Packaged Bread and Its Carbon Footprint: Balancing Convenience and Waste

Abstract

The growing market for pre-sliced and packaged bread, driven by convenience and extended shelf life, raises environmental concerns due to its reliance on single-use polyethylene (PE) bags. To evaluate this trade-off, a cradle-to-distribution-center Life Cycle Assessment (LCA) of white sliced bread in 4-slice modified atmosphere PE bags was conducted, following ISO 14040/14044 guidelines and using 2021–2022 factory data from Southern Italy. The initial carbon footprint (CF) of the packaged bread was estimated at 2.77 kg CO_2e/kg when using 100% Grid Electricity. The transformation phase was the largest contributor (41.5%), with electricity accounting for over 90% of this impact, followed by packaging (22.3%) and ingredients (19.4%). Allocation of by-products reduced the CF to around 2.68 kg CO_2e/kg, while the adoption of on-site renewable electricity significantly lowered impacts by up to 30% (to 1.95 kg CO_2e/kg). A key finding is the environmental trade-off between the product and its packaging: a wasted bread slice embodies approximately 70 g CO₂, whereas the production of the corresponding portion of the PE bag emits only about 5 g CO₂. This finding, which is confirmed to be statistically significant, demonstrates that the packaging’s footprint is substantially smaller than the potential impact of even a single wasted slice, proving its crucial role in preventing a larger environmental burden from food waste.

1. Introduction

The global bread market is a major part of the food industry, with consumption varying significantly by country due to different cultural traditions, food preferences, and economic factors. For example, the highest annual per capita bread consumption (88 kg) is in Romania [1]. This is followed by Germany with 80 kg, the Netherlands with 56 kg, Poland with 52 kg, Spain with 47 kgs, France with 44 kg, and Italy with 41 kg [2]. In the United Kingdom and the United States, sliced bread is a staple for many meals, with an annual per capita consumption of about 43 kg [1] and 24 kg [3], respectively.

The bread market is segmented by product type (e.g., loaves, baguettes, rolls, hamburger buns, sandwich slices, ciabatta, frozen bread, and other product types), distribution channel (e.g., convenience stores, specialty retailers, supermarkets and hypermarkets, online retail, various stores, and other distribution channels), and geography. This market is projected to have a Compound Annual Growth Rate (CAGR) of 3.55% over the next five years [4], driven by a rising demand for diverse, healthier options like whole grain and organic breads, and the expansion of e-commerce.

A key driver of this growth is the increasing popularity of pre-sliced and packaged bread. While this format offers convenience and a longer shelf life—which can help reduce household food waste — it also introduces a significant environmental challenge related to packaging. The global packaged bread market, valued at US$ 335 billion in 2023, is expected to grow at a CAGR of 2.6% from 2024 to 2030 [5]. The convenience of this product is at odds with its environmental footprint, primarily due to the use of single-use plastics and the energy-intensive processing required for slicing and packaging.

This study aims to evaluate the cradle-to-distribution center carbon footprint of pre-sliced bread packaged in polyethylene (PE) bags and to discuss strategies for impact mitigation, including the trade-off between packaging waste and the prevention of food waste.

2. Carbon Footprint of Bread and Packaged Sliced Bread: Analysis and Comparisons

Bread is generally considered a low-environmental-impact food, a view supported by numerous studies. Since 1999, various analyses of the Carbon Footprint (CF) of bread have reported values ranging from 0.256 to 2.3 kg CO_2e/kg [6,7]. An analysis by Notarnicola et al. [8] on the environmental sustainability of 21 types of bread consumed in the European Union found an even wider range, from 0.5 to 6.6 kg CO_2e/kg.

The CF of Barilla’s Pan Bauletto bianco, reported as 1.206 kg CO_2e/kg in the Barilla Product Environmental Declaration [9], also falls within this established range. CF estimates vary depending on the production scale and product type. For example, Andersson and Ohlsson [10] estimated a CF of 0.94–1.0 kg CO_2e/kg for industrial production and 0.63–0.64 kg CO_2e/kg for domestic production, considering cultivation, processing, packaging, and final consumption. Moudrý et al. [11] calculated a CF of 1.05–1.14 kg CO_2e/kg, considering conventional and organic agricultural systems but excluding transport. Espinoza-Orias et al. [7] estimated a CF of 1.22–1.55 kg CO_2e/kg for an 800-g loaf. Other authors have reported lower values: 0.625 kg CO_2e/kg [12]; 0.813 kg CO_2e/kg for French bread [13]; 0.950 kg CO_2e/kg [14]; 0.731 kg CO_2e/kg for rye bread [6].

The variations in bread CF values depend on multiple factors, such as the methodological assumptions adopted (e.g., databases, LCA methodology, emission factors); the system boundaries (e.g., inclusion or exclusion of agricultural, distribution, and consumption phases); the production scale (industrial, artisanal, or domestic); the type of energy used; and the climatic conditions in the geographical study areas.

Moresi [15] studied the environmental impact of the industrial production of 1 kg of soft-wheat bread packaged in polyethylene, using the PAS 2050 standard [16], SimaPro 7.2 software, and the 1997 Intergovernmental Panel on Climate Change (IPCC) emission factors [17]. The study estimated a CF of 0.75 kg CO_2e/kg, identifying the agricultural phase as the main emissions hotspot (47.8% of total emissions, primarily due to N₂O emissions from fertilized soils). This was followed by transport (57.2% of emissions, related to the movement of cereals from France, Austria, Hungary, and Germany) and the industrial processing.

Among the most recent assessments, an LCA of Galician bread [18] also demonstrated that agriculture is the dominant contributor to its carbon footprint, mainly due to N₂O soil emissions and diesel use in field operations, whereas baking contributes much less (approximately 0.25 kg CO_2e per kg of bread). Thus, agricultural practices play a crucial role in emission mitigation. Crosnier et al. [19] reached similar conclusions using a True Cost Accounting approach, showing that the hidden impacts of bread are overwhelmingly generated at the farming stage and that sustainability comparisons depend strongly on the functional unit (per hectare vs. per kilogram). Nevertheless, bread waste resulted in amplifying consumption-based impacts by up to 80%, because more than half of Swiss bread production is discarded. Consequently, reducing waste represents one of the most effective levers for improving bread system sustainability.

Considering a functional unit of 1 kg of bread, the Cumulative Energy Demand (CED) was found to vary from 9.1 MJ/kg to 32.9 MJ/kg (see Figure 3 of Ref. [8]). For the “average” bread in the European Union, this indicator was calculated to be 16.1 MJ/kg [20]. Other studies provide further context: Kulak et al. [21] found that the non-renewable resource indicator varied between 6 and 21.5 MJ/kg when considering the life cycle of different types of bread from alternative food networks. Andersson and Ohlsson [10] reported primary energy values from 13.5 MJ/kg for industrial bakeries to 6.5 MJ/kg for local bakeries. Braschkat et al. [22], in their LCA of various baking, milling, and agricultural practices, showed an energy demand varying from 4 MJ/kg for industrially baked bread to 6 MJ/kg for bread baked in small bakeries (excluding transport).

The CF of packaged sliced bread is expected to increase compared to unpackaged or simply wrapped bread due to several factors:

-

Packaging Materials: Sliced bread requires more complex, multi-layered packaging (plastic bags, transport cartons, labels, etc.), significantly increasing its environmental impact.

-

Additional Processing: Pre-slicing and hermetic packaging require extra industrial steps with higher energy consumption compared to traditional bread.

-

Ingredients and Additives: Sliced bread often contains emulsifiers, enzymes, and preservatives to maintain softness and extend shelf life, the production of which contributes to the overall CF.

-

Waste Amplification: While sliced bread has a longer shelf life, its packaging and portioning can generate more waste (though the overall impact depends on domestic consumption behavior vs. the waste reduction from extended shelf life).

Given these unique contributions, particularly from packaging and additional processing, a comprehensive and systematic LCA approach is required to accurately determine the product’s carbon footprint, as detailed in the following section.

3. Methodology

This work conformed to the life cycle assessment (LCA) procedure described by ISO norms 14040 [23] and 14044 [24] according to the guidelines established by the Publicly Available Specification (PAS) 2050 standard method [16].

3.1. Defining Objectives and Scope

This study assessed the carbon footprint (CF) from the cradle to the distribution centers for sliced white sandwich bread in modified atmosphere polyethylene (PE) bags. The focus was on 4-slice packs, with a total weight of 143 g. The selected Functional Unit (FU) was 1 kg of this conventional white sandwich bread in the specified 4-slice packs (see Figure S1 in the electronic supplement).

Figure 1 illustrates the system boundary for this study. Three different life cycle processes were included. More specifically, the upstream processes cover:

(U1)

The production of raw materials, auxiliary materials, and ingredients.

(U2)

The production of packaging materials.

(U3)

The transportation of raw materials, auxiliary materials, and packaging from their production sites (PS) to the company gate (FG).

The core processes involve:

(C1)

The preparation of sourdough, kneading, leavening, baking, refrigeration, slicing, and packaging of the sliced bread.

(C2)

The use of electricity and mains gas.

Finally, the following downstream processes were included:

(D1)

The end-of-life of waste and by-products generated during the production and packaging of sliced bread, as well as the disposal of wastewater.

3.1.1. Exclusions from the System Boundaries

The system boundary for this analysis excluded the production and disposal of capital goods (e.g., mixers, ovens, cold storage units, etc.), as well as personnel travel, in line with Sections 6.4.4 and 6.5 of the PAS 2050 standard method [16]. The potential environmental benefits from sliced bread by-products used as livestock feed were omitted. Following the principle of economic allocation, as the by-products were provided free of cost (zero economic value), all associated GHG emissions were entirely allocated to the main product.

3.1.2. Geographical, Temporal, and Technological Boundaries

This LCA study analyzed the production process of sliced sandwich bread at the reference bakery (e.g., Valle Fiorita Srl, Ostuni, Italy) in the years 2021 and 2022. The process technology used in this study is considered typical for Italy in the reference years.

3.1.3. Data Acquisition

Primary data concerning the sliced bread production process were provided by Valle Fiorita Srl (Ostuni, Italy). This comprehensive inventory included logistics for ingredients, auxiliary, and packaging materials, along with by-products and packaging waste resulting from the production lines and after delivery to distribution centers, as detailed in the following sections.

Primary energy consumption data (natural gas, grid electricity, and photovoltaic production) were also collected from detailed utility bills and meter readings for the period 2021–2022 (see Section 3.2.4 for a detailed description of the data breakdown).

Due to a lack of specific information on the exact ingredient varieties, a hybrid data acquisition approach was implemented to characterize the Carbon Footprint (CF) of the diverse inputs.

The average CF value and standard deviation for the cultivation and production subsystems (e.g., soft wheat, olive oil, salt, etc.) were extracted from the SU-EATABLE LIFE database. This database represents a meta-analysis by Petersson et al. [25,26], which combined findings from multiple scientific studies, including a previous review by Clune et al. [27].

The CF for packaging (e.g., PE bags, cardboard, heat-shrink film) and auxiliary materials (e.g., detergents, lubricating oil) was partly calculated using SimaPro Craft 10.2.0.2 software (PRé Consultants, Amersfoort, The Netherlands), integrating primary factory data with emission factors sourced from the Ecoinvent v. 3.10 database.

The remaining factors were extracted from other references [28,29,30].

Figure 1. System boundaries from cradle to distribution centers for conventional white sliced bread production. All symbols are given in the Nomenclature section.

Figure 1. System boundaries from cradle to distribution centers for conventional white sliced bread production. All symbols are given in the Nomenclature section.

3.1.4. Data Quality Evaluation

The dataset reliability of this LCA study was quantitatively assessed by employing the Pedigree matrix-based method [31], which adheres to the structure defined by the Data Quality Indicators (DQIs). The methodology, including the six assessment criteria (Reliability, Completeness, Time-related, Geographical, and Technological Representativeness, and Parameter Uncertainty), and the detailed Pedigree Matrix with individual flow scores are fully described in Appendix A. The resulting overall Data Quality Rating (DQR) [32] for the Life Cycle Inventory (LCI) data is reported in Section 4.1.

3.2. Life Cycle Inventory

The inventory analysis was performed to evaluate the consumption of materials, water, and energy, as well as the generation of waste.

With reference to the system boundaries shown in Figure 1, the main phases of the production process are described as follows:

(1)

Preparation and long kneading of the dough to promote the development of the gluten network and give the dough elasticity and tenacity. To produce conventional white sliced bread,

Table 1. Primary ingredient data for white sliced bread: Company registered quantities (±10% precision).

Ingredients	Mass Fraction (zi)	Mass Ratio (Zi/SWF0) *
	[% w/w]	[g/g]
Soft Wheat Flour Type 0 (SWF0)	60.0	1
Sourdough Starter	12.0	0.200
Block Brewer’s Yeast	2.0	0.033
Dextrose Monohydrate	1.5	0.025
Iodized Salt	1.5	0.025
Preservative (Propionic Acid)	0.1	0.0017
Whole Milk Powder	0.7	0.0117
Sunflower Seed Oil	0.6	0.010
Extra-Virgin Olive Oil	0.6	0.010
Soy Flour	1.0	0.017
Tap Water	20.0	-

* Referred to soft wheat flour type 0.

lists the ingredients with their mass fraction (z_i) and ratio (Z_i/SWF0) referred to soft wheat flour type 0 (SWF0).

Specifically, the sourdough starter was internally prepared by mixing water (55% w/w) and SWF0 (45% w/w) and inoculating 0.5% (w/w) of a lactic acid bacteria-based starter propagated in a laboratory fermenter and recovered by centrifugation from the culture media. Given the insignificance of the starter’s contribution (lyophilized lactic acid bacteria), this component will not be considered in the environmental impact assessment of the process under study. The preservative used, that is calcium propionate (E282), is especially effective at delaying mold growth both in baked goods (bread, biscuits, pastries) and for the surface treatment of aged cheeses. Its presence must be listed in the ingredients. Finally, the soy flour used consisted of hulled soybean seeds finely ground with a protein content of 38.6% (w/w).

(2)

Dough leavening: The dough is leavened at 37 °C and 80% relative humidity (RH) for 1.5–2 h. Each dough block of about 1.8 kg (see Figure S2a in the electronic Supplementary Material) was divided into equal portions (generally 7 loaves), as illustrated in Figure S2b. Each loaf is placed into a special rectangular pan (Figure S3a), lightly greased or lined with parchment paper, to give it the characteristic shape of sandwich bread. Figure S3b shows the seven loaves placed in each pan before being put into the proofing cabinet (Figure S4). Finally, Figure S3c shows the loaves at the end of the proofing process, where the increase in volume causes them to completely fill the pan.

(3)

Dough baking in a Rotor oven (Figure S5) at 210 °C for 20–25 min. The baking takes place in pans placed on a rotating rack, which allows the ventilated air to evenly touch the product being baked. The air is conveyed and distributed within the chamber through adjustable ducts. This promotes the evaporation of water and the formation of the crust and the characteristic soft crumb. The oven is also equipped with a large cascade steamer, composed of individual cast-iron elements that cover the entire back wall of the oven, designed to deliver water vapor to completely envelop the product. The oven is powered by natural gas.

(4)

Cooling of the product to room temperature. After baking, the racks with the pans are removed from the oven (Figure S6a) to allow them to cool to room temperature.

(5)

Removal of the sandwich loaves from the pans with the cooked loaves repositioned on the racks (Figure S6b) for storage in a cold room for 24 h.

(6)

Slicing the whole sandwich loaf into slices with a base of 12 cm, a height of 12 cm, and a thickness of 1.3 cm (Figure S7a) using automatic slicers (Figure S7b). The average moisture and protein contents of each sliced sandwich bread were 36.4 ± 0.6% (w/w), and 9.64 ± 0.14% (w/w), respectively. The initial and final slices are discarded because their crust is less appealing, their shape is irregular, and they are more exposed to heat, making them more susceptible to burning or uneven browning. In addition to the end slices, any trim and broken slices of bread are also discarded. All these scraps are not repurposed for products like breadcrumbs or croutons but donated to local livestock farms for reuse.

Table 2. Primary data for post-dough production: Mass (M_i) variation, scraps, and processing yield (M_i/M₁) from baking, cooling, and slicing phases (±3% precision).

Sandwich Bread	Mi [g]	Mi/M1 [%]
1—Dough put into the pan and ready to be baked	1800	100.00
2—Bread loaves removed from the oven and pan	1585	88.06
3—Bread loaves removed from the pans and cooled in a refrigerated cell	1553	86.28
4—End bread slices discarded	156	8.67
5—Bread slices discarded during packaging	75	4.17
6—Bread slices actually packaged	1322	73.44

shows the mass variation of the leavened dough after oven baking, cooling in the refrigerated cell, and slicing, which includes discarding the end slices and broken ones. The bread’s weight decreases by about 12% after baking and by approximately 14% after cooling. The total scraps (end and broken slices) account for about 13% of the weight, resulting in a final yield from production of pre-sliced bread of approximately 73.4% of the initial weight, with a potential variation of 3%. The higher the packaged slice yield, the smaller the overall Carbon Footprint (CF) will be. This is because a higher yield directly reduces the volume of by-products that need to be allocated for secondary uses like animal feed. While valorization into animal feed is beneficial, reducing the initial waste stream is always the primary way to minimize the CF and limit the need for downstream valorization efforts.

(7)

Primary packaging in a modified atmosphere: Four slices of bread are placed into a 5.8-g polyethylene (PE) food bag. An automatic machine then removes the air from the bag, sprays the slices with 1.5 mL of a 94% (v/v) ethanol solution, and fills the bag with a modified atmosphere (a gas mixture of 50% N₂ and 50% CO₂). The package is then sealed and an adhesive paper label weighing 1.2 g is added. This modified atmosphere packaging extends the sandwich bread’s shelf life by at least 45 days.

The hydroalcoholic solution is supplied in 10-L high-density polyethylene drums for food use, each with a unit mass of 350 g. This aqueous solution primarily consists of ethyl alcohol (94.0 ± 0.3% v/v), with minor quantities of other substances such as propylene glycol (0.36%) and glycerol (0.2%). Given the composition and density of the individual components at 20 °C (ethanol: 789 kg/m³; propylene glycol: 1036 kg/m³; glycerol: 1261 kg/m³; water: 1000 kg/m³), the mixture is estimated to have a density of 799.5 kg/m³, meaning 1.5 mL has a mass of approximately 1.2 g. According to Italian Ministry of Health Decree no. 312 of 13 July 1998 [33], special bread can be treated with ethyl alcohol (Art. 1b) when it is sold in sealed, waterproof packaging, either sliced or whole (Art. 2). This is permitted as long as the quantity of ethanol does not exceed 2% by weight of the food, expressed on a dry-matter basis (db), and the product label includes the phrase “treated with ethyl alcohol.” In this case, the addition of 1.5 mL of the hydroalcoholic solution (of which 1.41 mL, or 1.128 g, is ethanol) to a package of four slices weighing 143 g (equivalent to 90.96 g db) results in an ethanol/dry matter ratio of 1.124%, which is well within the legal limit.

The modified atmosphere packaging uses a gas mixture in packs of 16 cylinders. Each cylinder has a 50-L capacity and an empty weight of 80 kg. Loaded to a pressure of 100 bar, each cylinder contains about 7.5 kg of the gas mixture, resulting in a weight ratio between the full cylinder and the gas mixture of 11.67 kg/kg. Since each package of sandwich bread has a volume of 1365 cm³ (13 cm × 15 cm × 7 cm), approximately 2.1 g of the gas mixture is injected into it, while 0.2 g is released into the environment. A single 50-L gas cylinder can therefore package approximately 3261 primary packages, which corresponds to 466 kg of sliced bread.

Figure 2 shows the block diagram for the primary modified atmosphere packaging of sandwich bread slices.

(8)

Secondary packaging: Eight 143-g packages of sliced sandwich bread are placed into a 167-g recycled cardboard box. The box is then sealed with 4-g of PE tape, and a 2-g paper label is added, as shown in Figure 2.

(9)

Tertiary Packaging: The cardboard boxes are assembled onto a 25-kg EPAL wooden pallet for transport. Each pallet holds 16 cartons per layer, with 12 layers total. The boxes are secured to the pallet using 2.3 kg of heat-shrinkable PE film, and the pallet is then labeled with two 4.8-g paper tags, as shown in Figure 2.

Table 3. Complete packaging inventory for the sandwich bread system (based on 1 full pallet).

Packaging Type	Technical Specifications	Unit
Primary Packaging	PE bags
Mass of sandwich bread slices	143 ± 2	g
PE bag mass	5.8 ± 0.1	g
Width × Depth × Height	13 × 7 × 15	mm × mm × mm
Adhesive paper label mass of	1.20 ± 0.06	g
Hydroalcoholic solution mass	1.2 ± 0.1	g
CO2-N2 mixture mass	2.1 ± 0.2	g
Primary packaging overall mass	153.3 ± 2.5	g
Secondary Packaging	Cardboard carton
No. of primary packages	8	-
Length × Width × Height	290 × 200 × 155	mm × mm × mm
Carton mass	167 ± 2	g
Adhesive paper label mass	2.0 ± 0.1	g
PE scotch tape mass	4.0 ± 0.2	g
Mass of sandwich bread per carton	1.144 ± 0.020	kg
Secondary packaging overall mass	1.399 ± 0.020	kg
Tertiary Packaging	Wooden EPAL pallet
Pallet mass	25.0 ± 0.2	kg
Length × Width × Height	1200 × 800 × 144	mm × mm × mm
No. of cartons per layer	16	-
No. of layers per pallet	12	-
Overall height of pallet	2.004	m
Paper label per pallet	no. 2 × (2.4 ± 0.2)	g
Stretch-and-shrink PE film	2.3 ± 0.3	kg
Mass of sandwich bread per pallet	219.65 ± 2.00	kg
Tertiary packaging overall mass	295.97 ± 5.00	kg

presents the complete packaging inventory for the sandwich bread, specifying the mass and technical specifications for each component at three hierarchical levels: Primary Packaging (PE bags and their contents), Secondary Packaging (shipping cardboard cartons), and Tertiary Packaging (final loaded wooden EPAL pallet with stretch film).

Table S1 presents the average percentage of product and packaging material waste recorded during the company’s production. Additionally, Figure 2 illustrates the block diagram for the waste management process of the sliced bread production and packaging line. The waste is collected in separate bins as plastic waste (RPL: RPL_IN, PED, SPE, SSC, SFP), paper and cardboard waste (RCC: RCC_IN, SET, SCA, SEC, SEP), and wood waste (RL: SPAL). The organic residue (SPC) is collected and used as animal feed (MZ).

The system examined (see Figure 1) also includes:

(10)

Storage of the palletized product at room temperature.

(11)

Transportation of the product to distribution centers by truck.

(12)

The disposal of both processing and post-consumer packaging waste.

3.2.1. Material Balance of Sandwich Bread Production

The production of packaged sliced bread is visually represented in Figure 1, which illustrates the complete process flow. The sequence of steps involves preparing the sourdough starter and dough, leavening, baking, cooling, and the final stages of slicing and packaging (further detailed in Figure 2).

The material balance of these processes, including associated waste management, is summarized in Table S2. All material quantities are referenced to an input of 1000 kg of soft wheat flour type 0, as the Reference Flow for scaling the ingredient inventory.

3.2.2. Packaging Formats of Various Ingredients and Additives

Soft wheat flour type 0 (SWF0) is received in bulk via specific food-grade tanker trucks with an average capacity of 16–32 Mg. This flour is then unloaded into silos with a maximum capacity of 22 Mg. Brewer’s yeast (BY) is purchased in 500-g blocks, packaged as indicated in Table S3. Among the other ingredients, dextrose monohydrate (DM), iodized salt (IS), calcium propionate (PA), whole milk powder (WMP), and soy flour (FS) are purchased in 25-kg paper bags. These bags are 80 cm tall, 40 cm wide, and 7 to 14 cm deep with a unit mass of approximately 115 g, are stacked horizontally in a stable, interlocking pattern to prevent shifting during transport on EPAL wooden pallets (see Figure S8) and secured with heat-shrink PE film. Both sunflower seed oil (SO) and extra-virgin olive oil (EVOO) are purchased in 1000-L intermediate bulk containers. These are made of high-density polyethylene (HDPE) and are equipped with a top lid, a 2-inch butterfly discharge valve (with an anti-tamper seal), and a metal/plastic pallet, as shown in Figure S9. Table S4 summarizes the main geometric characteristics of this container, as well as the net and gross weights, from which the empty container weight can be calculated as 55 kg.

The consumption of these ingredients generates packaging waste, which is categorized as either paper-cardboard (RCC_IN) or plastic (RPL_IN). These waste streams are subsequently separated for recycling and managed through a differentiated disposal system.

Table 4. Mass and type of primary packaging per ingredient, including paper-cardboard (RCC_IN) and plastic (RPL_IN) waste generated and the mass of all components on the pallet (MPP_i) per unit mass of the ith ingredient.

Ingredient	Primary Packaging		RCCIN	RPLIN	MPPi
	Type	Mass [Kg]	[g/Kg]	[g/Kg]	[kg/Kg]
Soft wheat flour type 0	Bulk in tank	(13–15) × 103	-	-	-
Compressed yeast	Paper-Cellophane bags	0.5	35.86	1.43	1.059
Dextrose monohydrate	Paper bags	25	4.60	2.22	1.007
Calcium propionate	Paper bags	25	4.60	2.22	1.007
Iodized salt	Paper bags	25	4.60	2.22	1.007
Whole milk powder	Paper bags	25	4.60	2.22	1.007
Sunflower seed oil	IBCs	916	-	-	1.060
Extra-virgin olive oil	IBC	916	-	-	1.060
Soy flour	Paper bags	25	4.6	2.22	1.007
EtOH solution	HDPE drums	10	-	43.78	1.044
CO2-N2 mixture	Stainless steel cylinders	7.5	-	-	11.667
Detergents	PE drums	1.4	-	39.22	1.039
Lubricating oils	PE drums	1.4	-	53.03	1.053

IBC, Intermediate bulk container.

presents the mass of primary packaging for all consumed ingredients, the resulting waste, and the mass ratio of the palletized product to the packaged product (MPP).

3.2.3. Logistics of Input and Output Materials

Table 5. Logistics of input/output materials with indication of the means of transport used with the corresponding load capacity and distance travelled from different production sites to destination ones. All symbols are listed in the Nomenclature section.

Input/Output Materials	From	To	Means of Transport	Load Capacity [Mg]	Distance [km]
Soft Wheat Flour Type 0 (SWF0)	PS	FG	Euro 5 semi-trailer truck	16–32	788
Compressed Yeast	PS	FG	Euro 5 semi-trailer truck	16–32	854
Dextrose Monohydrate	PS	FG	Euro 5 semi-trailer truck	16–32	963
Iodized Salt	PS	FG	Euro 5 semi-trailer truck	16–32	157
Calcium Propionate	PS	FG	Euro 5 semi-trailer truck	16–32	963
Whole Milk Powder	PS	FG	Euro 5 semi-trailer truck	16–32	726
Sunflower Seed Oil	PS	FG	Euro 5 semi-trailer truck	16–32	142
Extra-Virgin Olive Oil	PS	FG	Euro 5 semi-trailer truck	16–32	142
Soy Flour	PS	FG	Euro 5 semi-trailer truck	16–32	870
EtOH solution	PS	FG	Euro 5 semi-trailer truck	16–32	960
CO2-N2 Cylinders	PS	FG	Euro 5 semi-trailer truck	16–32	92
PE Film Rolls	PS	FG	Euro 5 semi-trailer truck	16–32	103
Adhesive paper labels	PS	FG	Euro 5 semi-trailer truck	16–32	159
Cardboard boxes	PS	FG	Euro 5 semi-trailer truck	16–32	483
PE tape and heat-shrink film	PS	FG	Euro 5 semi-trailer truck	16–32	115
EPAL pallets	EPMC	FG	Euro 5 semi-trailer truck	16–32	92
	CD	EPMC	Euro 5 semi-trailer truck	16–32	805
Detergents	PS	FG	LCV	1.3	24
Lubricating oils	PS	FG	LCV	1.3	17
Palletized sliced sandwich bread	FG	DC	Euro 5 semi-trailer truck	16–32	713
Plastic waste RPL	FG	WCC	MWCS	13.9	26
Paper/cardboard waste RCC	FG	WCC	MWCS	13.9	26
Wood waste RL	FG	WCC	MWCS	13.9	26
Organic waste (MZ)	FG	CaF	LCV	1.3	30

CaF, Cattle farm; DC, Distribution center; EPMC, Euro pallet managing center; FG, Factory gate; LCV, Light Commercial Vehicle; MWCS, Municipal Waste Collection Service; PS, production site; WCC, Waste Collection Center.

shows the logistics of the input/output materials with the type and load of the means of transport used and overall distance travelled from the places of production to those of use/delivery, as mainly derived from the processing plant of reference.

For the transport of ingredients in paper bags, articulated lorries with a maximum cargo capacity of 16–32 Mg are used. By-products (MZ) from the sandwich bread production process are given free of charge to local farms within a 30-km radius of the production facility. These farms use their own vans to collect these waste materials. The company producing the sandwich bread has no waste from raw materials or uncooked dough, as production is initiated based on pre-established orders. Therefore, there is no storage of finished products. Cardboard and plastic waste are given to a specialized waste disposal company located near to the processing plant. EPAL pallets are supplied to the production company by a specialized firm, which retrieves them from distribution centers and refurbishes them when necessary.

For the distribution of the sandwich bread slices, the company uses articulated lorries with a load capacity of 16–32 Mg. The weighted average distance traveled by each articulated lorry, based on the number of packages transported, was 726 km in 2021 and 700 km in 2022. Based on the analysis of historical data for individual vehicle routes, the remainder of the study will assume each articulated lorry travels an average distance of 713 km, with a variation range between 700 and 726 km.

3.2.4. Energy Sources

The sandwich bread production process at Valle Fiorita Srl in Ostuni (Italy) primarily uses electricity and methane. A significant portion of the electricity demand is met by a photovoltaic system installed on the factory roofs. Any surplus electricity generated, such as on non-working days or during off-peak hours (as the company operates in two daily shifts), is fed into the national grid.

Natural gas, grid electricity (medium voltage), and the production and grid injection of photovoltaic (PV) electricity consumed by Valle Fiorita Srl were all sourced from detailed utility bills and meter readings. This provided data on the monthly and total consumption/production for 2021 and 2022. Furthermore, as shown in Table S5, these data were categorized by time slots: A1 (peak hours), A2 (intermediate hours), and A3 (off-peak hours). This temporal and tariff-specific breakdown was essential for accurately modeling the environmental impact of the different energy sources.

The refrigerated cells used for cooling the sandwich bread (after it is removed from the pans and placed on trolleys for 24 h) and for storing fillings have an overall nominal power of 18 kW. The sandwich bread occupies an average of 35% of the refrigerated cells’ volume, while the remaining 65% is used to store ingredients and semi-finished products for filling sandwiches. The operation of the industrial refrigerated cells is regulated by a duty cycle, which alternates between ON and OFF periods to maintain the desired temperature with maximum energy efficiency. Assuming the compressor and other refrigeration system components have an 80% duty cycle, the annual energy consumption would be:

18 (kW) × 24 (h/day) × 365 (days/year) × 0.8 = 126,144 kWh/year

Of this consumption, 65% (~82,000 kWh/year) is related to the storage of sandwich fillings and will be deducted from the total energy consumption.

Based on the lower calorific value of methane (35.8 MJ/m³) and the volumetric consumption data in Table S5, the thermal energy used was estimated to be between 324.66 MWh in 2021 and 495.63 MWh in 2022.

The total thermal and electrical energy consumption for the two-year period is summarized in

Table 6. Total consumption of thermal energy (TE) and grid electricity (GE), and energy from photovoltaic sources (PV) for the total production of baked goods at Valle Fiorita Srl (Ostuni, Italy) in 2021–2022, with related specific energy consumption.

Absolute and Relative Consumption	Unit	Year 2021	Year 2022
Thermal Energy (TE)	kWh/year	324,656	495,631
Grid Electricity (GE)	kWh/year	537,969	571,165
Grid Electricity referred to sandwich bread only (GEPC)	kWh/year	455,969	489,165
Photovoltaic Electricity (PV)	kWh/year	8,176,016	9,750,135
Total Production of Baked Goods	Mg/year	3772.5	3712.6
Specific Thermal Energy	kWh/kg	0.086	0.133
Specific Grid Electricity	kWh/kg	0.121	0.132
Specific Photovoltaic Electricity	kWh/kg	2.167	2.626

. By relating this consumption to the annual production of baked goods, excluding fillings, the specific energy consumption (thermal, grid electricity, and photovoltaic electricity) per kg of finished product was determined.

3.2.5. Fugitive Emissions of Refrigerant Gases

Although widely used in various commercial refrigeration applications, R404a has a high Global Warming Potential (GWP) of 3922 kg CO_2e/kg and a near-zero Ozone Depletion Potential [34]. Due to its high GWP, its use in new equipment is now banned, though it can still be used for the maintenance of existing systems until 2030. In the absence of specific data, an annual refrigerant loss of 5% was assumed [35], corresponding to a total loss of 3.15 kg. Of this total, approximately 1.1 kg (35%) was attributed to the refrigeration of the sandwich bread, equivalent to 5.41 g per kg of soft wheat flour used.

3.2.6. Process Water Consumption

Based on data from the reference firm, the consumption of process water for the dough was in the order of 0.486 L per kg of sliced sandwich bread (see Table S2). Total water consumption for sanitation—including washing of pans, mixers, equipment, and premises—was assumed to be approximately double the dough ratio, amounting to 0.97 L per kg of sliced sandwich bread.

3.2.7. Process Aids

The equipment for producing sandwich bread requires cleaning and lubrication. The average consumption of detergents was about 2920 kg in 2021 and 4424 kg in 2022, corresponding to a ratio of 0.8 and 1.2 g/kg of sliced sandwich bread, respectively. An average ratio of 1.0 g/kg was assumed. For lubricating oils, a ratio of 0.02 L/kg of sliced sandwich bread was assumed, based on consumption data from a pasta factory [36]. Both products are packaged in 30-L PE tanks, each having a mass of 1.4 kg. Finally, all other materials used in minimal quantities (i.e., other minor chemicals and waste from process equipment, etc.) were not included within the system boundaries because their potential influence on the analysis results was considered negligible, being less than 1% (PAS 2050: Section 6.3: [16]).

3.2.8. Overall Waste Management

All waste from the sandwich bread’s life cycle is collected in different colored bins, following municipal solid waste collection procedures. Specifically, waste was divided into the following categories:

-

Plastic, paper, and cardboard: This includes packaging from ingredients (like paper bags and HDPE drums), finished product packaging (such as PE bags, cardboard, labels, and heat-shrink PE film), and packaging used to deliver bread to stores.

-

Wood: This includes damaged pallets.

-

Food Scraps: Damaged bread, end slices, and trimming scraps are given to local livestock farms for free, and they handle the transportation.

Packaging waste was then disposed of based on the overall Italian solid urban waste management scenarios for 2020, as detailed in

Table 7. Overall waste management scenarios for packaging waste in Italy in 2020 [37].

Disposal Scenario	Landfill [%]	Recycling [%]	Incineration [%]
Paper and cardboard waste	5.2	87.3	7.5
Plastic waste	6.7	48.7	44.6
Wood waste	35.4	62.4	2.2

[37].

Wastewater from cleaning equipment and company premises, as well as from toilets and sinks, was disposed of in the municipal sewer system. Its volume was assumed to be equal to the total tap water consumption (see Section 3.2.6).

3.3. Environmental Impact Assessment

The environmental impact of the sandwich bread was assessed by focusing on the Climate Change impact category [16]. This category was quantified using the Global Warming Potential (GWP) metric, which relies on the IPCC 2021 method over a 100-year time horizon [38]. The resulting environmental burden is referred to as the Carbon Footprint (CF), which was estimated by summing the products of activity data and their corresponding emission factors, as defined by the following equation:

CF = ∑_i (Ψ_i × EF_i)

(1)

where Ψ_i represents the magnitude of each activity (expressed in mass, energy, or mass times km), and EF_i is the corresponding emission factor. Since each activity was referenced to the chosen functional unit (FU: 1 kg of sliced sandwich bread packaged in PE bags in a modified atmosphere), the resulting CF was expressed in kg CO_2e per kg of product.

The CF calculation followed a hybrid approach, combining primary data with emission factors sourced from external databases as described in Section 3.1.3. Specifically, the Ecoinvent v. 3.10 database employed the Allocation, cut-off, EN15804 system model [39]. This mandatory system model dictates that the producer is fully responsible for the disposal of their own waste and receives no credit for supplying recyclable materials. Consequently, all potential CO_2e credits from the recycling of renewable and non-renewable materials were excluded from the final CF calculation.

The key Carbon Footprint datasets used for this process are summarized in Table S6 of the electronic supplement.

3.4. Sensitivity Analysis

To assess the sensitivity of the LCA model defined by Equation (1), it was calculated how different sources of uncertainty in the i-th independent variable (x_i) affect the final Carbon Footprint (CF) value. Two methods can be used to achieve this:

-

Differential Approach: Mathematically determine sensitivity by taking the partial derivative of CF with respect to a single factor (x_i), while holding all other factors (x_j) fixed.

-

One-Factor-At-a-Time (OFAT) Approach: Systematically adjust a single factor (x_i) away from its nominal value, leaving all the other (x_j) at their baseline, to directly observe the change. This allows for easy comparison since all tests reference the same starting point.

The OFAT method, despite its ease, has a major drawback: it fails to capture simultaneous input variations or interactions. According to Saltelli et al. [40], this method is inappropriate unless the model is proven to be linear, like the simple form in Equation (1).

Thus, for a model assumed or proven to be linear, the sensitivity analysis can use a uniform test perturbation, such as ±100%, chosen independently of the actual physical variability bounds reported in any database. This is mathematically justified because, in a linear system, the influence of a factor (x_i) is constant across its entire range (∂CF/∂x_i is a constant). Therefore, the ±100% range serves as a standardized measure to isolate and quantify each factor’s contribution, allowing for a normalized comparison of how the dependent variable (CF) is affected by varying each independent variable.

For the mathematical computation of sensitivity, Equation (1) was differentiated with respect to each independent variable x_i, holding the other variables constant (x_j≠i = const). Each partial derivative (∂CF/∂x_i) was then used to determine the relative variation of CF (∆CF) with respect to a reference value (CF_R) using a first-order Taylor polynomial and assuming a given degree of relative variation for the i-th variable (Δx_i/x_iR), as follows:

$${\displaystyle \frac{∆\mathrm{C}\mathrm{F}}{{\mathrm{C}\mathrm{F}}_{\mathrm{R}}}}\approx {\displaystyle \frac{1}{{\mathrm{C}\mathrm{F}}_{\mathrm{R}}}}{({\displaystyle \frac{\partial \mathrm{C}\mathrm{F}}{\partial {\mathrm{x}}_{\mathrm{i}}}})}_{{\mathrm{x}}_{\mathrm{j}\ne \mathrm{i}}} ∆{\mathrm{x}}_{\mathrm{i}}$$

(2)

with

∆x_i = x_i − x_iR

(3)

ΔCF = CF − CF_R

(4)

Since the generic independent variable x_i can represent either the emission factor (EF_i) or the activity data (Ψ_i), the partial derivative (∂CF/∂x_i) can be estimated as

$${\left. {\displaystyle \frac{\partial CF}{\partial x_i}}\right|}_{x_{j\ne i}}={\mathrm{EF}}_{\mathrm{i}}\hspace{1em}\hspace{1em}\mathrm{for} {\mathrm{x}}_{\mathrm{i}}={\Psi }_{\mathrm{i}}$$

(5)

$${\left. {\displaystyle \frac{\partial CF}{\partial x_i}}\right|}_{x_{j\ne i}}={\Psi }_{\mathrm{i}}\hspace{1em}\hspace{1em}\mathrm{for} {\mathrm{x}}_{\mathrm{i}}={\mathrm{EF}}_{\mathrm{i}}$$

(6)

where x_iR is the reference value of the generic ith variable.

Thus, under the assumption of a uniform relative perturbation value (Δx_i/x_iR) for all factors, the relative change in the Carbon Footprint (ΔCF/CF_R) for the i-th component will be mathematically identical in form, regardless of whether the independent variable (x_i) is the Emission Factor (EF_i) or the Activity Data (Ψ_i). In fact, substituting the partial derivatives (Equations (5) and (6)) into the Taylor expansion (Equation (2)) yields the same structure for both cases:

$${\displaystyle \frac{∆\mathrm{C}\mathrm{F}}{{\mathrm{C}\mathrm{F}}_{\mathrm{R}}}}\approx {\displaystyle \frac{\left({EF}_{iR} {\Psi }_{iR}\right)}{{\mathrm{C}\mathrm{F}}_{\mathrm{R}}}} \left({\displaystyle \frac{∆x_i}{x_{iR}}}\right)$$

(7)

This means the sensitivity is directly proportional to the reference contribution of the i-th component (EF_iR × Ψ_iR) to the total Carbon Footprint (CF_R).

Consequently, to pinpoint the main contributors to the CF for the sliced sandwich bread studied, a sensitivity analysis was conducted by independently perturbing selected emission factors and activity data by ±100% relative to the default condition. This covered the effect of the emission factors characterizing the production of ingredients, packaging materials, electricity and thermal energy sources, as well as activity areas such as the delivery distance of soft wheat flour and final product.

Moreover, the CF was evaluated in the following alternative scenarios:

(i).

Use of electrical energy entirely sourced from the national medium-voltage grid.

(ii).

Use of grid and photovoltaic electrical energy obtained from solar panels installed on-site (the reference company case).

(iii).

Use of entirely photovoltaic electrical energy.

(iv).

Rail transport for ingredients and the finished product.

3.5. Dataset Uncertainty Estimation

Due to the constraints of using spreadsheet-based Life Cycle Assessments without dedicated Monte Carlo analysis software, the quantitative uncertainty of the final Carbon Footprint (CF) was estimated using a Data Quality Indicator (DQI)-based uncertainty propagation method. The uncertainty for each input flow’s contribution to the CF was quantified based on the Parameter Uncertainty (P) criterion (ranging from 1 to 5) of the Pedigree Matrix (Appendix A). This criterion is directly correlated to specific geometric standard deviations (σ_g). All individual uncertainty contributions are assumed to be independent and normally distributed. The overall standard uncertainty of the final CF is then calculated by combining these individual uncertainties using the Root Mean Square (RMS) method. This procedure for propagating the geometric standard deviations is based on established practice in LCA [31]. The specific correlation table, the full set of equations, and the detailed uncertainty propagation analysis are fully described in Appendix B.

4. Results

4.1. Data Quality Rating

The quantitative assessment of the LCI data quality, performed using the Pedigree matrix-based method described in detail in Appendix A, yielded an Overall Data Quality Rating (DQR) for the LCI data used in this study, its average value was 1.5 ± 0.4. According to the DQR scale [32], this score corresponds to an Excellent quality dataset (DQR ≤ 1.6). The full Pedigree Matrix detailing the scores for the main data flow items is provided in

Table A1. Data quality indicator (DQI) matrix referred to the main data used in this LCA study.

DQI	R	C	TiR	GR	TeR	P	Score
Data Flow Item
Wheat Flour (1st Ingr contributor)	1	1	1	1	1	1	1.00
	Data from factory inventory checks.	Data covers all the quantities yearly used.	Data refers to the study period.	Flour sourced from Italian suppliers.	Flour fits technology specification.	≤5%
Cartons (1st PM contributor)	1	1	2	3	2	2	1.83
	Data from factory inventory checks.	Data covers all the quantities yearly used.	≤5-years old secondary data.	Regional/national data.	Current industrial technology.	~10%
PE bags (2nd PM contributor)	1	1	2	2	2	2	1.67
	Data from factory inventory.	Data covers all the quantities yearly used.	≤5-years old secondary data.	Regional data.	Current industrial technology.	~10%
Thermal Energy	1	1	1	2	2	1	1.33
	Quantity from monthly gas bills.	Monthly metered data.	Data refers to the study period.	Sourced from the national gas grid.	Current oven Technology.	Gas-meter precision: 2–5%.
Grid Electricity	1	1	1	2	2	1	1.33
	Quantity from monthly energy bills.	Monthly metered data.	Data refers to the study period.	Sourced from the national MV grid.	Current equipment technology.	Power-meter precision: 2–5%.
Photovoltaic	1	1	1	1	1	1	1.00
Electricity	Quantity calculated from monthly energy bills.	PV production and export data from monthly bills.	Data refers to the study period.	Sourced from the on-site PV system.	Specific on-site PV system.	Calculated combined uncertainty: 7–10%.
Transport	1	1	2	3	1	3	1.83
	Quantity and transport means from factory data.	Primary data inputs for the transport modelling.	Data covers the study period with Ecoinvent v3.10 treated as up to date.	Delivery distances estimated via Micheline distance-meter.	Technologically specific transport LCI data.	Background LCI data with unknown Uncertainty.
Waste Disposal	3	3	1	2	1	3	2.17
	Quantities estimated using average waste rates.	Quantities estimated, not measured.	Data refers to the study period with LCI disposal scenarios treated as up to date.	Waste delivered to local centers using national waste disposal scenarios.	LCI waste disposal treatments regarded as up to date.	Typical uncertainty of secondary LCI data: 20–30%.

All symbols are reported in the Nomenclature section.

in Appendix A.

4.2. Carbon Footprint of Conventional Sliced White Bread

Table 8. Specific greenhouse gas emissions (SGHGE) and carbon footprint for a functional unit of 1 kg of sliced bread in 4-slice, modified atmosphere PE packaging: relative and total contributions for scenarios where electricity is sourced from the national grid, and from partial or total photovoltaic (PV) power.

Electricity	National Grid (100% Grid)		Partial Photovoltaic (95% PV)		Total Photovoltaic (100% PV)
Life Cycle Stage	SGHGE [kg CO2e/kg]	[%]	SGHGE [kg CO2e/kg]	[%]	SGHGE [kg CO2e/kg]	[%]
Ingredients (Ingr)	0.54	19.4	0.54	27.5	0.54	28.1
Processing (Proc)	1.15	41.5	0.33	16.9	0.29	15.2
Emissions to Air (EA)	0.03	1.0	0.03	1.4	0.03	1.4
Packaging (Pack)	0.62	22.3	0.62	31.6	0.62	32.3
Transportation (TR)	0.37	13.2	0.37	18.8	0.37	19.2
Waste disposal (WD)	0.07	2.7	0.07	3.8	0.07	3.9
Carbon Footprint (CF)	2.77	100.0	1.95	100.0	1.91	100.0

summarizes the specific greenhouse gas emissions (SGHGE) and resulting Carbon Footprint (CF) for a functional unit of 1 kg of sliced bread in 4-slice, modified atmosphere PE packaging, comparing the results across three distinct scenarios for electricity sourcing: 100% National Grid, partial photovoltaic (PV) supply (95% PV), and total PV supply (100% PV). The life cycle scope, which spans from cradle to distribution centers, includes the disposal of packaging waste (PE bags, cardboard, labels, tape, shrink wrap) and the post-consumption release of gases (N₂ and CO₂) from the PE bags into the air. The detailed analysis and implications of the partial and total PV scenarios are further discussed in Section 4.4.

Figure 3 illustrates the absolute and relative CO_2e contributions of the main life cycle phases (Ingredients, Processing, Emissions to Air, Packaging, Transportation, Waste disposal) to the overall Carbon Footprint (CF), providing a visual comparison among the three electricity sourcing scenarios: 100% National Grid (blue bars), partial Photovoltaic (PV) supply (yellow bars), and total PV supply (red bars).

The analysis shows that the processing phase is the main source of emissions, accounting for 41.5% of the total carbon footprint (CF). Specifically, the consumption of grid electricity (95.4%) and thermal energy (4.6%) are the factors with the greatest impact. This is followed by the production of packaging materials (22.3% of the CF) and the ingredients used for sliced bread (19.4% of the CF). Finally, the fourth critical point is the transportation of raw materials and the finished product (13.2% of the CF). The estimated Carbon Footprint of conventional sliced white bread, packaged in 4-slice PE bags, was estimated at 2.77 kg CO_2e/kg. However, a portion of the product, approximately 17.5% of the total sliced bread (PCF) by mass, is diverted as waste for animal feed (SPC) (see Table S2). Different allocation methods yield different results for the primary product:

i.

Economic allocation: Applying the economic allocation criterion based on the waste being given away for free results in assigning zero environmental impact to the feed portion, thereby attributing the entire 2.77 kg CO_2e/kg to the salable bread. This approach is generally criticized as it relies on a non-market or zero-price transaction rather than the material’s true economic value.

ii.

Material Balance Allocation: In contrast, the material balance allocation criterion assigns the impact proportionally to the mass split. This allocates 14.9% of the total CF to the waste stream and the remaining 85.1% to the packaged sliced bread, resulting in a CF of (0.851 × 2.77≅) 2.36 kgCO_2e/kg for the bread.

iii.

Substitution/System Expansion: Alternatively, a substitution approach can be used, which is generally preferred when by-product streams enter another system. By comparing the waste bread (protein content of 9.69%) to the conventional animal feed it replaces—de-oiled soybean meal (protein content of about 50% with a CF of 2.556 kg CO_2e/kg per Table S6—a CO_2e credit of approximately 0.5 kg CO_2e/kg can be attributed to the waste for displacing virgin feed. Consequently, the overall CF of the packaged conventional sliced white bread, derived from this system expansion and credit, is reduced to about 2.68 kg CO_2e/kg.

4.3. Sensitivity Analysis Results

The CF sensitivity was estimated by varying the i-th emission factor (EF_i) by +100% relative to its corresponding default value (EF_iR).

Table 9. Percentage relative change (ΔCF/CF_R) of the CF from cradle to distribution centers relative to the reference value (CF_R) for a +100% relative change in the generic parameter (Δx_i/x_iR).

Factor xi	Default Value				(ΔCF/CFR)
	Ψi	Unit	EFi	Unit	[%]
Soft wheat flour (SWF0)	1000	kg	0.43	kg CO2e/kg	12.25
Compressed yeast (BY)	30.2	kg	0.82	kg CO2e/kg	0.70
Monohydrate dextrose (DM)	22.7	kg	1.47	kg CO2e/kg	0.94
Calcium propionate (PA)	1.5	kg	1.81	kg CO2e/kg	0.08
Iodized salt (IS)	22.7	kg	0.10	kg CO2e/kg	0.06
Whole milk powder (WMP)	10.6	kg	10.52	kg CO2e/kg	3.13
Sunflower seed oil (SO)	9.1	kg	0.99	kg CO2e/kg	0.25
Extra-virgin olive oil (EVOO)	9.1	kg	3.84	kg CO2e/kg	0.98
Soy flour (SF)	5.8	kg prot.	6.00	kg CO2e/kg	0.50
Thermal energy (TE)	171.0	kWh	0.28	kg CO2e/kWh	1.33
Grid electricity (GE)	3530.6	kWh	0.40	kg CO2e/kWh	39.37
Deionized water	1.2	kg	0.48	kg CO2e/m3	0.02
Lubricating oil	17.9	kg	1.54	kg CO2e/kg	0.78
Detergents	1.3	kg	0.71	kg CO2e/kg	0.03
PE bags	52.5	kg	3.1	kg CO2e/kg	4.60
N2	8.0	kg	0.2	kg CO2e/kg	0.05
CO2	12.6	kg	0.6	kg CO2e/kg	0.20
Ethanol	10.8	kg	3.5	kg CO2e/kg	1.06
Bag labels (ET)	10.8	kg	2.8	kg CO2e/kg	0.86
Cartons (CA)	188.0	kg	2.6	kg CO2e/kg	13.82
Carton labels (EC)	2.2	kg	2.8	kg CO2e/kg	0.18
Scotch tape (SC)	4.5	kg	2.5	kg CO2e/kg	0.32
Pallet labels (EP)	0.03	kg	2.8	kg CO2e/kg	0.002
Shrink wrap film (FP)	13.4	kg	3.1	kg CO2e/kg	1.17
SWF0 transport distance	788	km	0.188	kg CO2e/(Mg km)	4.17
Palletized sliced bread transport distance	713	km	0.188	kg CO2e/(Mg km)	3.78
Carbon Footprint (CFR)	3559.8 kg CO2e/(1000 kg SWF0)

The data in italics highlights the parameters with a sensitivity on the CF higher than 5%.

shows the percentage relative change in CF (ΔCF/CF_R) for the sliced bread under examination.

This analysis reveals that grid electricity (GE) is by far the most influential factor. A 100% increase in the GE emission factor leads to a 39.4% rise in the CF, highlighting the production process’s strong dependence on electricity.

Immediately following, cardboard boxes (CA) emerge as the second critical factor. A 100% increase in their emission factor results in a 13.8 increase in the CF. This, along with the 4.6% increase due to PE bags, suggests the need to optimize packaging solutions to reduce the environmental footprint.

It is also interesting to note the impact of soft wheat flour (SWF0), with a 12.3% increase in the CF following a 100% increase in its emission factor. This underscores the importance of sustainability in the wheat supply chain.

Replacing grid electricity with renewable sources like solar power is a priority mitigation strategy. Installing on-site solar panels could drastically reduce the environmental impact. However, it is also crucial to optimize packaging, with a focus on recycled or biodegradable materials to reduce the overall carbon footprint. Finally, the impact of the wheat supply chain must also be considered, and the adoption of sustainable agricultural practices and the use of low-impact wheat varieties should be evaluated.

4.4. Effect of Using Photovoltaic Electricity on the Sliced Bread CF

The company meets a significant portion of its electricity needs with a rooftop photovoltaic (PV) system (see Table S5). Table 8 reports the resulting GHG emissions for the cradle-to-distribution center life cycle under these mitigated conditions. Figure 3 provides a comparison of the relative contributions across the scenarios (see yellow bars for this mitigated condition).

The decision to produce about 95% of the company’s consumed electricity on-site represents a major mitigation measure. This choice reduced the carbon footprint of sliced bread by approximately 30%, dropping from 2.77 to 1.95 kg CO_2e/kg (Table 8).

Critically, this shift changed the primary environmental hotspots. The production phase of packaging materials has become the new dominant hotspot, responsible for about 31.6% of emissions, followed closely by the production of ingredients (27.5%). The transportation of raw materials and the finished product ranks third, accounting for 18.8% of the CF. Consequently, the Processing step, which was previously the main source of emissions under the 100% Grid Electricity scenario, is now the fourth critical factor (16.9% of CF), as its emissions are primarily controlled by the consumption of low-carbon photovoltaic electricity (90.9% of the total processing energy). Further modeling of a 100% PV Electricity scenario (i.e., replacing the remaining grid consumption) shows a modest additional reduction of 2.1%, lowering the CF from 1.95 to 1.9 kg CO_2e/kg. While the overall ranking of the main hotspots remained unaltered in this scenario, the contribution of the Processing step would further reduce from 16.9% to 15.2% of the total CF (see red bars in Figure 3).

4.5. Uncertainty Assessment of the Estimated Carbon Footprint

The quantitative uncertainty of the final Carbon Footprint was estimated using the Data Quality Indicator (DQI)-based uncertainty propagation method detailed in Appendix B. The overall standard uncertainty (u_T) and the contribution of individual life cycle steps to this uncertainty are reported in

Table A2. Correlation matrix of parameter uncertainty (P) scores and assigned geometric standard deviations (σ_g) for dataset quality assessment.

P Score	Data Quality Description	σg
1	Verified, measured data (e.g., from supplier)	1.05
2	Estimated or literature data (good representativity)	1.10
3	Estimated or literature data (fair representativity)	1.20
4	Professional judgment or high uncertainty	1.50
5	Highly uncertain, estimated data (e.g., proxy)	2.00

(Appendix B) for all modeled scenarios.

For the 100% Grid Electricity scenario (Section 4.2), the calculated overall standard uncertainty (u_T) was 3.87%, corresponding to a 95% confidence interval of ±7.74% (approx. ±2 u_T).

The substitution of grid electricity with on-site photovoltaic (PV) power led to a marginal increase in overall uncertainty, with the u_T rising to 4.84% for the 95% PV scenario and 4.93% for the 100% PV scenario. This increase suggested that the introduction of site-specific PV data introduces greater data variability into the model, as noted in the DQI assessment (Appendix A).

A detailed analysis of the individual contributions (u_i) across all scenarios clearly identified Transport (contributing 2.64% to 3.83% of u_T across scenarios) as the single largest contributor to the overall uncertainty. This high contribution is primarily due to its lower data quality score (p = 3; σ_g = 1.20), as determined in the Pedigree Matrix (Appendix A). Conversely, Grid Electrical Energy, despite being a major CF contributor in the base case, contributed less to the final uncertainty due to its high data quality (p = 1; σ_g = 1.05).

These findings indicated that data quality improvement efforts should prioritize the Transport and Photovoltaic Electricity life cycle steps to narrow the confidence intervals of the LCA results.

The uncertainty analysis was then used to assess the statistical robustness of the differences observed between scenarios:

i.

100% Grid vs. 95% PV: The large reduction in CF when transitioning from 100% Grid to 95% PV (~30% reduction) was statistically significant at the 95% confidence level. This confirmed that the observed mitigation effect was robust and not due to data variability.

ii.

95% PV vs. 100% PV: The minor additional reduction in CF when transitioning from 95% PV to 100% PV was not statistically significant at the 95% confidence level. This indicates the calculated difference of 2.1% (from 1.95 to 1.91 kg CO_2e/kg) lies within the combined uncertainty range of the two estimates.

In summary, the uncertainty analysis confirmed that the major conclusion regarding the benefit of implementing the PV system was robust, while highlighting areas (Transport, PV data) where data quality improvement would be most impactful.

4.6. Potential Mitigation Opportunities and Future Analysis

The identification of key hotspots allows for the evaluation of potential future mitigation strategies. These strategies primarily focus on Packaging, Transportation, and Gas Management:

(a)

Packaging and Process Optimization: In the 100% Grid Electricity scenario, optimizing the sliced bread packaging is the first key opportunity, given that the production phase of packaging materials is responsible for approximately 22% of emissions. Further mitigation actions should focus on reformulating the packaging to reduce the ratio between gross weight and net weight, which is currently 1.35 kg/kg of sliced bread (Table 4). This would benefit both the reduction of packaging waste and the optimization of transportation. Despite the LCA results identify this as a priority, implementing any reduction in packaging mass is contingent upon external technical confirmation that the required modified atmosphere and product shelf life are unaltered.

(b)

Transportation Modal Shift: A quantified reduction in greenhouse gas emissions could be achieved by replacing road transport with rail transport for all ingredients (SWF0, BY, DM, SI, PA, WMP, SO, EVOO, and SF), the finished product, and the return of pallets from distribution centers to the reconditioning company. Based on this modal shift, the sliced bread’s carbon footprint might be supplementary reduced by an additional 8.7%, that is, from 2.77 to 2.53 kg CO_2e/kg. This reduction turned out to be statistically significant at the 95% confidence level on the assumption of an unchanged overall standard uncertainty (u_T) value.

(c)

On-site Gas and Refrigerant Management: Further opportunities focus on mitigating the environmental impact associated with auxiliary inputs and fugitive emissions.

(c1)

Modified Atmosphere Gases: The environmental impact of managing the gases used to create the modified atmosphere inside the PE bags could be reduced by adopting an on-site system for generating gaseous nitrogen. This system would filter, dehumidify, and compress air to feed a hollow-fiber membrane module, separating the air into an oxygen permeate and 99.5% pure nitrogen retentate. This solution, requiring an estimated energy consumption between 0.5 kWh/(STP m³) [41] and 0.556 kWh/(STP m³) [42], would eliminate the need to transport nitrogen cylinders. The potential transition to liquid carbon dioxide storage at low pressure (e.g., 2 bar) could also be considered. Supplying CO₂ via tanker trucks to on-site cryogenic tanks, rather than using high-pressure cylinders, could reduce the overall impact of transport. However, this switch must be carefully evaluated based on the associated acquisition and operational costs (e.g., for cryogenic tanks and receiving infrastructure), which can vary significantly depending on overall annual consumption.

(c2)

Refrigerants: To reduce the environmental impact of fugitive emissions, the refrigerated cells currently charged with the high Global Warming Potential (GWP) refrigerant R404a could be converted to use propane (R290), a refrigerant gas with a similarly negligible ozone depletion potential but a significantly lower GWP (approximately 3 kg CO_2e/kg) compared to the R404 blend [43].

5. Discussion and Conclusions

While traditional bread has a low environmental impact, its transformation into a pre-sliced and packaged product significantly increases its carbon footprint (CF). Our analysis of pre-sliced white bread in polyethylene bags, with a total energy consumption of 8.55–10.41 MJ/kg (see Table 6), places it on the lower end of the expected energy ranges for bread [8]. Crucially, the Life Cycle Inventory (LCI) data supporting these findings was rated as “Excellent” with an average Data Quality Rating (DQR) of 1.5 ± 0.4, confirming a high level of confidence in the dataset’s representativeness. Despite this reliable data and relatively low energy consumption, the product’s CF of 2.77 kg CO_2e/kg when using 100% Grid Electricity, is high, heavily influenced by factors beyond just energy use, such as:

-

Packaging materials: The multiple layers of packaging—from plastic bags to transport cartons—add substantially to the product’s environmental impact.

-

Additional processing: Slicing and hermetic packaging require extra industrial steps and energy consumption.

-

Ingredients and additives: The production of emulsifiers and preservatives used to extend shelf life also contributes to the overall CF.

-

Shelf life and waste: Although a longer shelf life can reduce household waste, the packaging and portioning process itself generates more waste at the industrial level.

A sensitivity analysis confirms that adopting renewable energy, such as solar or wind power, can reduce CF by up to 30% (1.95 kg CO_2e/kg). This reduction is statistically robust, confirmed by the uncertainty analysis which showed the difference between the 100% Grid and 95% PV scenarios is statistically significant at the 95% confidence level. However, this alone is not a sufficient solution. To effectively mitigate the environmental impact, a multi-faceted approach is necessary, encompassing packaging optimization, improved waste management, and the promotion of sustainable agricultural practices for wheat cultivation.

Furthermore, the uncertainty assessment identified Transport and Photovoltaic data as the main sources of data variability (u_T reached ~4.9% in the PV scenarios), indicating these are priority areas for data collection in future studies.

Transitioning to recyclable, mono-material, or biodegradable alternatives (such as polylactic acid, PLA, or polyhydroxyalkanoate, PHA) requires careful assessment of potential trade-offs. Adopting these alternatives must not compromise the extended shelf life of packaged sliced bread, and their environmental performance needs thorough scrutiny. Biodegradable options often entail higher upstream emission factors (e.g., in raw material production) or require specialized composting infrastructure that is not yet universally available. These factors must be explicitly considered and compared against conventional or mono-material options within the same LCA framework.

Crucially, this study reveals a key trade-off: the carbon footprint of food waste is significantly greater than that of the packaging designed to prevent it. A single wasted bread slice (~36 g) results in 69.7 ± 2.5 g of CO_2e emissions, a burden approximately 15 more impactful than the 4.5 ± 0.3 g of CO_2e associated with the PE packaging portion for that slice. This aligns strongly with the conclusions of Crosnier et al. [19], which highlighted that bread waste can amplify consumption-based impacts by up to 80% because a substantial portion of production is discarded. Consequently, reducing waste represents one of the most effective levers for improving bread system sustainability. This underscores that while optimizing packaging is vital, a far more impactful strategy is to ensure consumers finish the entire loaf, thereby preventing the waste of a product whose entire life cycle represents a much larger environmental footprint.

These results provide strong evidence for the need to shift sustainability efforts toward the consumption stage. Our findings robustly support consumer outreach campaigns that encourage simple actions, such as freezing unused slices or mindful consumption, demonstrating that preventing one wasted slice (~70 g CO_2e avoided) generates an environmental benefit significantly greater than the CF of its packaging portion.

In conclusion, while packaged bread offers convenience and a potential solution to food waste, a holistic approach is required that targets the entire supply chain—from sustainable farming and renewable energy to smarter packaging and, most critically, consumer behavior. Future research should expand the LCA to include other environmental impact categories, such as acidification, eutrophication, and land use, to provide a more comprehensive framework.