Life Cycle Greenhouse Gas Emissions Analysis of the Chlor-Alkali Process and By-Product Hydrogen in the United States

Abstract

Hydrogen is considered a key energy carrier for which interest has grown over recent years. Chlor-alkali plants in the United States (U.S.) can potentially recover and supply the by-product hydrogen at scale. However, there is a scarcity of standard analysis for energy use and emissions associated with products from chlor-alkali plants owing to lack of data and variations in chlor-alkali plant technology and operation. A rigorous life cycle analysis (LCA) is needed to quantify the emissions of by-product hydrogen and other products from chlor-alkali plants. In this study, we performed well-to-gate (WTG) emissions analysis of chlor-alkali products based on U.S. plant operating data gathered from the U.S. Environmental Protection Agency’s (EPA’s) Chemical Data Reporting database, the U.S. Energy Information Administration survey EIA-923 form, and the EPA’s Greenhouse Gas Reporting Program. We performed process-level mass allocation to allocate energy use and emissions to the chlor-alkali products. This study shows that the by-product hydrogen has WTG CO₂ emissions of 1.3–1.9 kgCO₂/kg H₂ for plants without combined heat and power (non-CHP) and 1.5–2.4 kgCO₂/kg H₂ for plants with combined heat and power (CHP). Furthermore, we identified that electricity upstream emissions are the key driver affecting the emissions of by-product hydrogen from non-CHP plants, while CHP emissions can be reduced by electricity export to grids with higher carbon intensity (CI). Finally, the study shows that chlor-alkali plants in the U.S. can potentially meet up to 320 kilotons of hydrogen demand (approximately 3% of total demand) annually.

1. Introduction

The chlor-alkali industry is a mature industry that produces chlorine (Cl₂), sodium hydroxide (NaOH), and hydrogen (H₂) through the electrolysis of brine. Three main technologies are used to perform electrolysis in chlor-alkali plants: mercury, diaphragm, and membrane technology. Mercury technology is the oldest of the three technologies, mainly used in Europe (holding a 17% production share) [1]. In the U.S., plants have either converted from mercury technology to membrane technology or phased it out because of environmental concerns related to mercury pollution. Diaphragm technology is prevalent in the U.S.; however, its share has been shrinking as it has been replaced by membrane technology. Until the year 2000, membrane technology was the most common technology in Japan [2].

The key advantages of mercury technology are a higher product quality (i.e., highly concentrated NaOH at 50% is the final product) with a lower brine quality requirement. However, the disadvantage is higher electricity consumption and greater environmental impact (especially related to the toxicity of mercury). The advantages of diaphragm technology are lower electricity consumption and lower quality of raw material required, with disadvantages including higher thermal energy input for concentrating NaOH owing to the production of a low concentration of NaOH (10–12 vol%) [2]. Additionally, some diaphragm facilities still use asbestos in the cells, which negatively impacts both human health and the environment [3].

With the increase in environmental concerns arising from mercury and asbestos emissions, mercury technology is being phased out. Similarly, obsolete diaphragm technologies are being either decommissioned or converted. Membrane technology is the latest commercial technology, having been introduced in 1970; all newly commissioned plants are based on membrane technology [4]. The advantages of membrane technology are lower electricity consumption and higher-quality NaOH (with NaOH at 33 vol%) [2]. However, membrane technology requires higher-quality brine, produces lower-quality chlorine, has higher thermal energy consumption, and requires higher capital investment. Additionally, developments have been made in chlor-alkali technology on the lab scale by using oxygen-depolarized cathodes instead of hydrogen-evolving cathodes, resulting in electrical energy savings of 28% over membrane technology [5]. However, there are major challenges: the current membrane technology cannot be retrofitted to use this technique because of the difference in operating conditions of oxygen-depolarized cathodes, and commercial-scale development would require significant capital investment. Other technical challenges include the requirement for a stable gas–liquid interface at the cathode, accelerated corrosion due to the presence of oxygen, and formation of hydrogen peroxide [5,6]. More details on each commercial-scale technology and a process figure are provided in the accompanying Supplementary Information (SI).

The electrochemical process that electrolyzes saturated brine (sodium chloride [NaCl]) or potassium chloride (KCl) to produce Cl₂, NaOH (or potassium hydroxide [KOH]), and hydrogen H₂ is shown in Equation (1).

2NaCl + 2H₂O → Cl₂ + H₂ + 2NaOH

(1)

Chlorine and sodium hydroxide are considered the main products from electrolysis, with 46.4% and 52.3% production share by mass, respectively, and hydrogen is produced with a remaining share of 1.3%. Both chlorine and sodium hydroxide have applications across various sectors. A major use of chlorine is the production of polyvinyl chloride (PVC), which is used in pipes, flooring, doors, and window frames. Additionally, chlorine is used as a disinfectant and in water treatment, plastics, insulation, and pesticides. Chlorine is also used as a solvent to degrease metal and adhesives, and in silicon rubber and lubricants [2]. Chlorine is usually produced near consumers because of storage and transportation challenges. Therefore, for long-distance transportation, chlorine is converted to its derivatives, i.e., ethylene dichloride (EDC), vinyl chloride monomer (VCM), and PVC. Unlike chlorine, sodium hydroxide is relatively easy to transport, mainly in the form of 50 wt.% caustic solution or, rarely, in the form of solid prills and flakes [2]. Sodium hydroxide has wide applications, including soaps, mineral oils, bleach, detergents, glass, and ceramics. In the metal industry, sodium hydroxide is used for steel hardening and manufacturing of car and airplane panels. Additional uses involve acid neutralization, gas scrubbing, and rubber recycling [2]. Sodium hydroxide is also important for the treatment of cellulosic biomass to enable its biochemical conversion to biofuels and bioproducts.

Hydrogen is produced as a by-product of brine electrolysis. Hydrogen has broad applications in various industries in modern society as a chemical reagent and/or fuel. For example, conventionally, hydrogen has been used as a chemical reagent in refining, ammonia production, and metal production; as liquid fuel for aerospace applications; and as heating fuel. In 2018, about 10 MMT hydrogen was produced in the U.S., of which about 90% was produced from the steam methane reforming process, while the rest came from electrolysis of water [7]. In terms of hydrogen application, typically about 68% of this hydrogen is used in the petroleum refining industry and 21% goes into fertilizer production [7].

The hydrogen demand in the U.S. in the next 2–3 decades is projected to exceed current market production capacity; thus, it is of interest to identify alternative hydrogen supply resources and tap into unrecovered hydrogen from existing sources to alleviate potential supply shortages in the near term. In particular, the hydrogen from chlor-alkali plants is expected to incur a low recovery cost because of the highly pure nature of the co-produced H₂ (99%) [1].

Chlor-alkali plants in the U.S. vary in complexity, unit processes, and product slates. In addition to the electrolysis-generated products of chlorine, caustic soda (NaOH)/caustic potash (KOH), and H₂, some facilities also have downstream processes to produce other chemical products, such as hydrochloric acid (HCl), sodium hypochlorite (NaClO), ethylene dichloride (EDC), and vinyl chloride monomer (VCM). Figure 1 depicts a general scheme of chlor-alkali plant processes.

Each chlor-alkali facility has distinct impacts because of differences in plant complexity, product output, hydrogen handling, and regional factors, making it challenging to standardize the LCA of chlor-alkali processes. Most facilities in the U.S. either export hydrogen to nearby facilities such as hydrogen liquefaction plants or vent the hydrogen, with possible indirect global warming implications. Additionally, by-product hydrogen requires different downstream activities, resulting in a different process-level system boundary from other chlor-alkali products, which complicates the environmental greenhouse gas (GHG) assessment of hydrogen. Furthermore, the literature evaluations of LCAs of chlor-alkali processes only cover the production of the main products from the electrolytic process, i.e., Cl₂, NaOH, and H₂ [1]. Chlor-alkali main products are further used (partially or completely) to produce other products such as HCl, NaClO, VCM, and EDC not evaluated in previous studies. Because of the complexity associated with different technologies, diverse product handling practices (e.g., combusting, exporting, or venting by-product hydrogen), and varying system boundaries based on final products, there is a need for a complete assessment of the environmental impacts of the chlor-alkali process and all of its final products.

This study aims to address the research gap by conducting a detailed LCA of chlor-alkali processes and their final products in the U.S., with a particular focus on GHG emissions from co-produced hydrogen. The LCA is carried out by gathering the facility-level data from several chlor-alkali plants in the U.S. and highlights the key parameters that govern the energy use and emissions of these plants. Additionally, the current state-of-the-art chlor-alkali plants generally vent the co-produced hydrogen and while advancements in chlor-alkali technologies have reduced their environmental impact, the underutilization of the high-purity by-product hydrogen remains a missed opportunity to support the growing hydrogen economy.

In light of the importance of hydrogen demand for the decarbonization of energy sectors, chlor-alkali facilities can play an important role by recovering and exporting the hydrogen and generating revenue. The exported hydrogen can support decarbonization through its use in the refining and fertilizer industries, and in emerging applications such as sustainable fuel and chemicals synthesis, biomass upgrading, blending in natural gas pipeline infrastructure, and fuel cell electric vehicles. Therefore, LCAs of the by-product hydrogen and other chlor-alkali products are needed. The scope of our study includes well-to-gate (WTG) energy use and CO₂ emissions for various products from U.S. chlor-alkali plants, and we derive the energy uses and emissions for the main plant products and by-product H₂ via the mass allocation method. The LCA was conducted using the chlor-alkali plants’ operation data collected from various databases, as described in Section 2.1. In addition, we conducted sensitivity analysis based on varying the electricity source from a regional mix to renewable electricity, i.e., solar and wind, to observe the impact on emissions associated with the production of chlor-alkali products.

2. Methodology

This section provides details about data sources, intermediate and final product amount, fuel consumption, facility emissions, and allocation methods.

2.1. Data Sources

This study compiles comprehensive inventories of GHG emissions in U.S. chlor-alkali facilities and information about plant operations (electricity generation and on-site usage) and products (including intermediate products). We rigorously matched chlor-alkali plant lists to ensure the same emissions datasets and plant product datasets, based on cross-checks of physical location, history of acquisitions and mergers, capacity, and other related factors.

2.1.1. Chemical Data Reporting (CDR): For Production Quantities (2011 and 2016 Datasets)

The facility-level production data are obtained from the CDR database [8]. The facility-level CO₂ emissions data are obtained from the Greenhouse Gas Reporting Program (GHGRP) database [9]. Additionally, facility-level electricity consumption is gathered from the U.S. Energy Information Administration’s EIA-923 form [10].

The CDR database is used to obtain the net and intermediate production data for each chlor-alkali facility. Since the database is updated every four years, we used CDR 2011 and 2016 data in the present study to account for production quantities at U.S. chlor-alkali facilities. The Pacific Northwest National Laboratory (PNNL) Merchant Hydrogen Plant Capacities dataset compiles hydrogen production capacities in the U.S. merchant market [11]. These data are compiled from company reports and other sources and are assumed in the present study to reflect current conditions.

2.1.2. GHGRP: For CO₂ Emissions from Stationary Combustion and Other Sources

Facilities in the U.S. report annual emissions separately to the GHGRP. Reporting is required for all stationary combustion emissions of CO₂, CH₄, and N₂O from industrial sources under GHGRP Subpart C [9]. Additional reporting is required in some sectors for industry-specific emissions sources. With limited sources of GHGs, however, chlor-alkali facilities are not regulated under any additional subparts. GHG emissions are reported by unit, and include information regarding the type of fuel combusted, if applicable. The chlor-alkali facility GHG emissions are listed in the Supplementary Information, with facility total emissions in Table S1 and facility CO₂ emissions breakdown for combustion of each fuel type in Table S2.

2.1.3. EIA-923: For Electricity Consumption and On-Site Energy Generation, Particularly in Facilities Using CHP

For facilities that use grid electricity, the electricity consumption was estimated by process energy demand, as such data were not publicly available. For facilities that generate electricity and heat on-site by CHP, the fuel consumption and electricity generation data are sourced from the EIA-923 form [10]. In this dataset, facilities report the quantity of fuel consumed for electricity generation as well as total fuel consumption. The difference is assumed to reflect fuel consumption for steam generation. Facilities also quantify the amount of electricity exported or resold if it is not all consumed on-site. In total, seven chlor-alkali facilities in this study were identified through these data as using CHP in one or both data years. However, only three of the CHP facilities had all datasets (e.g., emissions data, electricity data, and CDR data) needed to conduct the analysis for both 2011 and 2016.

2.2. Summary of Analysis Approaches by Steps

2.2.1. Determination of Production and Throughput in Chlor-Alkali Plants

The plant production amount is determined by using CDR-reported data for the facilities that produce Cl_2, NaOH or KOH, and H₂ as final products. Similarly, CDR-reported data are used for the facilities with derivative products (i.e., HCl, NaClO, EDC, and VCM), which report both intermediate and final products.

Otherwise, for facilities that do not report intermediate products, the intermediate amounts of Cl₂, NaOH (or KOH), and H₂ are calculated by element balance and chemical stoichiometry. Some products, such as EDC, are often produced from two different pathways of chlorination and oxychlorination. The approach of identifying the amount of EDC produced from individual pathways is shown in Supplementary Information Section S2.2. The derived amounts of facility intermediate and final products are shown in Tables S7 and S8 in the Supplementary Information.

2.2.2. Estimation of Energy Use

Allocating unit burdens to unit products requires a calculation of unit energy usage. The unit energy usage varies for each plant with variations in operation practices and plant sizes. Given the absence of such data in public information, they are estimated by unit energy intensity $U_I$ (MJ/kg), unit capacity $U_C$ (kg), and capacity factor $C_F$ as shown in Equation (2) and described and validated in a previous study [12].

$$Unitenergyusage=U_I\times U_C\times C_F$$

(2)

Unit energy intensity data are obtained from a literature survey. The product of unit capacity and capacity factor yields actual unit production data, which are estimated from CDR data. The unit energy intensity data used in the present study are listed in

Table 1. Energy intensities for chlor-alkali processes.

Direct Energy Input for the Chlor-Alkali Process			Diaphragm		Membrane		Mercury		Reference
Energy Input (MJ/kg Cl2)			Electricity	Steam	Electricity	Steam	Electricity	Steam
Brine Preparation				0.63		0.63		0.63	Brinkmann et al. [2]
Electrolysis			9.36		8.64		11.70		Brinkmann et al. [2]
Post-Processing	Cl2	Cooling/drying	0.39		0.39		0.39		Schmittinger et al. [13]
		Compression	0.63		0.63		0.63		Schmittinger et al. [13]
		Liquefaction	0.21		0.21		0.21		Brinkmann et al. [2]/Schmittinger et al. [13]
		Evaporation		0.45		0.45		0.45	Brinkmann et al. [2]
		Total for Cl2	1.23	0.45	1.23	0.45	1.23	0.45
	NaOH	Concentration		4.05		1.31		0.00	Brinkmann et al., 2014 [2]
		Cooling	0.27		0.27		0.27		Worrell et al., 2000 [14]
		Total for NaOH	0.27	4.05	0.27	1.31	0.27	0.00
	H2	Cooling	0.00088		0.00088		0.00088		Lee et al., 2018 [1]/O’Brien et al., 2005 [15]
		Compression for pipeline transmission (1.3–20 bar)	0.17598		0.17598		0.17598		Lee et al., 2018 [1]
		Total for H2	0.18	0.00	0.18	0.00	0.18	0.00
Total (MJ/kg Cl2)		Total system	10.93	4.68	10.21	1.94	13.27	0.63
ANL *			10.56	5.23	9.99	1.27			Lee et al., 2018 [1]
ACC 2011 **			9.57	5.19					ACC report 2011 [16]
Worrell 2000			11.3		9.8		12.3		Worrell et al., 2000 [14]

* The units are converted to per kg Cl₂ from per kg H₂ available in the source, on the basis of stoichiometric production shares of brine electrolysis products shown in Equation (1). ** The data are reported as capacity-weighted average of diaphragm (75%) and membrane (25%). The data for individual technologies were not presented in the reference. Most references report the energy usage as electricity and steam. In the present study, the steam is converted to fuel demand by assuming 80% boiler efficiency.

and

Table 2. Energy intensities of various downstream process units in chlor-alkali plants.

Product	Electricity	Fuel Use	Total Energy	Data Source	Note
	MJ/kg	MJ/kg	MJ/kg		--
VCM	--	--	4.88	Brueske et al., 2015 [17]	EDC cracking
	--	2.7		EDC/VCM-Process (N.D.) [18]	--
	0.50	3.193	3.69	Azapagic et al., 2006 [19]	--
	1.34 *	2.947	4.29	Average	--
EDC-Chlorination	0.20	6.424	6.62	Azapagic et al., 2006 [19]	--
	0.38	1.80	2.18	Althaus et al., 2007 [20]	Fuel is not defined
	0.29	4.11	4.40	Average
EDC-Oxychlorination	0.19	7.142	7.33	Azapagic et al., 2006 [19]	--
	--	--	7.92	Brueske et al., 2015 [17]	Oxychlorination
	3.28	5.23	8.51	Feraldi et al., 2011 [16]	Combining oxychlorination and EDC/VMC process
	1.73	6.19	7.92	Average
HCl	--	--	0.41	Brueske et al., 2015 [17]	--
	1.20	0.00	1.20	Althaus et al., 2007 [20]	--
	1.20	0.00	1.20	Average
NaClO	--	--	1.37	Brueske et al., 2015 [17]	--
	0.06	0.17	0.23	Althaus et al., 2007 [20]	--
			3.40	Brueske et al., 2015 [17]	PVC-AMO
	2.49	2.41	4.90	Feraldi et al., 2011 [16]	PVC-ACC

* Calculated as the difference between average total energy and average fuel for steam.

.

By using the unit energy intensity reported in the literature and the derived product amount, the electricity and steam consumption for each unit can be estimated. The results are shown in Tables S12–S15 in the Supplementary Information.

2.2.3. Facility Fuel Consumption

For facilities reporting CO₂ emissions for different fuel categories in the GHGRP, the fuel consumption is estimated by fuel emissions factors (EFs) using Equation (3). Here, $i$ refers to fuel type and the $EF$ (lower heating value basis) for various fuel types is obtained from GREET 2021 as listed in

Table 3. Emissions factors for fuels.

Fuel Type	kg CO2/mmBtu
Assorted Oil	76.22
Coal	93.28
Diesel	73.96
Fuel Gas	59
Natural Gas	53.06
Other	52.07

[21].

$${Fuel=\sum Fuel}_i=\sum {\displaystyle \frac{{C{O}_2}_i\left[\mathrm{k}\mathrm{g}\right]}{{EF}_i}}$$

(3)

2.2.4. Facility Electricity Consumption

CHP plants generate electricity and steam on-site. The EIA-923 form [10] reports the amount of total electricity generated, on-site electricity usage, and exported electricity. For CHP facilities, the exported electricity is not accounted for in the facility electricity use but receives a displacement credit. Non-CHP facilities purchase their electricity for facility use, with the amounts estimated by using product-specific process energy demand data.

The facility electricity demand can also be estimated using product energy intensity (provided in Table 1 and Table 2) and the production amount (covered in Section 2.2.1). This approach can further be used to estimate the facility-level energy intensity of each product. Such information is used to aid the allocation of reported facility energy use and CO₂ emissions to the facility final products based on unit-level energy consumption. The method shown in Equations (4) and (5) has been described in previous studies [12].

$${\displaystyle \frac{{Emission}_{uniti}}{\sum _i{Emission}_{uniti}}}={\displaystyle \frac{{Energy}_{uniti}}{\sum _i{Energy}_{uniti}}}$$

(4)

$${\displaystyle \frac{{Emission}_{producti}}{\sum _i{Emission}_{producti}}}={\displaystyle \frac{{Energy}_{producti}}{\sum _i{Energy}_{producti}}}$$

(5)

The two pathways of EDC production have different energy demands, and the energy demand for each pathway is calculated on the basis of the product amount from each pathway.

2.2.5. Additional CO₂ Emissions Data

For facilities with CO₂ emissions records, the reported emissions are used. For plants without GHGRP records, Equation (3) is used to estimate the CO₂ emissions, using fuel use records from EIA-923 (assuming the fuel type is NG). For non-CHP facilities, CO₂ emissions associated with electricity production are not facility-level and are not reported in the GHGRP. Therefore, CO₂ emissions data for electricity use were obtained using respective regional grid CO₂ intensity from GREET 2021 [21].

All the key information (i.e., amounts of intermediate and final products, unit-level energy usage, facility total energy consumption [fuel and electricity], and total CO₂ emissions) needed to allocate energy use and emissions to individual products is covered in Section 2.2. The allocation method is summarized in Section 2.3.

2.3. Allocation Method

2.3.1. Emissions Burdens Allocated to All Facility Products by Mass Allocation

At each unit level, the energy use or emissions burdens are allocated to unit products by the mass share. It is important to note, however, that other coproducts allocation methods can be utilized within life cycle analysis, which may result in different emissions associated with hydrogen production from chlor-alkali plants.

$$E_{Producti}\left({\displaystyle \frac{\mathrm{g}}{\mathrm{k}\mathrm{g}}}or{\displaystyle \frac{\mathrm{m}\mathrm{g}}{\mathrm{k}\mathrm{g}}}\right)={\displaystyle \frac{{Emission}_{unitprocess}\times {MassAF}_i}{{Mass}_i}}$$

(6)

$${MassAF}_i={\displaystyle \frac{{Mass}_i}{\sum {Mass}_i}}$$

(7)

2.3.2. Electricity Export Displacement Credits

Given the large electricity demand for the brine electrolysis process, some large chlor-alkali plants generate electricity on-site via the CHP process instead of purchasing grid electricity. Some plants generate excessive electricity for export sales after meeting their on-site electricity demand. For the exported electricity, plants may receive electricity export energy and emissions credits, which are subtracted from total energy use and total emissions of the facility, respectively, as a result of displacing grid electricity. The CO₂ emissions for the U.S. national grid mix and the regional grid mix are shown in

Table 4. CO₂ emissions for U.S. national grid mix and regional grid mix from GREET 2021 [21].

NERC Region *	ASCC	FRCC	HICC	MRO	NPCC	RFC	SERC	SPP	TRE	WECC	U.S. Mix
CO2 Emissions (gCO2/kWh)	564.3	447.2	884.2	587.3	231.5	403.9	384.6	475.3	385.4	330.9	410.1

* NERC: North American Electric Reliability Corporation, ASCC: Alaska Systems Coordinating Council, FRCC: Florida Reliability Coordinating Council, HICC: Hawaiian Islands Coordinating Council, MRO: Midwest Reliability Organization, NPCC: Northeast Power Coordinating Council, RFC: Reliability First Corporation, SERC: Southeastern Electric Reliability Council, SPP: Southwest Power Pool, TRE: Texas Reliability Entity, WECC: Western Electricity Coordinating Council.

, from GREET 2021 [21].

2.3.3. Facility Emissions Burdens Allocated to All Facility Products on the Basis of Unit-Level Energy Usage

In this unit-level allocation approach, the combustion emissions, reported under GHGRP Subpart C, are allocated to each process unit on the basis of the unit energy usage. The unit energy usage estimation method is shown in Equation (2). Subsequently, the obtained unit emissions are allocated to unit products on a mass basis. Intermediate products carry their upstream unit burdens to downstream unit processes, ensuring accurate attribution of emissions throughout the production processes. The method of unit burden allocation to unit products is shown in Figure 2.

The unit products (e.g., from Unit A) could have final products (e.g., P1) and intermediate products (e.g., I1) that are consumed in the unit downstream process (e.g., Unit B) for further conversion or combustion for energy supply. After Unit A, the intermediate Stream I1 will carry the allocated burden from Unit A (from using Fuel A and the upstream burden of producing Feedstock A) to Unit B. Subsequently, the inherited burden of I1, along with the Unit B burden from using Fuel B, as well as the upstream burden of Feedstock B, will be allocated to all Unit B products. The allocation calculation continues until all the burdens are allocated to all final products (P1, P2, P3, and P4).

3. Results and Discussion

3.1. Amounts of Chlor-Alkali Plant Products

Several chlor-alkali plants were identified, as discussed in the Methodology section. Please refer to Tables S3 and S4 in the Supplementary Information, which provide production (CDR database) data for intermediate and final products for the years 2011 and 2016, respectively. Although each database has an extensive collection of chlor-alkali plants, only seven facilities had sufficient data to conduct LCA. Additionally, utilizing the data for production years 2011 and 2016 expands the dataset to 14. The facility type, technology, and grid region are provided in

Table 5. Details for the chlor-alkali facilities under analysis.

Facility ID *	Type	Technology	NERC Region
C_1	CHP	Diaphragm	SERC
C_2		Membrane	TRE
C_3		Mercury	RFC
NC_1	Non-CHP	Membrane	SERC
NC_2		Membrane	SERC
NC_3		Membrane	SERC
NC_4		Membrane	MRO

* C represents CHP, NC represents Non-CHP.

.

For the chlor-alkali plants with CDR records, the production data are shown in Tables S3 and S4 for the year 2011 and 2016, respectively. Additionally, detailed product information regarding intermediate products being used in specific process streams is estimated and provided in Tables S7 and S8 in the Supplementary Information, for the year 2011 and 2016, respectively. These data can be used to allocate upstream burdens to downstream processes for appropriate attribution of emissions. Data for these products are used as unit throughput values for each process unit, and as such, the process unit energy consumption can be estimated by using energy intensity as described in Section 2.2.2.

In the present study, for the listed facilities, no hydrogen export record was found in publicly available information. From consultation with industry experts, we found that facilities either vent or export the by-product hydrogen and rarely combust it. The difference in practice has implications for energy use and emissions accounting. For the facilities that vent hydrogen, other products carry the energy use and emissions that would be carried by hydrogen if exported. Additionally, hydrogen is considered an indirect GHG with a global warming potential (GWP) that can further result in higher facility emissions impact if vented. For facilities that export hydrogen, hydrogen shares energy use and emissions burdens with other products and its indirect GWP impact is excluded. We note that recovering and compressing hydrogen for export require energy use and result in associated emissions.

In this study, we investigate the scenario that all facilities exported H₂. There were two facilities (C_3 and NC_4) that vent hydrogen, and their venting scenario results are shown in Supplementary Information Section S6.2. To compute the H₂ export scenario, we added an electricity burden for compression of H₂ from 1.3 bar (production pressure) to 20 bar (export pressure). The source of this electricity is considered to be the regional grid for non-CHP plants. For CHP plants, the electricity is supplied by cogeneration.

3.2. Upstream Burden for Salt Mining and Transportation

In this section, the energy use and emissions associated with brine recovery and transportation are estimated. We used the default values of energy use and emissions associated with brine recovery from GREET 2021 [21]. We consider the transportation contribution to be small and consequently its influence on overall WTG results as negligible.

Table 6. Energy use for salt mining and transportation.

	MJ/kg Salt	Electricity	Fuel	NG	Coal	Distillate	Diesel	Residual Oil
ACC [16]	Salt recovery	1.296	1.8	1.04	0.31	0.45	--	--
	Salt transportation	0.059	--	--	--	--	0.01137	0.0013
GREET 2021 [21]	Brine recovery	0.258	--	0.787	--	--	--	0.129

shows the energy use for brine recovery; the emissions associated with brine recovery are 90 gCO₂/kg brine. The emissions burden increases by 7 gCO₂/kg brine (~8%) if 50 miles of brine transportation is considered. The transportation burden is estimated using default GREET 2021 heavy-duty trucks running on diesel with 23-ton payloads and 7.2 mpg fuel economy from origin to destination and 9 mpg fuel economy for the empty return trip. In Section 3.5, we show that the contribution of brine emissions to the WTG emissions is approximately 10%. Therefore, the impact of transportation on the WTG emissions is less than 1% (0.08 × 0.1= 0.8%). If the distance were to be an order of magnitude higher, the impact would be substantial enough to be included. In this analysis, we have ignored the transportation of brine, since its impact on WTG emissions is minor.

3.3. Electricity Export Energy and Emissions Credits

Among the seven chlor-alkali plants studied, four plants use grid electricity and only combust fuels for steam generation. In contrast, the other three plants generate electricity on-site via the CHP process by combusting fuels, leading to much larger amounts of on-site fuel use and thus, air emissions. In addition, the CHP plants export electricity, for which they receive a displacement energy and emissions credit. The electricity export information is shown in Table S5. For each plant, the electricity displacement credit (energy use, CO₂ emissions) is based on the regional grid mix, which is based on the plant’s geographic location. Figure 3 and Figure 4 show the electricity export energy use and CO₂ emissions credits, respectively, for the CHP facilities for the years 2011 and 2016.

The electricity and emission credits for C_3 are higher compared to C_1 and C_2, as shown in Figure 3 and Figure 4, respectively, owing to its higher electricity export, i.e., approximately 1.5 times higher than C_1 and approximately 2.5 times higher than C_2, as shown in Table S5. Additionally, the production capacity of C_3 is approximately 10 times and 18.5 times smaller than C_1 and C_2, respectively, as shown in Table S5. Higher electricity export and lower capacity for C_3 result in significantly higher allocated electricity and emission credits.

3.4. Energy Consumption of Chlor-Alkali Plants Allocated to Plant Final Products

By using the energy consumption shares of facility final products in Tables S16 and S17, the chlor-alkali facility’s total electricity consumption and fuel consumption can be allocated to facility final products.

For non-CHP plants, the electricity consumption for final products is shown in

Table 7. The non-CHP chlor-alkali facilities’ on-site electricity consumption allocated to final products.

Facility ID	Year	Products’ Electricity Use (MJ/kg)
		Cl2	NaOH	H2	KOH	HCl	NaClO	EDC	VCM
NC_1	2011	5.4	4.4	10.4		6.5			3.5
NC_2		5.4	4.5	10.4
NC_3		5.1	4.6	10.1	3.6	6.3	5.5
NC_4		5.4	4.4	10.4		6.5
NC_1	2016		4.4	10.4		6.5			3.5
NC_2		5.4	4.4	10.4
NC_3		5.1	4.4	10.1	3.6	6.2	5.5
NC_4		5.0	4.4	10.0	3.6	6.2

. For these plants, fuels are only combusted on-site to generate steam as electricity is imported. Thus, the on-site energy use is allocated to the final product on the basis of product steam demand; the results are listed in

Table 8. The non-CHP chlor-alkali facilities’ on-site fuel consumption allocated to final products.

Facility ID	Year	Products’ Fuel Use (MJ/kg)
		Cl2	NaOH	H2	KOH	HCl	NaClO	EDC	VCM
NC_1	2011	1.6	3.2	0.6		1.6			12.4
NC_2		1.1	2.3	0.5
NC_3		1.2	2.6	0.5	2.4	1.9	2.3
NC_4		1.9	3.8	0.8		1.8
NC_1	2016		2.8	0.6		1.4			10.7
NC_2		1.1	2.3	0.5
NC_3		1.1	2.6	0.4	2.3	1.1	2.2
NC_4		0.9	1.9	0.3	1.8	0.9

. The breakdown of energy consumption (NG, H₂, coal, etc.) is listed in Tables S9 and S10. Total energy consumption allocated to each product can be calculated by adding the allocated electricity use from Table 7 and allocated fuel use from Table 8.

Table 7 shows that the allocated electricity use is similar among all non-CHP facilities, while Table 8 shows the difference in allocated fuel use between non-CHP facilities. The consistent electricity use can be attributed to membrane technology for all non-CHP plants, leading to similar electricity consumption. The small difference in fuel use can be attributed to differences in type of fuel, steam generation efficiency, and thermal energy utilization in the non-CHP facilities. For CHP facilities with electricity export, the energy use displacement credit is accounted for.

Table 9. The CHP chlor-alkali facilities’ energy consumption allocated to individual products.

Facility ID	Year	Products’ Fuel Use (MJ/kg)
		Cl2	NaOH	H2	KOH	HCl	NaClO	EDC	VCM
C_1	2011		22.4	25.9		18.2		12.7	25.3
C_2			23.2	39.5		27.4	26.7	26.0	49.3
C_3		25.6		42.6	19.7	30.0
C_1	2016		21.7	25.0		17.6		12.3	24.5
C_2			21.7	37.2		25.9	25.1	23.5	44.8
C_3		27.3		45.4	21.0	32.0

shows the energy consumption allocated to individual products for the CHP plants.

The difference in fuel use between C_1, C_2, and C_3 in Table 9 can be attributed to the type of technology. C_1 is based on diaphragm technology and therefore requires less energy consumption. However, it produces low-quality NaOH, therefore requiring additional energy for concentration. C_2 is based on membrane technology, which requires relatively higher energy consumption than diaphragm technology. C_3 is based on mercury technology, which requires the highest energy consumption.

3.5. Chlor-Alkali Plant CO₂ Emissions Allocated to Plant Final Products

The on-site CO₂ and criteria air pollutants (CAPs) emissions of these seven facilities are listed in Table S1 in the Supplementary Information. In the GHGRP database, the combustion CO₂ emissions for each fuel are reported, allowing the estimation of fuel consumption by using the EFs recommended by the U.S. Environmental Protection Agency (shown in Equation (3)). The derived fuel consumption data for each facility are listed in Tables S9 and S10 in the Supplementary Information.

3.5.1. Emissions Burdens Allocated to Non-CHP Plants

The emissions are allocated to the products by the method described in Section 2.3.3. Therefore, the total reported facility burden is fractioned to only account for the part corresponding to the energy used solely to produce the main products (electrolyzer and brine treatment, and post-treatment for hydrogen, chlorine, and caustic). For non-CHP plants, the results are shown for two cases, i.e., using the regional grid mix, as shown in Figure 5 and

Table 10. WTG emissions burden allocated to non-CHP plants using regional grid electricity.

Facility ID	Year	NERC Region *	CO2 Emissions (kg/kg)
			Cl2	NaOH	H2	KOH	HCl	NaClO	EDC	VCM
NC_1	2011	SERC	0.8	0.8	1.3		0.9			1.8
NC_2		SERC	0.7	0.7	1.3
NC_3		SERC	0.7	0.7	1.3	0.6	0.8	0.8
NC_4		MRO	1.0	1.0	1.9		1.3
NC_1	2016	SERC		0.8	1.4		0.9			1.7
NC_2		SERC	0.7	0.7	1.3
NC_3		SERC	0.7	0.7	1.3	0.6	0.8	0.8
NC_4		MRO	1.0	0.9	1.8	0.7	1.2

* The CO₂ intensities for the SERC and MRO mixes are 0.38 kgCO₂/kWh and 0.59 kgCO₂/kWh, respectively, from Table 4.

, and using renewable electricity, as shown in

Table 11. WTG emissions burden allocated to non-CHP plants using renewable electricity *.

Facility ID	Year	CO2 Emissions (kg/kg)
		Cl2	NaOH	H2	KOH	HCl	NaClO	EDC	VCM
NC_1	2011	0.20	0.30	0.21		0.22			1.4
NC_2		0.15	0.20	0.20
NC_3		0.15	0.21	0.19	0.18	0.17	0.21
NC_4		0.17	0.23	0.20		0.19
NC_1	2016		0.28	0.28		0.24			1.3
NC_2		0.15	0.19	0.20
NC_3		0.15	0.21	0.20	0.18	0.17	0.21
NC_4		0.13	0.17	0.18	0.16	0.15

* The CO₂ intensity for renewable electricity (e.g., wind, solar) is considered to be 0 kgCO₂/kWh, without accounting for embodied emissions.

, to assess the impact of electricity on the emissions.

In Table 10 and Figure 5, the emissions for NC_4 are higher than others, although NC_4 does not have significantly different individual electricity use relative to other facilities, as shown in Table 7. The difference in emissions is due to the higher CI of the regional electricity (MRO) in the case of NC_4.

From this analysis, the by-product H₂ has a CO₂ intensity of 1.3–1.9 kg CO₂/kg H₂. Since electricity has a significant impact on the CO₂ emissions of by-product chlor-alkali H₂, it is important to investigate the decarbonization potential of the process by using cleaner electricity. Therefore, we carried out a study to estimate the impact of using renewable electricity on the emissions allocated to the individual products, as shown in Table 11.

In Table 11, H₂ and HCl experience the most reduction in emissions because these products are electricity-intensive and have lower fuel use burdens. In the case of H₂, the emissions are upstream burdens associated with brine preparation and purification, since hydrogen is a by-product of electrolysis and then becomes further compressed using electricity. The associated fuel use burden comes from feedstock brine only. Using renewable electricity, the by-product hydrogen has average CO₂ emissions of 0.21 kg CO₂/kg H₂ from the facilities. Owing to the scarcity of CH₄ and N₂O data, it is challenging to determine the CI of H₂. However, a reasonable estimate can be made from the upstream CI of on-site NG use and the CI of brine to calculate the CI of H₂. This analysis results in a CI value for by-product hydrogen (using renewable electricity) of 0.22 kgCO₂e/kg H₂.

Production of VCM involves primarily fuel use rather than electricity use, and consequently, the lowest impact of electricity CI is observed. Finally, the key conclusion is that most of the emissions contribution for non-CHP facilities comes from electricity. The CI of electricity can vary significantly by region. Therefore, the emissions from non-CHP facilities are highly sensitive to region of operation. We performed a sensitivity analysis considering regional electricity mixes and potential use of renewable electricity to assess the impact on GHG emissions. If the facilities use renewable electricity or other clean power sources, there is a significant potential to reduce GHG emissions associated with Cl₂, NaOH, and H₂ production.

3.5.2. Emissions Burdens Allocated to CHP Plants

Allocated emissions for the individual products of the CHP plants are shown in

Table 12. WTG emissions burden allocated to CHP plants.

Facility ID	Year	NERC Region *	CO2 Emissions (kg/kg)
			Cl2	NaOH	H2	KOH	HCl	NaClO	EDC	VCM
C_1	2011	SERC		1.32	1.52		1.07		0.75	1.73
C_2		TRE		1.40	2.38		1.65	1.61	1.61	3.15
C_3		RFC	1.34		2.23	1.03	1.57
C_1	2016	SERC		1.29	1.49		1.05		0.73	1.69
C_2		TRE		1.30	2.24		1.55	1.51	1.46	2.87
C_3		RFC	1.42		2.35	1.09	1.66

* The CO₂ intensities for SERC, TRE, and RFC mixes from Table 4 are 0.38 kgCO₂/kWh, 0.39 kgCO₂/kWh, and 0.40 kgCO₂/kWh, respectively.

and Figure 6. Since some of the CHP plants export electricity, CO₂ emissions displacement credit is applied to each facility, depending on the grid region.

For CHP plants, the major contributor to the WTG emissions is on-site fuel use. Since the steam and electricity are cogenerated, the impacts of both are equally responsible. Additionally, the fuel use depends highly on exported electricity, as seen in case C_3, which experiences the highest emissions from on-site fuel use. Figure 6 also indicates that a higher electricity credit can be obtained by displacing higher-CI grid electricity, thereby further reducing WTG emissions for all products.

3.6. Energy Use and Emissions Comparison with the Literature

Chlor-alkali process emissions data are scarce in the literature. Additionally, there are various factors that can lead to differences in emissions, such as type of technology, cogeneration efficiency in CHP plants, regional grid CI, and upstream activities and transportation involved in procuring brine.

Table 13. Energy use ccomparison to reported data for chlor-alkali products.

Source	Year	Products’ Energy Use (MJ/kg)
		Cl2	NaOH	H2	KOH	HCl	NaClO	EDC	VCM
Euro-Chlor [22]	2013	19.90	18.10	15.70			39.60
PlasticsEurope [23]	2015								54.70
GaBi * [24]		21.3	21.3
GaBi ** [24]		18.8	18.8
USLCI [25]	2012			16.36
Ecoinvent *** [26]		4.9–5.9	5.0–6.0
Results: Non-CHP ****		5.2	7.1	10.8	5.2	7.8	7.8		15.1
Results: CHP ****		26.5	22.3	34.9	20.4	24.3	25.8	21.1	40.5

* Diaphragm technology, ** membrane technology, *** 2.97–3.58 kWh/kg Cl₂, **** weighted average based on production capacity including both years 2011 and 2016.

and

Table 14. Emissions Ccomparison with reported data for chlor-alkali products.

Source	Year	CO2 Emissions (kg/kg)
		Cl2	NaOH	H2	KOH	HCl	NaClO	EDC	VCM
Euro Chlor * [22]	2013	0.90	0.86	1.14			0.93
PlasticsEurope [23]	2015								1.71
GaBi ** [24]		1.16	1.16
GaBi *** [24]		0.97	0.97
USLCI [25]				0.92
Ecoinvent *** [26]		0.6–0.7	0.7–0.8
Results: Non-CHP ****		0.6	0.7	1.4	0.5	0.9	0.8		1.7
Results: CHP ****		1.4	1.3	2.1	1.1	1.4	1.6	1.3	2.6

* GHG emissions (kgCO₂e/kg), ** diaphragm technology, *** emissions estimated assuming U.S. mix, **** membrane technology.

show a comparison between the literature data and this study with respect to energy use and emissions associated with chlor-alkali products.

The data reported by Euro-Chlor and PlasticsEurope are only compared to CHP facilities’ energy use data in this study because the allocated energy use reported by Euro-Chlor and PlasticsEurope includes the fuel required both for generation of electricity and for thermal energy used in CHP plants. Therefore, it is reasonable to compare the reported energy use data of Euro-Chlor and PlasticsEurope to CHP facilities’ data in this study.

The energy use associated with the production of the main chlor-alkali products (Cl₂ and NaOH) is fairly consistent in the literature. Euro-Chlor and PlasticsEurope considered data from 50 participating facilities (EU27 + NO + CH) in member countries. The energy use of CHP facilities for Cl₂ and NaOH production in the present study is higher compared to Euro-Chlor data. The difference can be attributed to differences in both the technology share of the 50 participating facilities in the Euro-Chlor report and the cogeneration efficiency of CHP plants compared between Euro-Chlor and this study. It is reasonable to note that chlor-alkali technology in Europe is dominated by membrane technology (approximately 66%), which is the most efficient compared to diaphragm and mercury technology.

The energy use for hydrogen shows a significant difference between the Euro-Chlor data and the CHP results, mainly due to the energy used for compression of hydrogen. Our study considers hydrogen compression from 1.3 bar to 20 bar. Data reported in the literature do not include hydrogen compression to 20 bar. Compression leads to a 4.9 MJ/kg increase in the electricity use for hydrogen. Depending on the cogeneration efficiency (25–33%) of the CHP facility, the fuel use to provide 4.9 MJ/kg for hydrogen compression can be 15–20 MJ/kg higher, which eliminates the difference between our results and the literature data.

GaBi data account for energy use up to the electrolysis step, without considering post-processing for diaphragm and membrane technologies. In contrast, the CHP facilities in the present study include post-processing and have additional products, which slightly reduce plant energy efficiencies. As a result, the reported energy use in GaBi is slightly lower than that of the CHP facilities in the present study.

Ecoinvent data [26] only reported electricity used (for electrolysis) to produce Cl₂ and NaOH; therefore, it is fair to compare this value to the non-CHP facility energy use in our analysis because Cl₂ and NaOH production energy use is primarily for electricity. The range of Ecoinvent data [26] corresponds to the energy use range in membrane, diaphragm, and mercury cell technology. The energy use allocated to individual products for non-CHP facilities is in reasonable agreement with the reported data.

Comparing the results of Euro-Chlor, PlasticsEurope, and Gabi to CHP facilities emissions, we observe that the emissions for Cl₂ and NaOH are slightly higher than the reported emissions. The discrepancy can be attributed to the difference in cogeneration efficiency of CHP facilities under consideration, together with the fact that the Euro-Chlor and PlasticsEurope data are dominated by membrane technology, which has lower emissions. Hydrogen has twice the emissions, owing to additional compression energy included in our analysis, as discussed earlier. Ecoinvent data [26] include electricity use for producing chlorine and sodium hydroxide and are therefore compared with the non-CHP facility results. Similar to the energy use comparison between Ecoinvent data and non-CHP results, the emissions reported and estimated are also in close agreement.

3.7. Hydrogen Potential from Chlor-Alkali Plants

With the increase in interest in hydrogen for the decarbonization of the industrial, transportation, power, and residential sectors, chlor-alkali plants have the opportunity to export their by-product hydrogen. As observed in the previous section, hydrogen from chlor-alkali plants has CO₂ emissions of 1.25–1.9 kgCO₂/kg H₂ for non-CHP facilities and 1.5–2.4 kgCO₂/kg H₂ for CHP facilities. Additionally, if the non-CHP facilities utilize renewable electricity, the emissions range from 0.18 to 0.28 kgCO₂/kg H₂. Therefore, there are opportunities and economic incentives for chlor-alkali plants to export low-carbon hydrogen instead of venting it. Chlor-alkali plants may become hydrogen suppliers to support the regional hydrogen hubs in the U.S. Additionally, their integration into the existing supply chain can improve regional hydrogen availability and market efficiency to support global hydrogen growth trends. Figure 7 shows the various chlor-alkali facilities in the U.S. and their potential hydrogen production capacity (totaling 320 kilotons or ~3% of current total annual demand).

In Figure 7, regional grid intensities are provided to suggest that hydrogen from non-CHP plants from various grid regions may have significantly higher CI compared to hydrogen from renewable electricity or from a cleaner grid region. However, CHP plants located in the higher-CI grid region can take advantage of receiving credits by exporting electricity and displacing higher-CI grid electricity.

4. Conclusions

In this study, the energy use and CO₂ emissions of various products from chlor-alkali plants in the U.S. are investigated. Chlor-alkali plant production data were obtained from the CDR database. The CO₂ emissions data were gathered from the GHGRP and National Emissions Inventory databases, respectively. Electricity generation use or export data were obtained from the EIA-923 form. Seven chlor-alkali facilities (three CHP, four non-CHP) had complete datasets with which to carry out the analysis. Additionally, we analyzed the scenario in which all facilities exported hydrogen. Furthermore, we investigated the impact of the electricity grid by considering the regional grid mix and potential use of renewable sources, i.e., solar and wind electricity.

Analysis showed that non-CHP facilities are electricity-intensive and the major contribution to the GHG emissions was from the electricity supply chain. Emissions significantly decreased with the choice of renewable electricity. Additionally, use of cleaner electricity produced low-carbon Cl₂, NaOH, and H₂, which can be used to decarbonize the downstream processes that use these chemicals. We observed that CHP facilities benefit more if they are located in higher-CI grid regions. Through the export of surplus electricity cogenerated in a CHP facility, an emission credit can be obtained by displacing high-CI local grid electricity. A comparison of energy use and emissions with the literature values supported our study. The differences were attributed to using a different system boundary (i.e., considering hydrogen compression) or to differences in the regional electricity, the technology employed (e.g., membrane, diaphragm, mercury), and/or cogeneration efficiency. The product hydrogen export potential of chlor-alkali plants is approximately 320 kilotons per year in the U.S. This value represents approximately 3% of current annual U.S. hydrogen production.