Accepted Manuscript

Carbon footprint of the photovoltaic power supply chain in China

Xiaopeng Guo, Kai Lin, Han Huang, Yang Li

PII: S0959-6526(19)32060-8

DOI: 10.1016/j.jclepro.2019.06.102

Reference: JCLP 17271

To appear in: Journal of Cleaner Production

Received Date: 19 July 2018

Accepted Date: 10 June 2019

Please cite this article as: Xiaopeng Guo, Kai Lin, Han Huang, Yang Li, Carbon footprint of the
photovoltaic power supply chain in China, Journal of Cleaner Production (2019), doi: 10.1016/j.
jclepro.2019.06.102

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form.
Please note that during the production process errors may be discovered which could affect the
content, and all legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Improve raw
Raw material material
development development
technology

Comparing with Improve

thermal power Parts production and
development
manufacturing
efficiency

Calculating Carbon Suggestions are made

China vigorously Focus on the PV
Emissions in according to different
supports supply chain carbon Logistics and Optimize parts
Photovoltaic Supply links to reduce carbon
photovoltaics emissions transportation production process
Chain emissions.

Replace transport
Installation and
vehicles with
maintenance
electric vehicles

Improving waste
The waste recovery recycling efficiency
stage and waste disposal
methods
 ACCEPTED MANUSCRIPT

Carbon footprint of the photovoltaic power supply chain in

China

Xiaopeng Guo a.b*, Kai Lin a, Han Huang a, Yang Li c

a School of Economics and Management, North China Electric Power University; Hui Long

Guan, Chang Ping District, Beijing 102206, China;

b Beijing Key Laboratory of New Energy and Low-Carbon Development, North China Electric
Power University; Hui Long Guan, Chang Ping District, Beijing 102206, China;
c State Grid Fujian Electric Power Company Limited
*Correspondence: [email protected]; Tel.: +86-13520328997

ABSTRACT
With the increasing consumption of fossil fuels, environmental problems are
becoming increasingly serious. To solve these problems, we must vigorously
develop clean energy. To this end, China has promulgated many policies to support
the development of photovoltaic energy. To study the carbon footprint of the
photovoltaic power supply chain and calculate the reduction of carbon emissions,
this article establishes a carbon emission mathematical calculation model for
photovoltaic power generation systems during the production, transportation and
waste treatment processes. The carbon emission reduction model is established by
calculating the power consumption of the photovoltaic power supply chain and
power generation throughout the life cycle and by using the 1 kW photovoltaic
power generation system as an example to analyse the data. The results show that
from the perspective of the supply chain, it can effectively reduce carbon emissions
from photovoltaic power supply chains by improving raw material development
technology to provide high development efficiency, optimizing parts production
processes, replacing transportation vehicles with electric vehicles gradually,
increasing waste recycling efficiency, and improving waste disposal methods. In
2017, compared with thermal power generation in China, photovoltaic power
generation systems were used in areas where the solar radiation is effective for
1000 h-3000 h, the CO2 emission reduction could be considered to be between
1.738 GT and 3.078 GT, which have shown good carbon emission reduction effect.
Keywords: Photovoltaic power supply chain; Carbon emission; Carbon emission reduction；
Carbon footprint
 ACCEPTED MANUSCRIPT

1. Introduction
In recent years, the consumption of fossil fuels has increased dramatically
worldwide, and environmental pollution and other problems have become worse. It
is particularly important for China, a major energy-consuming country, to solve
these problems. The development of new clean energy can solve this problem
effectively. The many advantages of solar energy make it the world's first choice
for clean energy, so China has also begun to support photovoltaic power
generation in recent years. In 2016, the global capacity of newly installed
photovoltaics was 71.2 GW, an increase of 51.24% year-on-year, and cumulative
installed capacity was approximately 300 GW, an increase of 32% year-on-year.
China is the world’s largest market, and the newly installed capacity in 2016 was
34.2 GW, which accounted for 46% of the total. In the first three quarters of 2017,
China added approximately 43 GW of installed photovoltaic capacity, which was
an increase of 65.4% year-on-year, and the cumulative installed capacity was
approximately 120 GW, an increase of 2.6% year-on-year (Government report,
2017). To better develop photovoltaics, China has issued a number of policies and
regulations to promote the development of the photovoltaic power industry. For
example, the “Renewable Energy Law,” which was officially implemented on
January 1, 2006, was the earliest policy and regulation. In July 2017, the National
Energy Bureau of China issued the “Guidance on the Implementation of the 13th
Five-Year Plan for Renewable Energy Development” to guide the development of
the photovoltaic power industry from 2017 to 2020. Ground-level power plants
(including the Leader project) will maintain an average annual installed capacity of
20-23 GW, and distributed projects will adopt a record-based system, which is not
subject to indicators. In addition, centralized power stations, distributed
photovoltaic, village-level poverty alleviation power stations, and cross-provincial
and cross-district power transmission stations in 7 provinces (district and city),
including Beijing, Shanghai and Tianjin, are now planning to support 20-23 GW of
installed capacity, which accelerated the photovoltaic poverty alleviation project.
In conclusion, China's photovoltaic power generation industry is developing
rapidly, but we must also realize that photovoltaics are not zero emissions. From
the perspective of supply chain, there are carbon emissions in photovoltaic power
generation. This paper takes the typical photovoltaic power supply chain in China
as an example to calculate the carbon emissions of the PV supply chain and
 ACCEPTED MANUSCRIPT

compare it with China's thermal power carbon emissions to evaluate the emission
reduction effect. According to different supply chain links, the paper puts forward
relevant suggestions to reduce photovoltaic carbon emissions and further improve
the clean photovoltaic energy to ensure that it can provide healthy and low-carbon
energy security for people's life and social development. The specific structure of
article is shown in Fig 1.

Improve raw
Raw material material
development development
technology

Comparing with Improve

thermal power Parts production and
development
manufacturing
efficiency

Calculating Carbon Suggestions are made

China vigorously Focus on the PV
Emissions in according to different
supports supply chain carbon Logistics and Optimize parts
Photovoltaic Supply links to reduce carbon
photovoltaics emissions transportation production process
Chain emissions.

Replace transport
Installation and
vehicles with
maintenance
electric vehicles

Improving waste
The waste recovery recycling efficiency
stage and waste disposal
methods

Fig 1. The structure of the article.

2. Background and literature review
Many scholars studied carbon footprint in the early years, such as Barthelmie
et al. (2008) suggested that the carbon footprint refers to the total amount of CO2
produced by a product or activity throughout its life cycle. Larsen and Hertwich
(2009) proposed a meaning of carbon footprint at the core of the product and
considered the carbon footprint to refer to the sum of greenhouse gases produced
by the product throughout its life cycle. Wiedmann and Minx (2008) believed that
the carbon footprint includes both direct and indirect emissions and that when the
quality of CO2 is used as a unit, the remaining greenhouse gases can be converted
to CO2. This article uses ISO14067 content (ISO14067), the product carbon
footprint refers to the total amount of greenhouse gas emissions directly and
 ACCEPTED MANUSCRIPT

indirectly generated during the process of obtaining, manufacturing, transporting,

selling, using and discarding raw materials.
The carbon footprint has been recognized as an effective tool for measuring
carbon emissions from various activities associated with products (Schmidt, 2009).
Lombardi et al.(2018) applied the urban carbon footprint (UCF) and relative
carbon footprint (RCF) to a small city in the south of Italy. The study focused only
on carbon dioxide emissions. The results showed that the UCF came from
electricity and gas consumption in 2015. Prehoda and Pearce (2017) reviewed the
data and quantified the lives saved by the replacement of U.S. coal-fired electricity
with solar PV systems. The results showed that the conversion of coal-fired power
generation to solar power has great benefits to peoples’ health and the
environment. Jin et al. (2018) used the traffic flow data of toll stations in Jiangsu
Province of China to estimate the carbon footprint of self-driving travel between
counties. Khan et al. (2018)found the effects of financial development, income
inequality, energy usage, and per capita GDP on carbon dioxide emissions and
studied the environmental Kuznets curve (EKC) for three developing Asian
countries – Bangladesh, India, and Pakistan. Suo et al. (2017) developed a
two-stage, type-2 fuzzy stochastic programming method for supporting the clean
production of energy systems with carbon and pollutant mitigation under
uncertainty. Sukumaran and Sudhakar (2017) analysed the first-year operational
data of the Solar Power Airport of India Cochin International Airport Limited, and
the economic and environmental aspects confirmed its effectiveness in reducing
carbon footprint. Laine et al. (2017) took the 20 largest cities in Finland as the
research object to analyse the role of the local municipal energy system in the
carbon footprint of urban residents. The study complemented previous carbon
footprint assessments of urban residents through energy system analysis. Lee
(2017) studied the carbon footprint of large-scale antenna systems and showed that
the expected energy efficiency gain of large-scale antenna systems had sufficient
practicality and carbon emission reduction functions. Bazán et al. (2018)
re-evaluated urban rooftop areas that were not fully utilized to determine the power
supply and greenhouse gas reduction potential of three middle cities in Peru.
Boamah et al. (2018) made use of carbon footprint to empirically investigate the
causal effects of urban population growth and international trade on environmental
pollution in China from 1980 to 2014 and further established the presence of an
N-shaped environmental Kuznets curve relationship between the economic growth
 ACCEPTED MANUSCRIPT

and environmental pollution of China. In summary, the concept of carbon footprint

has been enriched since its initial conception. It has become an effective tool for
measuring carbon emissions and has been widely used in carbon emissions
research in various industries.
Carbon footprint research methods include the input-output method, life cycle
assessment method and IPCC method (Wang et al., 2017), which are the three
most widely applied. For example, Dong and Geng (2012) used the input-output
method to study the characteristics of the direct carbon footprint and indirect
carbon footprint of Beijing residents' in 2007. According to the analysis of the
data, they obtained the total carbon footprint of rural residents and urban residents
in Beijing. Cansino et al. (2016) used an enhanced structural decomposition
analysis supported by detailed input–output tables from the World Input–Output
Database for the period of 1995–2009 to analyse the structural decomposition of
changes in CO2 emissions in Spain. Liu et al. (2011) used the whole life cycle
method to perform a bottom-up calculation on the carbon footprint of a coal-fired
power plant in Luliang City, Shanxi Province and obtained a detailed distribution
of the plant’s carbon footprint. Arvesen et al. (2015) also used the life cycle
method to evaluate and analyse power transport in Lager County, northern
Norway. The carbon footprint of its total transmission and distribution was 7.8
kg/MWh, and the local transmission accounted for approximately 60% of its
largest share. Zhao et al. (2014) used the IPCC method to analyse the carbon
footprint of Shanxi Province in China from 1999 to 2010. Based on the analysis
results, the overall carbon footprint changes, energy intensity, and per capita
changes were summarized. Singh et al. (2016) used a combination of the
fundamental mass-balance approach for energy consumption and the methodology
defined by IPCC for the carbon emissions, and, using mainly data from 50 sewage
treatment plants, he gave a detailed account of the energy consumption and carbon
footprint of different technologies and different scales of sewage treatment.
China's research on the photovoltaic supply chain began in 2006. The research
content includes product markets, corporate relations, and coordination
mechanisms in the supply chain. It contains qualitative and quantitative analyses,
but the depth and accuracy of the research on the balance between supply and
demand are very limited. Scholars in other countries have formed very rich
theoretical achievements in the study of the risk pool effect of the supply chain,
 ACCEPTED MANUSCRIPT

involving inventory management, warehouse location, supply chain/network

design, product distribution and many other links (Sun et al., 2014).
Chen and Su (2018) studied the business dynamics of the multi-photovoltaic
supply chain system theoretically and numerically to get better conclusions and put
forward appropriate industrial policies and business strategic directions for the
government and enterprises. Su Qing-zhong and Yuan (2009) believed that each
node in the supply chain plays a different role in the supply chain system and bears
different functions. Each node provided different contributions and relevance to the
supply chain system. Using analytic hierarchy process (AHP), he made a
quantitative analysis of the contribution degree and correlation degree of the
supply chain node enterprises. Cucchiella et al. (2015) used net present value,
internal rate of return, discounted payback period, discounted total cost return rate,
and carbon dioxide emission reductions to determine the number of PV systems
needed to achieve the target. Chen Zhi-song et al. (2011) used the game model and
the supply chain coordination mechanism to quantitatively analyse the benefits of
photovoltaic supply chain companies under competition, cooperation, and
coordination mechanisms. According to the analysis results, he proposed a
coordination mechanism for upstream and downstream profit distribution in the
supply chain, which would help improve the overall performance of the supply
chain. Marigo (2007) studied the development of photovoltaic supply chains in
China. He mainly studied the current development status of photovoltaic
production capacity. Through quantitative analysis, it was identified that the main
factors that affected the sustainable development of photovoltaic capacity were the
shortage of silicon material upstream and the lack of technological innovation in
the middle reaches. According to the risk pool effect, Dai et al. (2008) analysed
two retailers with independent demand and normal distribution and concluded that
centralized warehouse supply can reduce cost.
From the perspective of supply chain, this paper studies the carbon footprint
of photovoltaic power industry, and calculates the sum of direct carbon emissions
and indirect carbon emissions, which are generated by various energy, materials
and manpower consumed in the whole process of production, use, maintenance and
scrap recovery of photovoltaic power supply chain by using life cycle assessment
method. Direct carbon emissions refer to carbon dioxide emissions directly into the
atmosphere in the production process, and indirect carbon emissions are carbon
emissions from energy consumption in the production process.
 ACCEPTED MANUSCRIPT

3. Material and methods

3.1 Theoretical model

The carbon footprint of the photovoltaic power supply chain mainly involves
the production of photovoltaic panels. Based on each node in the life cycle of
photovoltaic panels, this article constructs a one-way carbon chain structure: raw
material development, parts production and manufacturing, logistics and
transportation, installation and maintenance, and the waste recovery stage.
In the raw material development stage, various resources related to the
production of materials are mined, and these activities generate gases such as CO2.
In the manufacturing stage, carbon emissions mainly consist of direct and indirect
emissions in the production process, and from the current research, indirect CO2-eq
emissions account for more than 90% of the total emissions (Reich et al., 2007).
Transportation exists in each stage of the supply chain. After the raw material is
mined, it is transported to the manufacturing plant for production. It is then
transported to the designated place for installation and use. Finally, the reverse
transportation of waste recycling is carried out. All of the above are components of
logistics and transportation in the supply chain context. The installation and
maintenance phase includes inverters, maintenance cost, feasibility studies and so
on, and eventually converts them into carbon emissions. In the waste recovery
stage, a series of related activities will produce CO2 and other gases through the
existing technology for recycling.
The production process of the photovoltaic power generation system is very
complex, and the complete photovoltaic power generation system includes the
system controller, the photovoltaic module, the inverter and many other
accessories and equipment. There are dozens or even more processes for producing
photovoltaic modules. Therefore, in the photovoltaic supply chain, carbon
emissions mainly focus on the carbon emissions from activities such as
exploitation of resources related to production materials, the carbon emissions
produced by the fuel in the transportation process, the carbon emissions produced
by the consumption of equipment, and the carbon emissions from the disposal of
waste (Zhan, 2011).
 ACCEPTED MANUSCRIPT

Table 1.
CO2-eq emissions in different stages of photovoltaic supply chain.
Stages of PV supply chain Sources of CO2-eq emissions
Raw material development Emissions from mining production materials
Emissions from the production of photovoltaic
Parts production and manufacturing
modules and other components
Emissions from transportation of materials at each
Logistics and transportation
node
Maintenance, feasibility studies, etc. will
Installation and maintenance
eventually be converted to carbon emissions
Waste recovery Emissions from recycling activities
The direct carbon emission of photovoltaic power is directly emitting carbon
dioxide into the atmosphere in the production process. For example, when
extracting coarse silicon, the reaction is:
SiO2（quartz sand）+C→Si+CO2
It can be seen from the equation that the extraction of coarse silicon produces
a large amount of carbon dioxide. Therefore, we need to pay attention to the
calculation of carbon footprint. Indirect carbon emissions refer to the carbon
emissions generated by energy consumption in the production process. According
to the ISO14064-1 (ISO14064-1, 2006), strict carbon emissions must also include
plant building, production equipment, personnel and any other related carbon
emissions, but this part of carbon emissions is very small compared to the above
production process. This article omits the calculation of this part, which will not
affect the accuracy of the whole calculation.
The calculation model of the total emissions of photovoltaic power generation
systems in the production process refers to the calculation model in Gong’s article
(Gong et al., 2013). The influence factor is introduced in the calculation process to
make the carbon footprint calculation more accurate. We also add the carbon
emission calculation of transportation and waste treatment to increase the integrity
of the supply chain carbon footprint calculation model. In sum, the calculation
model is as follows:
C = NC𝑝 +T + W…… (1)
In the model: C is the total emission of a photovoltaic power generation
system;
 ACCEPTED MANUSCRIPT

C𝑝 is the total emission of PV power generation systems for unit power during
the production process.
N represents total power;
T is the total discharge of the transport process; and
W represents emissions from waste.
The calculation formula for the total emissions of the unit power PV system
during the production process is as follows:
𝑛
C𝑝 = ∑1（𝐷E𝑛 + 𝑐𝐵𝑛）𝑓𝑛 …… (2)
As shown in Eq. (2): 𝐷E𝑛 represents the direct carbon emissions per unit
product in the production process;
𝐵𝑛 indicates the amount of electricity consumed by the production unit
product;
𝑐 represents the carbon emissions generated by the unit electricity;
𝑓𝑛 indicates the influence factor. And the specific content is shown in Table 2.
Table 2.
Carbon emission in the production process of photovoltaic power system.
System Two level Production Direct carbon Power Influence
components classification process emissions consumption factor
Preparation of
DE1 B1 f1
coarse silicon
Preparation of
high purity DE2 B2 f2
Battery sheet polysilicon
Silicon wafer
Photovoltaic DE3 B3 f3
preparation
module
Preparation of
DE4 B4 f4
battery sheet
Glass DE5 B5 f5
Backplane DE6 B6 f6
Frame DE7 B7 f7
EVA DE8 B8 f8
Junction box DE9 B9 f9
 ACCEPTED MANUSCRIPT

Table 2.
Carbon emission in the production process of photovoltaic power system.
System Two level Production Direct carbon Power Influence
components classification process emissions consumption factor
Photovoltaic
Photovoltaic
module DE10 B10 f10
module
preparation
Controller DE11 B11 f11
Inverter DE12 B12 f12
Others DE13 B13 f13

The total emission of the transport process is as follows:

T = T1 + T2 + T3 = D ∗ W ∗ ∑(𝐺𝑊𝑃𝑖 ∗ 𝜉𝑖 ∗ 𝑆𝑖 ∗ 𝑅𝑆𝑖) ……… (3)
As shown in Eq. (3), T1 represents the direct emission of CO2 during
transportation (kg), T2 represents the equivalent carbon emission of CH4 during
transportation (kg), T3 indicates the equivalent carbon emission of N2O during
transport (kg), D indicates the transport distance (km); W represents the
transport weight(t), 𝐺𝑊𝑃𝑖 represents global warming potential of greenhouse
gases, 𝜉𝑖 represents the conversion coefficient of other greenhouse gases to
CO2-eq, 𝑆𝑖 indicates the consumption intensity of fuel during transport (L/(t•
km)), and 𝑅𝑆𝑖 indicates the coefficient of greenhouse gas emissions emitted by
fuel (kg/L). And W = 𝐶𝑎 + 𝐶𝑏,𝐶𝑎 is the CO2 emission of wastewater treatment
process.
𝐶𝑎 = 𝛼 ∗ ∑𝑁𝑖……… (4)
𝛼 indicates the CO2 emission coefficient (kg/m3) in the wastewater process,
and 𝑁𝑖 indicates the amount of waste water produced by each process (m3).
𝐶𝑏 is the CO2 emission of the waste treatment process. There are two means of
mainstream waste disposal: incineration and landfills. In this article, we mainly
consider incineration, and its carbon emission model is as follows:
44
𝐶𝑏 = 12𝑄𝑏 ∗ 𝜀𝑏……… (5)
𝐶𝑏 is the carbon emission of incineration treatment (kg); 𝑄𝑏 is the waste
discharge of the unit products (kg); 𝜀𝑏 is the proportion of carbon in waste; and
44/12 is the reciprocal of C in the CO2 molecular formula.
 ACCEPTED MANUSCRIPT

To sum up, the carbon emissions model is:

𝑛 44
C = N∑1（𝐷E𝑛 + 𝑐𝐵𝑛）𝑓𝑛 +D ∗ W ∗ ∑(𝐺𝑊𝑃𝑖 ∗ 𝜉𝑖 ∗ 𝑆𝑖 ∗ 𝑅𝑆𝑖) + 𝛼 ∗ ∑𝑁𝑖 + 12𝑄𝑏 ∗ 𝜀𝑏……… (6)

3.2 Data collection

To study the carbon emissions of photovoltaic power generation systems, we
calculate the direct and indirect carbon emissions during the preparation of coarse
silicon. To make the calculation simple, the direct carbon emissions are
transformed into the corresponding amount of coal electricity. The power
consumption per unit product is based on the typical level of the current
technology. In November 4, 2016, the State Council issued the “13th Five-Year
programme for the control of greenhouse gas emissions.” For the first time, it was
clearly stated that "large power generation group unit power supply carbon dioxide
emissions are controlled within 550 g CO2/kWh."(Zhejiang Energy Conservation,
2016). According to the national regulation, the carbon emissions 𝑐 should not be
higher than 550 g CO2/kWh. Because the specific data are so difficult to collect,
this paper takes 550 g CO2/kWh for 𝑐. Table 3 is carbon emissions from typical
photovoltaic systems in China. In the table DE1 is purified from crude silicon,
which will produce direct carbon emissions. The other production processes will
not produce direct carbon emissions, so the direct emissions are zero.
Table 3.
Carbon emissions from the production process of a typical photovoltaic power generation system
in China (Gong et al., 2013).
Direct
Power consumption Impact factor
carbon emissions
DE1=1.57 B1=13kWh/kg f1=12.35
B2=150kWh/kg f2=8.32
B3=1.5kWh/kg f3=252
B4=0.98kWh/kg f4=252
B5=0.15kWh/kg f5=1000
B6=0.15kWh/kg f6=1000
B7=0.15kWh/kg f7=1000
B8=0.15kWh/kg f8=1000
B9=0.15kWh/kg f9=1000
B10=5.6kWh/kg f10=3.5
B11=0.31kWh/kg f11=1000
 ACCEPTED MANUSCRIPT

Table 3
Carbon emissions from the production process of a typical photovoltaic power generation system
in China (Gong et al., 2013).
Direct
Power consumption Impact factor
carbon emissions
B12=0.31kWh/kg f12=1000
B13=0.31kWh/kg f13=1000

In the production and transportation of PV modules, the use of fossil fuels is

the main source of carbon emissions. The carbon emission estimation model is:
𝐶𝑓 = 𝛽1𝐶𝑓1 + 𝛽2𝐶𝑓2 + 𝛽3𝐶𝑓3……… (7)
As shown in Eq. (7), 𝐶𝑓 is the carbon emission, 𝐶𝑓1 is standard coal
consumption, 𝛽1 is the carbon emission conversion factor of coal consumption,
𝐶𝑓2 is the standard coal consumption of petroleum, 𝛽2 is the carbon emission
conversion factor of petroleum consumption, 𝐶𝑓3 is the standard coal consumption
of natural gas, and 𝛽3is the carbon emission conversion factor for natural gas
consumption. The carbon emission factors of the three types of energy are shown
in Table 4. The standard coal conversion factors of the three types of energy are
shown in Table 5.
Table 4.
Table of carbon emission factors for three types of energy (Energy Research Institute, 2003).
Energy coal oil gas
βi (10,000 t carbon/10,000 t standard
0.7476 0.5825 0.4435
coal)

Table 5.
Standard coal conversion factor table for three energy sources (National Bureau of Statistics,
2006).
Coal Oil Gas
Energy
（kg coal /kg） （kg coal /kg） （kg coal /m3）
conversion factors 0.7143 1.4286 1.1－1.33

In China, the fuel used for transport vehicles is gasoline and diesel. To
calculate the carbon emissions of them, it is necessary to calculate the emissions of
 ACCEPTED MANUSCRIPT

CO2, CH4 and N2O. Among them, CH4 and N2O can be converted into CO2-eq for
carbon emission calculation. CO2, CH4, and N2O emission factors for gasoline and
diesel are listed in the following tables.
Table 6.
CO2 emission factors of gasoline and diesel (IPCC, 2006).
Suggested uncertainty
Emission source emission factor（kgCO2/L）
（95% confidence interval）
gasoline 2.26 -2.6%~5.3%
diesel 2.73 -2.0%~0.9%

Table 7.
CH4 emission factors for gasoline and diesel (IPCC, 2006).
Suggested uncertainty
Emission source emission factor（kgCH4/L）
（95%confidence interval）
gasoline 8.16*10-4 -66.7%~233.3%
diesel 1.44*10-4 -59.0%~143.6%

Table 8.
N2O emission factors for gasoline and diesel (IPCC, 2006).
Suggested uncertainty
Emission source Emission factor（kgCH4/L）
（95%confidence interval）
diesel 1.44*10-4 -66.7%~233.3%
gasoline 2.61*10-4 -66.7%~233.3%

Table 9.
Greenhouse Gas GWP Values (IPCC, 2001).
Greenhouse Gas GWP Value
CO2 1
CH4 23
N2O 296

4. Results and Discussion

In this article, we take a 1 kW photovoltaic power system as an example. As
shown in Eq. (1), we can obtain the total emission of unit power PV system in the
 ACCEPTED MANUSCRIPT

production process, and calculate the power consumption (𝐸𝑢) in the production
process as follows:
𝑛
C𝑝 ∑1（𝐷E𝑛 + 𝑐𝐵𝑛）𝑓𝑛
𝑛
𝐸𝑢 = c = 𝑐 = C𝑝 = ∑1（𝐷E𝑛/𝑐 + 𝐵𝑛）𝑓𝑛……… (8)

The data in Table 2 is brought into the carbon emission estimation model to
obtain 𝐸𝑢 = 3768.364 kWh. For a 1 kW photovoltaic system, the power
consumption during the production process is 3786.364 kWh. In this article, the
carbon emissions involved in the typical photovoltaic power generation system in
China are calculated as examples, which lack the actual transportation distance and
waste disposal data. However, the model is still practical because we only take the
data into the model to get more accurate carbon emissions from photovoltaic
systems. Considering that there are no actual data on transportation, waste
disposal, etc., this article uses a more general approach to estimate an increase of
20% on the basis of its power consumption (Gong et al., 2013) and conclude that
the system's power consumption is 4522.0368 kWh.
By calculating the power consumption of the supply chain of photovoltaic
power generation systems and comparing the generation capacity throughout the
life cycle, the carbon emissions reduction is calculated. Its electric power model is
as follows:
𝑚
𝐸𝑔 = 𝑒𝐻𝑃∑1 (1 ― 𝜁𝑚) ……… (9)
As shown in Eq. (9), 𝐸𝑔 is the total amount of electricity generated during
the life cycle of the PV system; 𝑒 is the photovoltaic generation efficiency of a
variety of factors, such as integrated component attenuation, temperature and dust;
𝐻 is the annual effective light time of the installation area; 𝑃 is the maximum
power of the photovoltaic module; m is the service life of the power generation
system; and 𝜁𝑚 is the attenuation rate of the power of the photovoltaic power
system for m years.
In summary, 𝐶𝑟, the total carbon emission reduction of photovoltaic power
generation system in its life cycle, is equal to （𝐸𝑔 ― 𝐸𝑢） ∗ c.
The average annual effective light duration in different regions of the country
varies greatly; for example, in the Qinghai Tibet Plateau and southern Xinjiang it is
approximately 3000 h, while in some parts of Sichuan it is only approximately
1000 h. In this article, we use the average annual effective illumination durations
of 3000 h, 2000 h and 1000 h to calculate and analyse. Generally, it is believed that
 ACCEPTED MANUSCRIPT

the effective service life of a photovoltaic power generation system is 25 years, and
the efficiency of a PV grid connected power station is usually 80%, which
represents that 𝑒 is 80%, and the annual average attenuation rate of the PV power
system is 2% (Carol, 2017). By entering these data into the power consumption
model, we can obtain the generation capacity of a PV system over the whole life
cycle under different annual effective light duration as follows:
Table 10.
Power generation of photovoltaic power generation system throughout its life cycle under
different effective annual light durations.
Total power generation under different light durations（kWh）
Years
3000 h 2000 h 1000 h
1 2400 1600 800
2 4752 3168 1584
3 7056.96 4704.64 2352.32
4 9315.8208 6210.5472 3105.2736
5 11529.50438 7686.336256 3843.168128
6 13698.9143 9132.609531 4566.304765
7 15824.93601 10549.95734 5274.97867
8 17908.43729 11938.95819 5969.479097
9 19950.26854 13300.17903 6650.089515
10 21951.26317 14634.17545 7317.087724
11 23912.23791 15941.49194 7970.74597
12 25833.99315 17222.6621 8611.331051
13 27717.31329 18478.20886 9239.10443
14 29562.96702 19708.64468 9854.322341
15 31371.70768 20914.47179 10457.23589
16 33144.27353 22096.18235 11048.09118
17 34881.38806 23254.25871 11627.12935
18 36583.7603 24389.17353 12194.58677
19 38252.08509 25501.39006 12750.69503
20 39887.04339 26591.36226 13295.68113
21 41489.30252 27659.53501 13829.76751
22 43059.51647 28706.34431 14353.17216
23 44598.32614 29732.21743 14866.10871
24 46106.35962 30737.57308 15368.78654
 ACCEPTED MANUSCRIPT

25 47584.23243 31722.82162 15861.41081

According to the calculation, in the place where the solar energy is abundant
and the effective illumination time is 3000h, the power generation of the
photovoltaic system exceeds the power consumption of the entire supply chain
within two years, that is, the recovery period of carbon emission is two years.
When the effective light is 3000 h, the total carbon emissions of the photovoltaic
power system with 1 kW of power during its life cycle is about 𝐶𝑟3 =
（𝐸𝑔 ― 𝐸𝑢3） ∗ c = 23.684 t; when the effective light is 1000 h, the total amount
of carbon emission reduction of the photovoltaic power system with 1 kW of
power during its life cycle is about 𝐶𝑟1 = （𝐸𝑔 ― 𝐸𝑢1） ∗ c = 13.374 t.
By the end of December 2017, accumulated installed capacity of photovoltaic
in China had exceeded 130 GW, and accumulated installed capacity of
photovoltaic in China ranked first in the world (Gao, 2018). In 2017, the CO2
emission reduction of China's photovoltaic power generation system could be
considered as between 1.738 Gt and 3.079 Gt. The results show that the
photovoltaic power supply chain has a good emission reduction effect compared
with thermal power. The energy recovery time and carbon footprint of commercial
roof photovoltaic system are the same. The energy recovery period is obviously
shorter than the life expectancy of 30 years. Compared with the carbon footprint of
fossil fuel power generation, the carbon footprint of photovoltaic power generation
is advantageous (Mariska, 2013).
Although the photovoltaic power supply chain reduces a large amount of
carbon emissions compared with the thermal power supply chain, we cannot ignore
the amount of carbon emissions produced in the production process.
Therefore, further technological improvements in photovoltaic module
production and system balance module manufacturing, improving the operational
period of time of PV systems, and widespread use of renewable energy as the main
energy can effectively reduce the carbon footprint of photovoltaic system
(Nikolaos and Christopher, 2014).
There is no specific data on transportation and waste recovery, so we only
made a model and did not do a detailed calculation in the example for this part.
However, the carbon emission model can be calculated more accurately according
to the specific photovoltaic power generation system. At the same time, we can
adjust the production part of the carbon emission model in this article and keep the
other parts unchanged to calculate the carbon emissions of low carbon energy
 ACCEPTED MANUSCRIPT

power generation systems, such as wind power and hydropower. Therefore, the
carbon emission model in this article has a certain popularity.
Due to China strongly supporting the photovoltaic industry over the past few
years, a large number of photovoltaic generating units emerged, and the problems
of subsidy gaps and low production technology levels also appeared. With the
development of low-carbon energy in China, the concept of the energy internet has
also begun to receive attention. To promote the transformation of the photovoltaic
industry from scale growth to high-quality development, it is essential to better
integrate photovoltaics into the energy internet and develop them in synergy with
various energy sources. To achieve the maximum degree of emissions reduction,
the National Development and Reform Commission, the Ministry of Finance, and
the National Energy Administration jointly issued the "Notice on Matters Related
to Photovoltaic Power Generation in 2018" on May 31, 2018, proposing that the
2018 ordinary PV power plant indicators should not be arranged temporarily and
strictly control the scale of distributed photovoltaics, to reduce the subsidy
intensity and to uniformly reduce the on-grid tariffs of newly launched projects by
0.05 yuan/kWh. This policy is also an important measure for implementing
structural reforms on the supply side and promoting high-quality economic
development. It is an important measure for alleviating prominent contradictions
and outstanding problems, such as the subsidy gap and the abandonment of light
and power curtailment currently faced by the photovoltaic industry. The
introduction of the policy indicates that China's photovoltaic industry has gone
through the initial stage of large-scale subsidies. After the early "million roofs"
plan, China now wants to expand photovoltaics to the "PV +" strategic project, not
only on the roofs but also on the roads and other buildings, making full use of
urban solar energy resources, so that the per capita carbon emissions of the city can
be reduced greatly. And PV power plants can be located on marginal lands and
brownfields. They can also be used on higher-quality lands in conjunction with
grazing livestock and crops. According to the examples of Brazil and Germany,
China can consider possible interchange scenarios for production and operation of
the PV systems considering the carbon dioxide intensity of the local electricity
grids and install different kinds of power plants during the operation time of a PV
system that might change the CO2 intensity of the grid (Krauter, 2004). Chinese
enterprises can consider using photovoltaics to conduct better GHG management
that can be more likely to increase value added through an increase in demand
 ACCEPTED MANUSCRIPT

because customers require green products and processes avoiding risks stemming
from less environmentally conscious suppliers (Cucchiella et al., 2017).
5. Conclusion
The main source of carbon emissions in the photovoltaic power supply chain
is the carbon emissions generated during various aspects of the production process.
The data of 1 kW photovoltaic power generation system in this paper are typical
data of photovoltaic power generation system in China. The calculation shows that
the power consumption of this system is about 4522.0368 kWh. Photovoltaic
power generation capacity is affected by many factors, among which the solar
radiation conditions in the installation area are the most important. The 1 kW
power generation system was calculated to have an effective illumination duration
of 3000 h and 1000 h, and the total power generation during the entire life cycle
was 47584.23 kWh and 15861.41 kWh with carbon emission reductions of 23.684
t and 13.374 t, respectively. At the same time, we calculated that the carbon
recovery period of China's photovoltaic power generation system as between 2 and
6 years in different average annual effective illumination duration from 1000 h to
3000 h. The solar radiation level in the installation area of visible photovoltaic
power generation system has a significant impact on its carbon emission reduction
and carbon recovery effect. Recent trends show that releases of these gases are on
the decline, which may be attributed to more efficient manufacturing processes and
the use of alternative substances. Cadmium is a natural by-product of zinc mining,
and studies suggest that an environmentally friendly means to sequester elemental
cadmium is to use it in PV modules. Decommissioning waste for PV is the stage
expected to result in the largest environmental impact when PV is evaluated on a
full life cycle basis. That said, recycling of spent PV modules has now begun and
has shown potential to improve the environmental profile of PV technologies
(Environment Canada, 2012). Thus, the carbon emissions of photovoltaic power
supply chain can be reduced effectively by improving raw material development
technology to provide high development efficiency, optimizing parts production
process, gradually replacing transport vehicles with electric vehicles, improving
waste recycling efficiency and waste disposal methods. By calculating the carbon
emissions of the photovoltaic supply chain, we hope that all policymakers will be
made aware of the importance of reducing emissions and make more policies
focused on the continued reduction of carbon emissions.
 ACCEPTED MANUSCRIPT

Acknowledgments
Project supported by the National Social Science Fund of China (17BGL136)
References
Arvesen A, Hauan I B, Bolsøy B M, et al., 2015. Life cycle assessment of transport of electricity via different

voltage levels: A case study for Nord-Trøndelag county in Norway[J]. Applied Energy. 157:144-151.

Barthelmie R J, Morris S D, Schechter P, 2008. Carbon neutral Biggar: calculating the community carbon footprint

and renewable energy options for footprint reduction[J]. Sustainability Science. 3(2): 267-282.
Bazán J, Rieradevall J, Gabarrell X, et al., 2018. Low-carbon electricity production through the implementation of
photovoltaic panels in rooftops in urban environments: A case study for three cities in Peru[J]. Science of the

Total Environment. 622:1448-1462.

Boamah K B, Du J, Boamah A J, et al., 2018. A study on the causal effect of urban population growth and

international trade on environmental pollution: evidence from China[J]. Environmental Science & Pollution

Research. 25(6):1-13.

Cansino J M, Román R, Ordóñez M, 2016. Main drivers of changes in CO 2, emissions in the Spanish economy: A

structural decomposition analysis[J]. Energy Policy. 89:150-159.

Carol, 2017. Efficiency of PV grid connected power station system.

https://news.solarbe.com/201701/26/108278.html (accessed 14 May 2018)

Chen, Z. and S.I. Su, Multiple competing photovoltaic supply chains: Modeling, analyses and policies. JOURNAL

OF CLEANER PRODUCTION, 2018. 174: p. 1274-1287.

Chen Zhi-song, Wang Hui-min, Ma Ping, 2011. Competition, Cooperation and Coordination of Polysilicon

Photovoltaic Supply Chain [J]. Chinese Journal of Management Science. 19(1):55-61.

Cucchiella F, D'Adamo I, Koh S C L, 2015. Environmental and economic analysis of building integrated

photovoltaic systems in Italian regions[J]. Journal of Cleaner Production. 98:241-252.

Cucchiella, F., M. Gastaldi and M. Miliacca, The management of greenhouse gas emissions and its effects on firm

performance. JOURNAL OF CLEANER PRODUCTION, 2017. 167: p. 1387-1400.

Dai Y, Fang S C, Ling X, et al., 2008. Risk pooling strategy in a multi-echelon supply chain with price-sensitive

demand[J]. Mathematical Methods of Operations Research. 67(3):391-421.

Dong Hui-juan, Geng Yong, 2012. Study on Carbon Footprint of the Household Consumption in Beijing Based on

Input-Output Analysis [J]. Resources Science. 34(3):494-501.

Energy Research Institute of China National Development and Reform Commission, 2003. Scenario analysis of

sustainable energy and carbon emissions in China [R]. https://www.docin.com/p-78064180.html (accessed 6

March 2019)
 ACCEPTED MANUSCRIPT

Environment Canada, 2012, Assessment of the Environmental Performance of Solar Photovoltaic Technologies: a

report funded under the Clean Energy Fund, Cat. No.: En84-88/2012E-PDF, Ottawa - Ontario, ISBN

978-1-100-21269-2

Gao Ge, 2018. Total volume of PV installed in China in 2017.

http://guangfu.bjx.com.cn/news/20180126/876884.shtml/ (accessed 14 May 2018)

Gong Dao-ren, Chen Di, Yuan Zhi-zhong, 2013. Mathematics calculation model and application of CO2 emission of

photovoltaic(PV) power generation system [J]. Renewable Energy Resources. 31(9):1-4.

Government report, 2017. An overview of the development of China's photovoltaic industry in 2017

http://www.china-nengyuan.com/news/122016.html/ (accessed 25 March 2018)

IPCC, 2001. Climate Change 2001: The Scientific Basis[M]. Houghton J T, Ding Y, Griggs D J, et al.,

Eds. Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press, 881.

IPCC, 2006. IPCC Guidelines for National Greenhouse Gas Inventories[R]. Prepared by the National Greenhouse

Gas Inventories Programme, EgglestonHS, Buendia L, Miwa K, Nhara T and Tanabe K(eds).

ISO14064-1-2006, Greenhousegases-Part1: Specification with guidance at the organization level for quantification

and reporting of greenhouse gas emissions and removals [S].

ISO14067. https://baike.baidu.com/item/ISO%2014067/4951575?fr=aladdin (accessed 8 January 2019)

Jin, C., et al., Self-driving tourism induced carbon emission flows and its determinants in well-developed regions: A

case study of Jiangsu Province, China. JOURNAL OF CLEANER PRODUCTION, 2018. 186: p. 191-202.

Khan A Q, Saleem N, Fatima S T, 2018. Financial development, income inequality, and CO2 emissions in Asian

countries using STIRPAT model.[J]. Environmental Science & Pollution Research International. 25(7):6308.

Laine J, Ottelin J, Heinonen J, et al., 2017. Consequential Implications of Municipal Energy System on City Carbon

Footprints[J]. Sustainability. 9(10):1801.

Larsen H N, Hertwich E G, 2009. The case for consumption-based accounting of greenhouse gas emissions to

promote local climate action[J]. Environmental Science & Policy. 12(7): 791-798.

Lee B, 2017. Energy Efficiency Gain of Cellular Base Stations with Large-Scale Antenna Systems for Green

Information and Communication Technology[J]. Sustainability. 9(7):1123.

Liu Yun, Shi Hua-ding, Zeng Xian-gang, 2011. A Life-Cycle Carbon Footprint Assessment of Electric Power

Companies [J]. RESOURCES SCIENCE. 33(4):653-658.

Lombardi, M., et al., 2018. Toward urban environmental sustainability: The carbon footprint of Foggia's

municipality. JOURNAL OF CLEANER PRODUCTION. 186: p. 534-543.

Marigo N, 2007. The Chinese silicon photovoltaic industry and market: a critical review of trends and outlook[J].

Progress in Photovoltaics Research & Applications. 15(2):143-162.

Mariska J. De Wild-Scholten. Energy payback time and carbon footprint of commercial photovoltaic systems[J].

Solar Energy Materials and Solar Cells, 2013, 119(Complete):296-305.

 ACCEPTED MANUSCRIPT

National Bureau of Statistics: Manual of energy statistics.

https://wenku.baidu.com/view/7e8553553b3567ec102d8adc.html/ (accessed 6 March 2018)

Nikolaos Stylos, Christopher Koroneos, 2014, Carbon footprint of polycrystalline photovoltaic systems, Journal of

Cleaner Production Volume 64, 1 February 2014, Pages 639-645, https://doi.org/10.1016/j.jclepro.2013.10.014

Prehoda E W, Pearce J M, 2017. Potential lives saved by replacing coal with solar photovoltaic electricity

production in the U.S[J]. Renewable & Sustainable Energy Reviews. 80:710-715.

Reich N H , et al. CO2 Emissions of PV in the Perspective of a Renewable Energy Economy[C]// WIP-Renewable

Energies, 2007.

Schmidt M, 2009. Carbon accounting and carbon footprint-more than just diced results? [J]. International Journal of

Climate Change Strategies and Management. 1(1): 19-30.

Singh P, Kansal A, Carliell-Marquet C, 2016. Energy and carbon footprints of sewage treatment methods[J]. Journal

of Environmental Management. 165:22-30.

Su Qing-zhong, Yuan Jun-bin, 2009. To Evaluate on Contribution Degree and Relation Degree for Nodical

Enterprises in the Supply Chain [J]. Value Engineering. 28(11):60-62.

Sukumaran S, Sudhakar K, 2017. Fully solar powered airport: A case study of Cochin International airport[J].

Journal of Air Transport Management. 62:176-188.

Sun Jing-xiang, Cai Xin, Xu Qing-chuan, 2014. Research review on the risk problems of photovoltaic supply chain

[J]. Journal of Xi'an Jiaotong University (Social Sciences). 34(2):37-42.

Suo, C., et al., Identifying optimal clean-production pattern for energy systems under uncertainty through

introducing carbon emission trading and green certificate schemes. JOURNAL OF CLEANER

PRODUCTION, 2017. 161: p. 299-316.

Wang Yong-qin, Zhou Ye, Zhang Rong, 2017. A Review of The Research Methods on Influencing Factors of

Carbon Emissions and Carbon Footprint [J]. Environmental Engineering. 35(1):155-159.

Wiedmann T, Minx J, 2008. A definition of ‘carbon footprint’[J]. Ecological economics research trends. 1: 1-11.

Zhan Xiao-yan, 2011. Life Cycle Research on Carbon Emissions of Polycrystalline silicon-Photovoltaic System [D].

Yangzhou University.

Zhao Xian-gui, Xiao Ling, Ma Cai-hong, et al., 2014. Dynamic analysis of carbon footprint and evaluation of

carbon emission degree in Shanxi province [J]. Journal of Arid Land Resources and Environment. 28(9):21-26.

Zhejiang Energy Conservation, 2016 Circular of the State Council on Issuing the Work Program for Greenhouse

Gas Emission Control during the 13th Five-Year Plan - "Work Program for Greenhouse Gas Emission Control

during the 13th Five-Year Plan" [J]. (4): 1-6.

http://www.gov.cn/zhengce/content/2016-11/04/content_5128619.htm (accessed 9 June 2019)

 ACCEPTED MANUSCRIPT

Highlights
•Calculate carbon emissions from photovoltaic supply chain with life cycle
method.
•Photovoltaic carbon recovery period is 2 years when average annual beam
is 3000 h.
•In 2017, photovoltaic carbon emission reduction in China is 1.738 GT -
3.079 GT.
• Increasing waste recovery can effectively reduce photovoltaic carbon
emissions.