS business models primarily consisted of several operational modes, including state-owned enterprise operation, joint operation, independent transporter, and operator modes. These models could be fully funded and managed by a single entity, or a project company can be established for joint financing and collaboration among multiple stakeholders. Currently, the investment in China’s pilot demonstration projects mainly comes from key state-owned enterprises, particularly in the oil, gas, and electric power sectors. Yao et al. (2018) [37] identified four CCUS business models and found that the vertically integrated model is suitable for China’s early demonstration phase, recommending storage subsidies and stable carbon pricing to support large-scale deployment. Jarvis and Samsatli (2018) [38] assessed performance metrics of CO₂ conversion technologies, while Mikhelkis and Govindarajan (2020) [39] compared CCS and CCU investments at Exergi’s waste-to-energy plant, showing that a combined CCS-CCU approach has greater investment potential. Wang (2023) [40] used a system dynamics approach to determine that the vertically integrated model is most viable for coal-fired plants.

While research predominantly targets the coal power sector, the petrochemical industry, especially refineries, has received less attention. Refineries generated significant carbon emissions that could be used in carbon trading systems and as feedstocks for chemical processes. On the basis of previous studies, this paper establishes a project-level techno-economic evaluation model based on real option theory. The aim of this research is to address a gap in the academic literature and to guide investors in the petrochemical industry on how to apply the real options approach. This approach is intended to enhance the scientific rigor of investment decisions and to increase the flexibility value of management.

3. Problem Formulation and Model Establishment

3.1. Description of Issue

This study focuses on a refining–chemical integrated plant (RCIP) as the primary case study (Figure 1). China’s national carbon market currently includes key sectors such as power generation, steel, cement, and aluminum, and is rapidly expanding to the petrochemical and chemical industries. Based on this trend, this study assumes that the RCIP participates in the national carbon trading scheme. An integrated CCUS business model is adopted. The system encompasses catalytic and steam cracking (as the main CO₂ sources), CO₂ capture and transport, and subsequent utilization via two pathways: methanol synthesis and enhanced oil recovery. The steam cracker of refinery generates high-concentration, low-cost CO₂ suitable for chemical conversion. CO₂ is captured at the flue gas outlet using dedicated equipment, then purified, liquefied, and stored in on-site tanks. (1) Methanol Synthesis: The process consists of hydrogen production, CO₂ purification, catalytic methanol synthesis, and distillation. Hydrogen is generated via water electrolysis powered by local solar energy. Sixty percent of the captured CO₂ (150,000 t/a) is purified and hydrogenated under optimized temperature and pressure conditions. The methanol production system includes synthesis reactors (100,000 t/a), distillation columns, and compressors. Crude methanol is first synthesized and then refined through distillation to obtain high-purity methanol. The project achieves an annual methanol output of 110,080 tons. (2) EOR: The remaining 40% of CO₂ (100,000 t/a) is transported via tank trucks to nearby oilfields (within 100 km) for injection into reservoirs, increasing crude oil output. This integrated approach leverages upstream–downstream synergies to create a closed-loop carbon utilization system. It maximizes the use of point-source CO₂ emissions and enhances resource value through both EOR and chemical conversion.

3.2. Theoretical Assumptions

3.2.1. Carbon Price

Assuming that the carbon emission allowance price, hereinafter referred to as the carbon price, P_c(t) follows a geometric Brownian motion, it satisfies the following equation:

$$d{P}_{c\left(t\right)}={\mu }_c{P}_{c\left(t\right)}\mathit{dt}+{\sigma }_c{P}_{c\left(t\right)}\mathit{dz}$$

(1)

where μ_c represents the expected growth rate of the carbon price, σ_c denotes the volatility of the carbon price, and dz is a standard Wiener process that follows a standard normal distribution. Based on the average monthly carbon price data, we have the following:

$$r_t={\displaystyle \frac{P_{c\left(t+1\right)}}{P_{c\left(t\right)}}},(t=1,2\dots ,n)$$

(2)

Let P_c(t+1) and P_c(t) represent the average carbon prices in months t + 1 and t, and r_t represents the monthly growth factor of the carbon price. The parameters μ_c and σ_c can be estimated using the following equation:

$$\begin{array}{c}\left\{\begin{array}{c}\&\overline{r}={\displaystyle \frac{1}{n}}{\displaystyle \sum _{t=1}^n}\left(r_t-1\right)\\ \&S^2={\displaystyle \frac{1}{n-1}}\left[{\displaystyle \sum _{t=1}^n}{\left(r_t-1\right)}^2-n\overline{r}\right]\end{array}\right.\\ \left\{\begin{array}{c}\&\overline{r}={\mu }_{\mathit{month}}\Delta t\\ \&S^2={{\sigma }^2}_{\mathit{month}}\Delta t\end{array}\right.\Rightarrow \left\{\begin{array}{c}{\mu }_{\mathit{month}}=\&\overline{r}/\Delta t\\ {{\sigma }^2}_{\mathit{month}}={\&S}^2/\Delta t\end{array}\right.\Rightarrow \left\{\begin{array}{c}{\mu }_c=12{\mu }_{\mathit{month}}\\ {\sigma }_c={\sigma }_{\mathit{month}}\sqrt{12}\end{array}\right.\end{array}$$

(3)

Denote (r_t − 1) as the simple monthly return for month t,$\&\overline{r}$ as the sample mean of these n simple monthly returns, and S² as the sample variance of these n simple monthly returns. The time interval for the monthly statistics is represented by Δt, where Δt = 1 corresponds to one month. Let μ_month represent the expected monthly simple return (estimated by $\&\overline{\mathrm{r}}$ when Δt = 1). Its standard deviation, σ_month, is given by the square root of the variance of monthly simple returns. Finally, μ_c and σ_c denote the estimated annualized expected growth rate and annualized volatility, respectively, derived from these monthly statistics.

3.2.2. Oil Price

Suppose the crude oil price P_c(t+1) follows a geometric Brownian motion, it satisfies the following equation:

$$d{p}_{\mathit{oil}\left(t\right)}={\mu }_{\mathit{oil}}{p}_{\mathit{oil}\left(t\right)}\mathit{dt}+{\sigma }_{\mathit{oil}}{p}_{\mathit{oil}\left(t\right)}\mathit{dz}$$

(4)

where μ_oil represents the expected growth rate of the oil price, σ_oil denotes the volatility of oil prices, and dz refers to a standard Wiener process that follows a standard normal distribution. The calculation methods for μ_oil and σ_oil are similar to those for μ_c and σ_c, so further details will not be provided here.

3.2.3. Equipment Investment and Operating Costs

Assuming that CCUS equipment investment is a one-time expense, the retrofit cost is denoted by I, with an initial value of I_C,0. The variable Y denotes the operating cost, with Y_C,0 as its initial value. As technological progress continues, both the investment and operating costs decrease over time, following the learning curve effect. The costs associated with CCUS equipment investment and operation follow the equations below:

$$I=I_{C,0}\times *e^{-\mathit{it}}$$

(5)

$$Y=Y_{C,0}\times e^{-\mathit{lt}}$$

(6)

The variable i represents the effect of technological advancements on equipment retrofit costs, and l indicates the influence of technological progress on operating costs after retrofitting.

3.2.4. Assumptions on Project Lifespan and Investment Period

The project period, also known as the option validity period, T is defined as the timeframe within which the RCIP must make an investment decision. The construction period is expected to last two years, and the operational lifespan of the facility after completion is specified as TL.

3.3. Model Framework

3.3.1. Investment Return Model

Based on the analysis of the above indicators and the model assumptions, the additional annual revenue CI_t generated by the CCUS project can be expressed as follows:

$$C{I}_t=C{I}_t^C+C{I}_t^O+C{I}_t^M$$

(7)

Define $C{I}_t^C$ as the revenue generated from carbon emission trading as a result of carbon reduction efforts. Therefore

$$C{I}_t^C=p_{c\left(t\right)}\times q_c\times w_1$$

(8)

The term $C{I}_t^O$ represents the incremental crude oil revenue resulting from EOR. Thus

$$C{I}_t^O=p_{\mathit{oil}\left(t\right)}\times q_c\times w_1\times w_2\times \alpha +q_c\times w_1\times w_2\times p_y\times \alpha$$

(9)

where p_oil denotes the crude oil price; q_c refers to the carbon dioxide reduction (the amount of carbon dioxide captured); w₁, w₂ represent the carbon capture rate and the proportion of captured CO₂ used for EOR, respectively; and α indicates the oil displacement efficiency. Here, p_oil(t) * × q_c × w₁ × w₂ × α refers to the revenue generated from oil displacement, while p_c(t) represents the price of carbon emission allowances. Additionally, p_c(t) × q_c × w₁ denotes the revenue from reducing CO₂ emissions. The variable q_c × w₁ × w₂ × p_y × α indicates the government subsidy for EOR. p_y represents the unit incremental subsidy for EOR.

$C{I}_t^M$ denotes revenue from the synthetic methanol module:

$$C{I}_t^M=q_M\times p_M$$

(10)

Here, q_M denotes the volume of methanol production, p_M refers to the market unit price of methanol, and q_M × p_M reflects the revenue generated from the chemical utilization of methanol production.

The additional annual cash outflow, CO_t, can be expressed as follows:

$$C{O}_t=C{O}_t^O+C{O}_t^H+C{O}_t^M$$

(11)

Let $\mathrm{C}{\mathrm{O}}_{\mathrm{t}}^{\mathrm{O}}$ refer to the cost of the EOR module, $\mathrm{C}{\mathrm{O}}_{\mathrm{t}}^{\mathrm{H}}$ to the cost of the hydrogen production module, and $\mathrm{C}{\mathrm{O}}_{\mathrm{t}}^{\mathrm{M}}$ to the cost of the synthetic methanol module. Then

$$C{O}_t^O=q_c\times w_1\times u_c+q_c\times w_1\times u_t\times S+q_c\times w_1\times w_2\times u_E+Y_{C,t}$$

(12)

Here, u_c denotes the unit capture cost, and q_c × w₁ × u_c indicates the total capture cost. u_t refers to the unit transportation cost. S indicates the transportation distance from the RCIP to the oil field; q_c × w₁ × u_t × S captures the total transportation cost. u_E signifies the unit oil displacement cost, and q_c × w₁ × w₂ × u_E corresponds to the total oil displacement cost. Finally, Y_C,t denotes the operational cost of the capture device, taking into account the learning curve effect.

$$C{O}_t^H=q_H\times u_H+Y_H+W_H$$

(13)

where q_H refers to the amount of hydrogen consumed for methanol production, u_H refers to the unit hydro-power cost for hydrogen production, and q_H × u_H denotes the total cost of the hydrogen production process. Y_H indicates the operational and maintenance costs associated with the hydrogen production unit. W_H signifies the employee welfare expenses related to the hydrogen production unit.

$$C{O}_t^M=U_M+q_M\times u_t^M+Y_M+W_M$$

(14)

In this context, U_M represents the hydroelectric cost associated with methanol production, while q_M denotes the amount of methanol output. Additionally, $u_t^M$ refers to the unit transportation cost for selling methanol. The term q_M × $u_t^M$ indicates the total transportation costs incurred for the sales of methanol products. Y_M encompasses both the operational and maintenance costs of the methanol production facility, as well as the hydroelectric expenses. Lastly, W_M represents the employee welfare costs related to the methanol synthesis unit.

Thus, the net annual cash flow NCF_t generated from investing in the CCUS project can be expressed as follows:

$${\mathit{NCF}}_t=C{I}_t-C{O}_t$$

(15)

Using the discounted cash flow formula with continuous compounding interest, if ${\mathrm{e}}^{\mathrm{r}}=1+\mathrm{r}$ is provided, the NPV of the CCUS project at time t can be expressed as follows:

$$\mathrm{NPV}={\displaystyle \sum _{i=t}^T}{\mathit{NCF}}_t\times e^{-\left(i-t\right)r}-(1-\theta )I_{C,t}-I_M-I_H$$

(16)

Let θ represent the government subsidy coefficient for the RCIP’s investment in CCUS, which primarily covers the costs associated with modifying equipment. (1 − θ)I_C,t refers to the investment required for the RCIP’s equipment modifications at a specific time t. Let r denote the risk-free rate. Define I_M as the investment amount allocated to methanol production facilities, and I_H as the investment amount associated with hydrogen production infrastructure.

The value of the investment decision made by the investor at any given time should reflect the expected value of the project under various uncertain conditions. Therefore, the investment value V_t of the project in year t can be expressed as follows:

$$V_t=\mathrm{E}\left(\sum _{\mathrm{i}=\mathrm{t}}^{\mathrm{T}}{\mathit{NCF}}_t\times e^{-\left(i-t\right)r}-(1-\theta )I_{C,t}-I_M-I_H\right)$$

(17)

3.3.2. Investment Decision Model

The decision to invest in CCUS can be likened to a time-continuous American call option. This means that investors have the flexibility to defer their investment for an indefinite period, depending on prevailing conditions. In this context, we denote the investment decision variable as ${\chi }_k$:

$${\mathrm{\chi }}_{\mathrm{k}}=\left\{\begin{array}{c}1,\mathrm{make}\mathrm{the}\mathrm{investment}\mathrm{in}\mathrm{year}\mathrm{k}\\ 0,\mathrm{defer}\mathrm{the}\mathrm{investment}\mathrm{in}\mathrm{year}\mathrm{k}\end{array}\right.$$

(18)

An investor has the option to invest now or postpone their decision by comparing the present value of an immediate investment with the expected future value of the project. Assuming rational behavior, the investor will choose to invest when the project’s investment value is at its highest. The specific form should be the following:

$$\mathrm{F}=\underset{1\le t\le T}{\max }\left[\max (V_t,0)\times e^{-\mathit{rt}}\right]$$

(19)

where F denotes the maximum investment value of the project.

4. Model Solution

4.1. Solution Method

This study utilizes the least squares Monte Carlo (LSM) simulation method, which combines Monte Carlo simulation, least squares regression, and dynamic programming. LSM is particularly effective for determining continuous values and optimal investment timing, showcasing high accuracy and efficiency, especially in the pricing of American options. By solving recursively in a backward manner, LSM effectively identifies the optimal exercise boundary, making it a powerful tool for pricing financial derivatives.

4.2. Solution Process

Based on the basic approach previously outlined for the LSM method, the steps of the algorithm for the entire model are as follows:

Simulate the path of the carbon price P_c(t) by dividing the option’s maturity period T into N − 1 investment intervals. Including the initial investment point, there are N investment points to consider, with each interval having a length of $\Delta t={\displaystyle \frac{T}{N}}$.

The carbon price follows a geometric Brownian motion, as described in Equation (1). To better capture its trajectory, Ito’s Lemma is applied to transform the equation. We start by defining a smooth function that closely approximates the geometric Brownian motion.

$$\mathrm{f}(P_{c\left(t\right)},t)=\ln (P_{c\left(t\right)},t)$$

(20)

By applying the chain rule of Ito’s Lemma, the following expression can be derived:

$$\mathrm{df}={\displaystyle \frac{\partial \mathrm{f}}{\partial P_{c\left(t\right)}}}\mathrm{d}{P}_{c\left(t\right)}+{\displaystyle \frac{\partial \mathrm{f}}{\partial \mathrm{t}}}\mathit{dt}+{\displaystyle \frac{1}{2}}{\displaystyle \frac{{\partial }^2\mathrm{f}}{\partial P_{c\left(t\right)}^2}}{\mathrm{d}}^2{P}_{c\left(t\right)}$$

(21)

Using the expression of function $\mathrm{f}(P_{c\left(t\right)},t)$, the solution is derived as follows:

$$\left\{\begin{array}{c}{\displaystyle \frac{\partial f}{\partial P_{c\left(t\right)}}}={\displaystyle \frac{1}{P_{c\left(t\right)}}}\\ {\displaystyle \frac{\partial f}{\partial t}}=0\\ {\displaystyle \frac{{\partial }^{2}f}{\partial P_{c\left(t\right)}^2}}=-{\displaystyle \frac{1}{P_{c\left(t\right)}^2}}\end{array}\right.$$

(22)

Substituting into Equation (1) and performing the integration, the following result is obtained:

$$P_{c\left(t+\Delta t\right)}=P_{c\left(t\right)}\exp \left[({\mu }_c-{\displaystyle \frac{{{\sigma }_c}^2}{2}})\mathrm{\Delta }\mathrm{t}+{\sigma }_c\mathrm{z}\sqrt{\mathrm{\Delta }\mathrm{t}}\right]$$

(23)

In the subsequent steps of the solution process, σ_c and μ_c are constant. Therefore, for any given time t, the following holds:

$$P_{c\left(t\right)}=P_{c\left(0\right)}\exp \left[(u_c-{\displaystyle \frac{{{\sigma }_c}^2}{2}})\mathrm{\Delta }\mathrm{t}+{\sigma }_c\mathrm{z}\sqrt{\mathrm{t}}\right]$$

(24)

Since V_t is a function of P_c(t), the corresponding path for V_t can be obtained based on the simulated carbon and oil price paths.

Simulate multiple investment value paths based on the volatility of carbon and oil prices. For each path j, it is necessary to iterate backward from the end investment time point (excluding the terminal node) to the starting investment time point. At each step, the expected return from delaying the investment is compared to the return from making an immediate investment, and the decision information, along with the maximized investment value, is updated. Thus, for path j, we have the following:

$${\mathrm{F}}_{\mathrm{t},\mathrm{j}}=\underset{1\le \mathrm{t}\le \mathrm{T}}{\max }\left[V_{t,j},E_t\left(F_{t+1,j}\right)\times e^{-r}\right]$$

(25)

Define F_t,j as the maximum investment value achieved along the path after the iteration process; V_t,j indicates the current investment value of the project at a specific point on path j and $E_t\left(F_{t+1,j}\right)\times e^{-r}$ refers to the expected value of continuing to hold the investment opportunity, which is calculated using a basis function fitted through least squares after the decision is made to delay investment at the current time point.

At this stage, the decision variable ${\chi }_{j,t}$ at time t on the corresponding path j can be expressed as follows:

$${\chi }_{j,t}=\left\{\begin{array}{c}1,V_{t,j}>E_t\left(F_{t+1,j}\right)\times e^{-r}\\ 0,V_{t,j}<E_t\left(F_{t+1,j}\right)\times e^{-r}\end{array}\right.$$

(26)

The variable Q denotes the total number of paths. Repeat the previous steps for all paths to identify the optimal investment time point and the corresponding maximum investment value for each path. Let t_max,j represent the optimal investment time point on path j. The final optimal investment time point t_max should be the mode of the set {t_max,j}. The maximum investment value F_max of the project is the average investment value at time t_max across all Q paths. The specific formulation is as follows:

$$\left\{\begin{array}{c}{t}_{\max }=\mathit{inf}\left\{t\right|{\chi }_{j,t}=1\},t\in [1,T]\\ F_{\max }=E({\displaystyle \sum _{j=1}^Q}{F}_{t_{\max ,j}}*e^{-r*t_{\max }}),j\in [1,Q]\end{array}\right.$$

(27)

5. Case Analysis

5.1. Model Parameters

Using the parameters and their specified values listed in

Table 2. Related parameters.

Parameter	Unit	Value	Data Source
μc	Carbon Price Drift Rate (%)	0.01726	Using China’s carbon market price data (up to 2024), calculate using Formula (4)
σc	Carbon Price Volatility (%)	0.01419
μ0	Crude Oil Price Drift Rate (%)	0.0092	With reference to Brent crude oil prices over the past 5 years, perform calculations based on a modified version of Formula (4)
σ0	Crude Oil Price Volatility (%)	0.0252
poil	Initial Crude Oil Price (USD/barrel)	72	Using the 5-year average Brent crude price, converted at the 2024 USD to CNY exchange rate
pc	Initial Carbon Price (CNY/ton)	92	With reference to the average carbon price in China’s carbon market in 2024
qc	Carbon Dioxide Reduction (ton)	280,000
w1	Carbon Dioxide Capture Rate (%)	90	Wang and Gao, 2025 [41]
w2	Proportion of Carbon Dioxide Used for EOR (%)	40
uc	Unit Carbon Dioxide Capture Cost (CNY/ton)	250	Zhang, X., Yang, X., Lu, X. et al., 2023 [6]
$u_E$	Unit Carbon Dioxide EOR Cost (CNY/ton)	349
ut	Unit Carbon Dioxide Transportation Cost (CNY/(ton·km))	1	Cai, B. F., Li, Q., Zhang, X. et al., 2021 [42]
IC,0	Initial carbon capture unit retrofit cost (CNY 10,000)	22,644	Yang et al., 2019 [43]
YC,0	Initial operating cost for carbon capture system (10,000 CNY/year)	377
i	The impact of technological advancements on the retrofit cost of carbon capture units	0.021
l	The effect of technological progress on the cost of retrofitting carbon capture equipment	0.057
py	Unit EOR incremental subsidy (CNY/barrel)	0	Set the parameters based on current Chinese policies
α	Enhanced the oil recovery rate (%)	15	Nikolova, C., and Gutierrez, T., 2020 [44]
θ	Government’s Initial Subsidy Coefficient for CCUS (%)	0	Set the parameters based on current Chinese policies
S	Transportation distance of tank trucks from the plant to the oil displacement site (kilometers)	100
$T$	Investment period for the CCUS project (years)	20	Tan et al., 2024 [45]; Lin and Tan, 2021 [46]
$T_L$	Project operational lifespan (years)	30
$r$	Risk-free interest rate (%)	6	Rubin et al., 2007 [47]
qH	Hydrogen consumption for methanol production (tons)	21,000	This parameter is primarily based on the green methanol synthesis unit of the 100,000 tons per year Liquid Sunshine Project in Ordos, Inner Mongolia, China
uH	Water and electricity cost per unit of hydrogen produced via water electrolysis (CNY/ton)	3526	Zhang et al., 2023 [48]
WH	Employee welfare costs for the hydrogen production unit (10,000 CNY/year)	1200
IH	Investment cost for the hydrogen synthesis unit (CNY 10,000)	81,000	Zhang et al., 2023 [48]. The hydrogen production capacity is up to 48,000 m3/h, with an annual hydrogen output of 20,000 tons. Alkaline water electrolysis equipment was selected as the primary hydrogen production technology
YH	Operation and maintenance costs of the hydrogen production facility (10,000 CNY/year)	1620	$\mathrm{Calculate}\mathrm{at}2\%\mathrm{of}{\mathrm{I}}_{\mathrm{H}}$
qM	Methanol production volume (tons)	110,080	A standard commercial unit size for methanol plants
pM	Unit price of methanol (CNY)	4300	Zhu et al., 2023 [49]; the price represents a reasonable mid-term expectation and falls within the cost range of CO2 hydrogenation-based methanol production in China
${\mathrm{u}}_{\mathrm{t}}^{\mathrm{M}}$	Transportation cost per unit of methanol sold (CNY/ton)	145	Wang et al., 2024 [50]
IM	Investment cost for the methanol synthesis unit(CNY 10,000)	60,000
YM	Operation and maintenance costs of the methanol production facility (10,000 CNY/year)	900	Shu et al., 2024 [51]
WM	Employee welfare costs for the methanol synthesis unit (10,000 CNY/year)	800	40 staff, with an annual average salary of CNY 200,000 per person
UM	Water and electricity costs for methanol production (10,000 CNY/year)	11,516.8	Power usage: 13,961 kWh/h; water usage: 435 tons/h

, the model is calculated to identify the optimal timing for investment and the associated investment value under the given conditions. Following this, a sensitivity analysis is conducted on the key parameters.

5.2. Calculation Results

Figure 2 showcases 100 sample paths, which are selected from 1,000,000 simulated paths generated using Matlab(R2021b). Each sample path is represented by a line in a different color, illustrating the potential variations in carbon and oil prices over time. The carbon price demonstrates a general upward trend as time progresses, while the oil price displays a range of patterns, including increases, decreases, and periods of stability.

The ideal timing for investing in CCUS projects is determined by specific parameters. Across all simulated scenarios, the most favorable time for investment is consistently identified as the year 2033, with a probability approaching 100%. Therefore, this study concludes that 2033 is the optimal year for making investments in the entire project. In a delayed investment scenario, the year represents the point at which the maximum investment value can be achieved.

$$E\left(NP{V}_i\right)=20520000yuan(i=1,2,3\dots \dots 1000000)$$

(28)

When the investment project’s option value is not considered, and the traditional NPV method is applied, the value of investing immediately in the project in 2025 is calculated as follows:

$$E\left(NP{V}_j\right)=7162100yuan(j=1,2,3\dots \dots 1000000)$$

(29)

The calculations indicate that the additional profit from delaying the investment in the project amounts to CNY 13.3579 million when considering the option value. The decision to postpone the investment is primarily influenced by four factors: carbon price, oil price, initial investment costs, and operating costs.

When the investment is delayed, the stable additional return from postponing it for one year arises from reductions in initial investment costs and operating expenses, referred to as A. Conversely, the risk associated with delaying the investment primarily stems from the missed profits that could have been gained from carbon dioxide capture and utilization, labeled as B.

If A exceeds B, it implies that the stable additional return from delaying the investment is greater than the opportunity loss incurred by not investing and missing out on the benefits of carbon dioxide capture and utilization. In this scenario, delaying the investment is more profitable than proceeding with it immediately.

The optimal time to invest is identified when the cumulative difference between A and B reaches its maximum. Beyond this point, the annual return from reductions in initial investment costs and operating expenses will consistently be lower than the additional profits generated by carbon dioxide capture and utilization activities.

To ensure the reliability of the LSM model, this paper evaluates its effectiveness through a convergence test. By setting different numbers of simulation paths (10,000; 100,000; 500,000; 1,000,000), the stability of two key indicators, the optimal investment time and the maximum investment value, was tested. The results show that the optimal investment time is stable in the eighth year under each simulation path, and the relative difference in the maximum investment value is less than 0.01% under 500,000 and 1,000,000 paths. This shows that the 1,000,000 simulation paths used in this study can ensure the convergence of the results.

5.3. Sensitivity Analysis

5.3.1. Sensitivity of Carbon Price

Figure 3 presents the relationship between initial carbon price levels and project investment decisions, assuming ceteris paribus. A strong correlation exists between carbon pricing and investment decision-making for the project. Three critical carbon price thresholds are identified: at 103, 115, and 123 CNY/ton, the optimal investment timing is to advance progressively from 2033 to 2032, 2031, and 2030, respectively. As the carbon price rises to CNY 125 per ton or higher, it triggers the conditions for immediate investment, allowing the investor to proceed. Consequently, targeted policy incentives should be implemented to stimulate near-term investment decisions at prevailing carbon price levels.

5.3.2. Sensitivity of Carbon Price Volatility

Figure 4 illustrates the investment decisions for the project under varying levels of carbon price volatility, while keeping other conditions constant. It is evident that when the carbon price volatility is approximately 0.058, the optimal time for investment is projected to be in 2033. As the carbon price volatility increases to 0.2, the optimal investment time is delayed until 2036, and the maximum investment value declines. When carbon price volatility exceeds 0.25, the optimal time for investment is further postponed to 2039, and the decrease in the maximum investment value becomes more significant. This suggests that more stable carbon price volatility can enhance investment in the project, resulting in higher returns.

5.3.3. Sensitivity of Oil Price Levels

The project’s investment decisions across various initial oil prices are illustrated in Figure 5, while keeping all other conditions constant. We observe that as the initial oil price increases, the maximum investment value also rises. Price increases below USD 70 per barrel have minimal impact on investment schedules. A phase transition is observed in the USD 70–105 per barrel range, where the optimal investment year advances from 2034 to 2033. This temporal optimization continues as prices reach USD 106–133 per barrel, with investment timing shifting to 2032. The most notable change occurs when the price surpasses the USD 134 per barrel threshold, pushing the optimal investment year to around 2031. These results quantitatively demonstrate that oil price escalation effectively drives investment acceleration in CCUS projects.

5.3.4. Sensitivity of EOR Subsidies

Based on varying EOR subsidies, with all other conditions held constant, Figure 6 illustrates the project’s investment decisions. The data indicates that when the EOR subsidy is less than or equal to CNY 25 per barrel, its effect on stimulating investment is limited. However, when the subsidy increases from CNY 25 per barrel to CNY 75 per barrel, the optimal investment timing shifts to 2031. If the subsidy exceeds CNY 75 per barrel, immediate investment conditions are activated, further advancing the investment timing to 2025. These findings suggest that policy support proves crucial for securing prompt investor engagement given existing conditions.

5.3.5. Sensitivity of Equipment Investment Subsidy Coefficient

The investment dynamics of the project, in relation to varying levels of equipment investment subsidies, are illustrated in Figure 7. The immediate effect occurs when the equipment investment subsidy coefficient reaches or exceeds 0.2, indicating that the incentive policy is effective even at relatively low subsidy levels. Furthermore, once the subsidy coefficient exceeds a certain threshold, the maximum investment value of the project continues to increase while the optimal timing for investment stabilizes. At this stage, the marginal benefit of further increasing the subsidy coefficient begins to decline. Therefore, it is essential to consider the trade-off between the enhanced incentive effects of higher equipment investment subsidies and the associated fiscal costs to optimize the policy’s effectiveness.

5.3.6. Sensitivity of Photovoltaic Levelized Electricity Cost Coefficient

Given that the LCOE from photovoltaic (PV) sources is a key determinant of hydrogen production costs in this project, the impact of hydrogen cost on investment decisions is assessed indirectly by analyzing variations in PV electricity prices. Currently, the average on-grid price for centralized PV power in China ranges from CNY 0.3 to 0.5 per kWh; the lower bound is adopted for this analysis. According to Figure 8, when the LCOE remains relatively high (e.g., CNY 0.30/kWh), the optimal investment timing is significantly delayed. As the LCOE steadily declines from CNY 0.30 to approximately CNY 0.22/kWh, the investment window advances markedly, shifting from around 2037 to 2032. A further reduction to just above CNY 0.12/kWh continues to bring investment forward. Notably, when the LCOE reaches CNY 0.12/kWh or below, immediate investment becomes economically viable, and the optimal investment year moves up to as early as 2025. The findings indicate that the LCOE plays a critical role in determining the optimal timing of project investment. Declining electricity costs are shown to significantly accelerate investment decisions.

6. Conclusions

This study applies real options theory to evaluate the maximum investment value and optimal timing for refining–chemical integrated plants in the context of CCUS. It considers uncertainties related to carbon prices, the volatility of oil prices, and the reduction in both initial investment and operating costs due to technological advancements. The model is solved using the Matlab tool, followed by a sensitivity analysis to assess the effects of carbon prices, their volatility, oil prices, EOR subsidies, and coefficients for initial investment subsidies. The main findings are summarized as follows:

First, carbon pricing emerges as a primary determinant of CCUS investment behavior. Prices in China’s national carbon market remain relatively low and are insufficient to incentivize large-scale deployment; investment becomes viable when the initial carbon price reaches CNY 125 per ton. Compared to oil prices—where investment is triggered only when prices exceed USD 134 per barrel—carbon pricing demonstrates a stronger influence on decision-making, particularly under the “CO₂-to-methanol + EOR” model.

Second, volatility in carbon pricing significantly affects investor confidence and strategy. Large fluctuations may delay or deter investment, underscoring the need for a stable and transparent carbon market. Strengthening carbon price signals through policy refinement and market development is essential to de-risk CCUS investment and accelerate adoption.

Third, targeted subsidies—especially for enhanced oil recovery (EOR) and initial equipment investment—play a catalytic role. When EOR subsidies reach CNY 75 per barrel or equipment subsidies exceed a coefficient of 0.2, immediate investment becomes economically viable. The results reinforce the necessity of strong government support through well-designed financial incentive frameworks.

Lastly, the LCOE from photovoltaic sources is a critical determinant of hydrogen production economics. A sustained decrease in LCOE significantly accelerates the optimal investment window. When the LCOE falls to approximately CNY 0.12/kWh or lower, projects become economically viable, and investment can be brought forward—potentially as early as 2025. These findings underscore the strategic importance of lowering renewable electricity costs to enable the large-scale deployment of green hydrogen and CCUS technologies.

This study offers important insights into CCUS investment decisions by simulating the behavior of investors in integrated refining–chemical plants. While valuable, the analysis remains subject to limitations. Real-world conditions are more complex than the model assumes, and volatile factors such as crude oil, carbon, and methanol prices introduce additional uncertainty. In particular, the methanol price is treated as a fixed value sufficient to cover production costs, without reflecting its market volatility. Furthermore, key risk dimensions—including market, technological, and policy risks—are not comprehensively addressed. Future research should adopt more advanced risk modeling to capture these dynamics and develop a structured policy scenario framework with parameterized sub-scenarios to systematically assess the economic viability and investment risks of CCUS under different policy pathways. Finally, as the study is based on a business model that integrates refining and chemical production, certain data inputs may require adjustment depending on commercialization levels and specific business configurations.