SCHOOL OF

INDUSTRIAL AND INFORMATION ENGINEERING

Master of Science in

Management Engineering

Manufacturer’s Information Value in De-manufacturing

Oriented Circular Economy

Supervisor: Prof. Marcello Colledani

Co-Supervisor: Dott. Francesco Baiguera

Authors:

Omid Nejat Matr. 835511

Masoud Esnaashari Esfahani Matr. 863998

Academic Year 2017 - 2018

P a g e | 2

Acknowledgment

We would like to express our deep gratitude to Prof. Colledani and Dr. Baiguera,

our thesis supervisors, for their patient guidance, enthusiastic encouragement and

useful critiques of this work.

We would also like to thank Eng. Pinci from Italtel Company, for his advice and

assistance for supporting us by providing needed data. Finally, we wish to thank

our parents for their support and encouragement throughout our study.

P a g e | 3

Table of Contents

Acknowledgment .................................................................................................. 2

List of figures ........................................................................................................ 5

List of tables .......................................................................................................... 8

List of Abbreviations ........................................................................................... 10

Abstract ............................................................................................................... 12

1 Introduction .................................................................................................. 13

1.1 Definition, Conceptualizations, and Definitions of the CE ..................... 13

1.2 Opportunities in moving towards CE and its relation to sustainable

business ................................................................................................... 13

1.3 CE framework and its successes and challenges ..................................... 16

1.4 The circular economy of electronic and electrical equipment ................ 20

1.5 Manufacturing and De-manufacturing in the circular economy ............. 24

1.6 The role of information in the circular economy .................................... 25

1.7 Problem statement and motivation of the thesis ...................................... 29

1.8 Thesis outline .......................................................................................... 30

2 State of the art in PCB manufacturing and de-manufacturing ..................... 32

2.1 PCB Structure .......................................................................................... 32

2.2 Manufacturing of PCB ............................................................................ 37

2.3 De-manufacturing framework and its implication on EOL PCBs .......... 43

2.3.1 Disassembly in de-manufacturing .................................................. 44

2.3.2 Remanufacturing in de-manufacturing .......................................... 52

2.3.3 Recycling and recovery processes ................................................. 57

2.4 The flow of needed information in de-manufacturing ............................ 64

3 Defining and modeling an economic model in optimizing PCB De-

manufacturing ................................................................................................. 73

3.1 Market analysis........................................................................................ 73

3.2 General PCB treating work description in de-manufacturing ................. 79

3.2.1 Main PCB’s treatment flows in the work description .................... 98

3.2.2 The need for a collected database about each PCB ..................... 101

3.2.3 The structure of the database and its different sections ............... 102

3.3 Economic model .................................................................................... 106

4 Case study and application of the model on the case ................................. 113

4.1 Italtel and Italtel’s boards ...................................................................... 113

P a g e | 4

4.2 Preliminary analysis and the review of the work which had been done on

the data gathering of the PCBs .............................................................. 114

4.3 Data gathering and data structure .......................................................... 117

4.4 Economic model application ................................................................. 124

4.5 Results ................................................................................................... 140

4.6 Sensitivity analysis ................................................................................ 153

5 Conclusion ................................................................................................. 159

6 Bibliography ............................................................................................... 161

P a g e | 5

List of figures

Figure 1. 1 Scheme for the circular economy ..................................................... 14

Figure 1. 2 Comprehensive CE framework ......................................................... 16

Figure 1. 3 Closed circular economy loop .......................................................... 19

Figure 1. 4 Illegal WEEE treatment and health concerns ................................... 21

Figure 1. 5 Waste PCBs from all kind of electronic equipment ......................... 23

Figure 1. 6 Manufacturer-centric Circular Economy framework ....................... 27

Figure 1. 7 Information and good’s flow in the circular economy ..................... 28

Figure 2. 1 A Printed Circuit Board with various high value components ......... 32

Figure 2. 2 Single-layer PCB Diagram ............................................................... 33

Figure 2. 3 Double-layer PCB Diagram .............................................................. 34

Figure 2. 4 Multi-layer PCB Diagram ................................................................. 34

Figure 2. 5 Different PCB Classifications ........................................................... 35

Figure 2. 6 A part of PCB Manufacturing line .................................................... 37

Figure 2. 7 Making the Substrate ........................................................................ 38

Figure 2. 8 Component Assembly ....................................................................... 39

Figure 2. 9 Inkjet PCB pattern printing on a substrate ........................................ 40

Figure 2. 10 Through Hole Technology & Surface Mount Technology ............. 42

Figure 2. 11 Repairing or Recycling the PCB ..................................................... 43

Figure 2. 12 Products and materials Life Cycle .................................................. 44

Figure 2. 13 Infrared heating application ............................................................ 47

Figure 2. 14 Structure schematic diagram of the experiment facility ................. 51

Figure 2. 15 PCB Inspection ............................................................................... 54

Figure 2. 16 PCB cleaning with degreaser spray ................................................ 55

Figure 2. 17 Reconditioning the PCB to be reused ............................................. 56

Figure 2. 18 Single shaft shear shredder, and Chamber of a cutting mill ........... 59

Figure 2. 19 density-based material separation ................................................... 60

Figure 2. 20 Electrostatic material separation ..................................................... 61

Figure 2. 21 Eddy current material separation .................................................... 62

Figure 2. 22 Current product flow ....................................................................... 65

Figure 2. 23 Future centralized information flow for product environmental

profile .................................................................................................................. 66

P a g e | 6

Figure 2. 24 Future decentralized information flow for product environmental

profile in which information is embedded in the product ................................... 67

Figure 2. 25 Product Lifecycle ............................................................................ 68

Figure 2. 26 1D and 2D Barcode Labels ............................................................. 69

Figure 2. 27 Contact Memory Buttons ................................................................ 70

Figure 2. 28 RFID Label ..................................................................................... 70

Figure 3. 1 Trends in global PCB production [Prismark(2012)] ........................ 74

Figure 3. 2 PCB Production in various countries ................................................ 74

Figure 3. 3 PCB Production and Market price index 2007-2017 ........................ 76

Figure 3. 4 The work description of PCB treatment’s possible streams ............. 82

Figure 3. 5 An Automatic PCB Testing machine in STIIMA-CNR ................... 88

Figure 3. 6 Before and after PCB cleaning ......................................................... 91

Figure 3. 7 Automatic PCB component disassembly station at STIIMA-CNR .. 92

Figure 3. 8 Before and after simultaneous disassembly of the PCB .................. 95

Figure 3. 9 Board’s partial cleaning and preparation for assembly .................... 97

Figure 3. 10 The main assumed flows of the waste PCBs treatment .................. 98

Figure 4. 1 The sample of Italtel’s User-interface and Power-supply PCBs .... 114

Figure 4. 2 PCB’s analyzed image by the software .......................................... 116

Figure 4. 3 Italtel PCB S7338-L6030 ............................................................... 118

Figure 4. 4 Front and rear of an Italtel PCB code S7338-L6-30-A2-A1 .......... 119

Figure 4. 5 The PCB’s stream allocation, regarding the 8th, 9th, and 10th streams

........................................................................................................................... 130

Figure 4. 6 The PCB’s Stream Allocation, Regarding the 1st, 2nd, and 3rd Streams

........................................................................................................................... 134

Figure 4. 7 The PCB’s Stream Allocation, Regarding the 4th, 5th, 6th, and 7th

Streams .............................................................................................................. 139

Figure 4. 8 Italtel’s Power-supply PCB S7201-X6001-U4-A1 ........................ 141

Figure 4. 9 Treatment streams allocation for (PCB S7201-X6001) .................. 142

Figure 4. 10 The treatment streams’ associated profit (PCB S7201-X6001) ... 143

Figure 4. 11 Italtel’s Power-supply PCB S7338-L6030-U2-A2 ....................... 144

Figure 4. 12 Treatment streams allocation for (PCB S7338-L6030) ................ 145

Figure 4. 13 The treatment streams’ associated profit (PCB S7338-L6030) .... 146

Figure 4. 14 Italtel’s Power-supply PCB S7510-L6023-U1-A3 ....................... 146

Figure 4. 15 Treatment streams allocation for (PCB S7510-L6023) ................ 147

Figure 4. 16 The treatment streams’ associated profit (PCB S7510-L6023) .... 148

P a g e | 7

Figure 4. 17 Italtel’s Power-supply PCB S7510-L6047-U1-A3 ....................... 149

Figure 4. 18 Treatment streams allocation for (PCB S7510-L6047) ................ 150

Figure 4. 19 The treatment streams’ associated profit (PCB S7510-L6047) .... 151

Figure 4. 20 Profit’s sensitivity analysis regarding the PCB’s condition-related

variables ............................................................................................................ 154

Figure 4. 21 Profit’s sensitivity analysis regarding the market-related variables

........................................................................................................................... 156

Figure 4. 22 Profit’s sensitivity analysis regarding the provided-data level ..... 157

P a g e | 8

List of tables

Table 2. 1 Comparisons between different heating methods for de-soldering ... 52

Table 2. 2 Comparison of ID Technologies ........................................................ 71

Table 3. 1 Mobile phone PCB price (New and Used PCB) ................................ 77

Table 3. 2 Computer motherboard PCB price (New and Used PCB) ................. 78

Table 3. 3 Electronic Package price .................................................................... 78

Table 3. 4 List of operating and decision-making steps’ abbreviation ....... Error!

Bookmark not defined.

Table 3. 5 The needed data for the “Board Inspection” step .............................. 83

Table 3. 6 The needed data for the “Image Process” step ................................... 85

Table 3. 7 The needed data for the “1st Cost Analysis” step ............................... 87

Table 3. 8 The needed data for the “PCB Testing” step ..................................... 88

Table 3. 9 The needed data for the “2nd Cost Analysis” step .............................. 90

Table 3. 10 The needed data for the “Board Final Cleaning” step ..................... 91

Table 3. 11 The needed data for the “Selective Disassembly of Intact

Components” step ............................................................................................... 93

Table 3. 12 The needed data for the “Selective Destructive Disassembly of HCM

Components” step ............................................................................................... 94

Table 3. 13 The needed data for the “Simultaneous Disassembly” step ............. 95

Table 3. 14 The needed data for the “Component Quality Control” step ........... 96

Table 3. 15 The needed data for the “Selective Disassembly of Defective

Components” step ............................................................................................... 96

Table 3. 16 The needed data for the “Board’s Partial Cleaning” step ................ 97

Table 3. 17 The needed data from different players and the output data from the

............................................................................................................................. 98

Table 3. 18 Package Peak Temperature for SnPb Solder Materials [37] .......... 105

Table 3. 19 Package Peak Temperature for Pb-free Solder Materials [37] ...... 105

Table 4. 1 Sample of the database regarding the PCB S7338-L6030 ............... 119

Table 4. 2 Dimension and weight of the Italtel PCB S7338-L6030 ................. 120

Table 4. 3 The Production related data of the Italtel PCB S7338-L6030 ......... 120

Table 4. 4 Deviation of the component codes for the same component ........... 121

Table 4. 5 Sample of Italtel PCB S7338-L6030 database, as the component

introduction data ................................................................................................ 121

P a g e | 9

Table 4. 6 The data regarding the De-soldering and Picking tasks ................... 122

Table 4. 7 The formative material content of the components ......................... 123

Table 4. 8 A part of the database regarding the components’ places ................ 124

Table 4. 9 The conditional coefficient and assumed data for the 1st condition of

PCB S7338-L6030 ............................................................................................ 127

Table 4. 10 The conditional coefficient and assumed data for the 2nd condition of

PCB S7338-L6030 ............................................................................................ 127

Table 4. 11 The conditional coefficient and assumed data for the 3nd condition of

PCB S7338-L6030 ............................................................................................ 127

Table 4. 12 Assumed Constant Cost Related Factors ....................................... 128

Table 4. 13 Assumed Revenue Related Factors for PCB S7338-L6030 ........... 128

Table 4. 14 Cost and Revenue Influential Factors for PCB S7338-L6030 ....... 128

Table 4. 15 The 8th, 9th, and 10th treatment streams’ related revenue, cost, and

profit comparison .............................................................................................. 129

Table 4. 16 The conditional coefficient and assumed data for the 4th condition of

PCB S7338-L6030 ............................................................................................ 131

Table 4. 17 The conditional coefficient and assumed data for the 5th condition of

PCB S7338-L6030 ............................................................................................ 131

Table 4. 18 The conditional coefficient and assumed data for the 6th condition of

PCB S7338-L6030 ............................................................................................ 132

Table 4. 19 Assumed constant cost related factors ........................................... 132

Table 4. 20 Assumed revenue relate factors for PCB S7338-L6030 ................ 132

Table 4. 21 Cost and revenue influential factors for PCB S7338-L6030 ......... 132

Table 4. 22 The 1st, 2nd, and 3rd treatment streams’ related revenue, cost, and profit

comparison ........................................................................................................ 133

Table 4. 23 The conditional coefficient and assumed data for the 7th condition of

PCB S7338-L6030 ............................................................................................ 135

Table 4. 24 The conditional coefficient and assumed data for the 8th condition of

PCB S7338-L6030 ............................................................................................ 136

Table 4. 25 The conditional coefficient and assumed data for the 9th condition of

PCB S7338-L6030 ............................................................................................ 136

Table 4. 26 The conditional coefficient and assumed data for the 10th condition of

PCB S7338-L6030 ............................................................................................ 136

Table 4. 27 Assumed Constant Cost Related Factors ....................................... 137

Table 4. 28 Cost and Revenue Influential Factors for PCB S7338-L6030 ....... 137

Table 4. 29 The 4th, 5th, 6th, and 7th treatment streams’ related revenue, cost, and

profit comparison .............................................................................................. 138

P a g e | 10

Table 4. 30 Information provided de-manufacturing profit vs. recycling profit

comparison ........................................................................................................ 152

List of Abbreviations

1st BT First Board Test

1st CA First Cost Analysis

1st CARB First Cost Analysis Result for Board Reuse

1st CARC First Cost Analysis Result for Component Reuse

1st CARR First Cost Analysis Result for Recycling

2nd BT Second Board Test

2nd CA Second Cost Analysis

2nd CARB Second Cost Analysis Result for Board Reuse

2nd CARC Second Cost Analysis Result for Component Reuse

2nd CARR Second Cost Analysis Result for Recycling

AIDC Automated Identification and Data Capture

ANC Assembly of New Components

BBR Bare Board Recycling

BBV Bare Board Value

BDS Board Demand Status

BFC Board Final Cleaning

BGC Board General Cleaning

BI Board Inspection

BINT First Board Result

BIR Board Inspection Result

BP Board Packaging

BPC Board Partial Cleaning

BRR Board Repair Result

CE Circular Economy

CMV Component Material Value

CP Component Packaging

CR Component Recycling

CQC Component Quality Control

DIP Dual In-line Package

ELV End of Life Vehicle

EOL End of Life

GDP Gross Domestic Product

HCM High Concentrated Material

P a g e | 11

HCMCV High Concentrated Material Components Value

IC Integrated Circuit

IP Image Process

LW Labor Wage

massQCfailedC Mass of Quality Control Failed Components

massDDC Mass of Disassemble Defective Components

MCC Material Content of Component

MEANCMV Mean Component Material Value

MP Material Price

NDC Number of Defective Components

NIC Number of Intact Components

NPDC Number of Present Demanded Components

NQCpassedC Number of Quality Control Passed Components

PCB Printed Circuit Board

POB Price of Board

POC Price of Component

PTH Pin Through Hole

QFP Quad Flat Package

RFID Radio Frequency Identification

SDDHCMC Selective Destructive Disassembly of High Concentrated Material Component

SDIC Selective Disassembly of Intact Component

SMT Surface Mount Technology

SOIC Small Outline Integrated Circuit

THT Through Hole Technology

WEEE Waste of Electronic and Electrical Equipment

P a g e | 12

Abstract

With the sharp slop of the technology growth and presenting the new models,

versions, and technologically improved products, we are observing a new trend

of products’ usage, which means the products’ life in the consumption phase is

becoming shorter than before. With a high-level of production rate, and product’s

short life, it can be understood that we are facing a huge number of end-of-life

products. The concept of circular economy has been generated to cope with this

problem. It tries to make a closed loop of manufacturing, usage, end returning the

End-of-Life (EOL) products to the first point of manufacturing state to be used as

a resource of raw materials or even be reused. The feasibility of the circular

economy fulfilment in each kind of industry, is limited to some constraints, such

as the lack of needed information and the proper way of exploiting the relevant

information. In order to close the loop of circular economy, there should be some

motivations for different influential parties who are present in the whole life-cycle

of a product in circular economy concept. The most powerful motivation for the

main players who are able to either provide or exploit the information, is the

economic value which could be achieved by providing or exploiting it. Hence the

aim of this thesis is to assess what the needed data and information are that would

be beneficial for treating the EOL products and also how they could be exploited

in a way that the manufacturers as the data providers and the de-manufacturers or

recyclers as the data and information users, could take the advantage of and

generate a margin from?

The work has been done by focusing on the Printed Circuit Boards (PCBs) as one

of the most important products in the electric and electronic industry. The PCB

manufacturing and EOL PCB’s possible de-manufacturing procedures have been

studied and as a result, all the needed data and information regarding their

treatments are presented. As far as the data exploitation is concerned, a general

algorithm related to EOL PCB’s possible treatments is designed to optimize the

EOL PCB’s treatment and based on which, an economic model is designed. It

firstly suggests the most profitable procedure to treat the different PCBs with

different EOL conditions.

Consequently, as a result of evaluating a big number of various conditions of a

few types of PCBs by the “Economic Model”, the generated margin of treating

the EOL PCBs with the presence of data is compared with the basic treatment

(recycling) of PCBs without the presence of any data.

P a g e | 13

1 Introduction

1.1 Definition, Conceptualizations, and Definitions of the

CE

The term circular economy (CE) has both a linguistic and a descriptive meaning.

Linguistically it is an antonym of a linear economy. A linear economy is one

defined as converting natural resources into waste, via production. Such

production of waste leads to the deterioration of the environment in two ways: by

the removal of natural capital from the environment (through

mining/unsustainable harvesting) and by the reduction of the value of natural

capital caused by pollution from waste. Pollution can also occur at the resource

acquisition stage. This is a one-way system and an economy based on such a

system has been referred to as a cowboy economy by Boulding (1966).

The term linear economy was brought into popular use by those writing on the

Circular Economy and related concepts. Thus, in many ways, the origin has been

deliberately set, in framing the antonym, to promote the term circular economy.

By circular, an economy is envisaged as having no net effect on the environment;

rather it restores any damage done in resource acquisition while ensuring little

waste is generated throughout the production process and in the life history of the

product. [1]

1.2 Opportunities in moving towards CE and its relation

to sustainable business

In recent years, the circular economy has figured prominently in political,

economic, and business dialogues. But the concept remains eclectic and lacks a

scientifically endorsed definition. For this economic analysis, the circular

economy is defined as an economy that provides multiple value-creation

mechanisms which are decoupled from the consumption of finite resources

(Figure 1.1). This definition rests on three principles:

• Preserve and enhance natural capital by controlling finite stocks and

balancing renewable resource flows – for example, replacing fossil fuels with

renewable energy or returning nutrients to ecosystems.

P a g e | 14

• Optimize resource yields by circulating products, components, and materials

in use at the highest utility at all times in both technical and biological cycles –

for example, sharing or looping products and extending product lifetimes.

• Foster system effectiveness by revealing and designing out negative

externalities, such as water, air, soil, and noise pollution; climate change; toxins;

congestion; and negative health effects related to resource use.

Figure 1. 1 Scheme for the circular economy

Narrower notions of the circular economy, limited to material reuse and

sometimes regeneration, exist. But the modern economy requires applying all

three principles to reintegrate the economy into our planet’s system, which is the

ultimate ambition of circular thinking. Thus, applying these principles means

creating an economy that is restorative and regenerative, that preserves

ecosystems and increases their return over time, that creates prosperity, and that

fuels growth by capturing more value from existing infrastructure and products.

System change is also crucial to the circular economy. In a circular economy, one

system’s waste is the next system’s input, and the aim is to maximize total utility

from the products and materials in use. This requires taking a system view of large

value chains better-performing resources.

P a g e | 15

A circular economy also enhances natural capital by encouraging flows of

nutrients within the system and creating the conditions for regeneration of soil.

The second pillar is to optimize resource yields, "by circulating products,

components, and materials at the highest utility at all times in both technical and

biological cycles." This means designing for remanufacturing,

refurbishing, and recycling to keep components and materials circulating in the

economy, creating positive effects. A circular system uses inner loops whenever

they preserve energy and other value, such as embedded labor. The system should

also extend product life and optimize reuse, gaining in turn increased resource

utilization. Circular systems also maximize use of waste materials, extracting

valuable biochemical stocks and cascading them into different low-grade

applications.

Finally, the last action is to Foster system effectiveness "by revealing and

designing out negative externalities." This includes reducing damage to human

utility, such as food, mobility, shelter, education, health, and entertainment, and

managing externalities, such as land use, air, water and noise pollution, the release

of toxic substances, and climate change.

These principles define the line of actions to create a circular economy, entailing

major and difficult changes of current society, which still delay the proposed

switch. [2]

The relation of the CE to sustainable business

A true circular economy would demonstrate new concepts of system, economy,

value, production, and consumption, leading to sustainable development of the

economy, environment, and society. The ultimate objective of this approach

would be to achieve the decoupling of economic growth from natural resource

depletion and environmental degradation. As such, the Circular Economy might

be thought of as a general term covering all activities that reduce, reuse, and

recycle materials in production, distribution, and consumption processes. Murray

resource (2007) describes the Circular Economy as a mode of economic

development based on the ecological circulation of natural materials, requiring

compliance with ecological laws and sound utilization of natural resources to

achieve economic development. He also, explains that there is a feedback process

of product renewed resource, and that the ultimate objectives of optimum

production, optimized consumption, and minimum waste can be achieved in

production. Murray resource (2011) stresses that the focus of the Circular

Economy is on resource productivity and eco-efficiency improvement, and they

adopt the 4R approach: reduce, reuse, recycle and recover. [1]

P a g e | 16

1.3 CE framework and its successes and challenges

A comprehensive framework of CE

CE perspective resource scarcity, environmental impact, and economic benefits

emphasize a combined view. Figure1.2 illustrates a comprehensive framework for

CE based on these three perspectives including their relationships.

Figure 1. 2 Comprehensive CE framework

Economic benefits in CE: Each individual company strives for gaining economic

benefits in order to secure profitability and a competitive edge. This requires an

integrative approach towards business models, product design, supply chain

design and choice of materials.

Resource scarcity in CE: Social prosperity depends on planet earth's finite

resource supplies which makes regenerative use of resources mandatory for CE

realization. The underlying factors in this context concern circularity of resources,

material criticality, and volatility of resources in the light of the globally

increasing number of industrial activities.

P a g e | 17

Environmental impact in CE: A society with minimum environmental impacts

is a desirable state of nations, governmental bodies, and individuals around the

globe. CE aims at reducing solid waste, landfill, and emissions through activities

such as reuse, remanufacturing and/or recycling.

Resource scarcity, environmental impact, and economic benefits cover short-term

as well as long-term objectives and are seamlessly linked in this CE framework.

As each of the three perspectives has a systemic view of specific boundaries the

relationships among the perspectives need to be described:

To gain economic benefits industrial activity fundamentally depends on resources

necessary to perform manufacturing operations and transform raw materials into

products. In return, resource price volatility and supply risks have a direct

influence on the competitive edge of companies and their capability of performing

their industrial activity in sustainable and profitable manner.

At the same time, while pursuing economic benefits, industrial activity influences

the natural environment, as e.g. through waste generation in a linear system.

Perceiving end-of-life products as resources rather than waste, involves

management of resource values as part of standard business operations. The

approach of resource value management stands in contrast to the conventional

“waste management” approach in the prevailing linear economy and does not

differentiate between “waste” and “resources” (as nature does not distinguish

between these two either). In response to waste generation and negative influences

on the natural environment industrial activity is restricted by legislation through

e.g. directives, thus putting constraints and affecting competitiveness.

However, in this scenario legislation is not a burden.

In a CE the speed of resource depletion and waste generation are reduced.

Assuming a rapidly growing population the speed of resource depletion will be

greater in a linear economy than in a CE. Similarly, the speed of waste generation

will be higher in a linear economy than in a CE.

Circular economy successes and challenges

The CE concept has influenced policy and innovation in some of the World's

largest economies such as China, Germany, Japan, and the UK. Whereas many

new CE-related projects fail, others have operated for decades, for example, in

China and Denmark. In some cases, there is an opportunity to learn from the

projects that succeed as well as those that fail. Appropriate policy instruments

contribute to the success, innovation and network synergies that help stakeholders

P a g e | 18

to meet the multiple objectives or environmental, economic, societal/managerial,

and topological challenges of CE-related initiatives. A policy that supports

standardized use and recycling of products (or materials) is required to encourage

industries to adopt the CE concept.

Specific value chains, material flows, and products need to be assessed to show

the value of applying the CE concept. There are potential barriers to product-level

use and reuse in a closed-loop system, including the lack of information about

specific products and the perceived risks associated with refurbishing or reusing

materials like plastics and food wastes. However, there is evidence that the

consumer demand and the market for reused and recycled products are increasing,

and dialogue between procurers and suppliers can further support a business

model for this market that can be sustained.

A key challenge related to the use and reuse of materials (e.g., steel) in an

application of the CE concept is the quality of these materials over time. This

challenge was discussed in the 1990s by Leontief who considered the value of

materials over time after use and reuse using mathematical principles. Leontief

concluded that economic, as well as physical material value, can be estimated

depending on the stakeholder need(s). [3]

A 'circular economy' would turn goods that are at the end of their service life into

resources for others, closing loops in industrial ecosystems and minimizing waste.

It would change economic logic because it replaces production with sufficiency:

reuse what you can, recycle what cannot be reused, repair what is broken and

remanufacture what cannot be repaired. A study of seven European nations found

that a shift to a circular economy would reduce each nation's greenhouse-gas

emissions by up to 70% and grow its workforce by about 4% — the ultimate low-

carbon economy.

Circular-economy business models fall in two groups: as shown in Figure 1.3

those that foster reuse and extend service life through repair, remanufacture,

upgrades and retrofits; and those that turn old goods into as-new resources by

recycling the materials. People of all ages and skills are central to the model.

Ownership gives way to stewardship; consumers become users and creators. The

remanufacturing and repair of old goods, buildings, and infrastructure create

skilled jobs in local workshops. The experiences of workers from the past are

instrumental.

P a g e | 19

Figure 1. 3 Closed circular economy loop

Yet a lack of familiarity and fear of the unknown mean that the circular-economy

idea has been slow to gain traction. As a holistic concept, it collides with the silo

structures of academia, companies, and administrations. For economists who

work with the gross domestic product (GDP), creating wealth by making things

last is the opposite of what they learned in school. GDP measures a financial flow

over a period of time; circular economy preserves physical stocks. But concerns

over resource security, ethics, and safety as well as greenhouse-gas reductions are

shifting our approach to seeing materials as assets to be preserved, rather than

continually consumed.

In the past decade, South Korea, China, and the United States have started

research programs to foster circular economies by boosting remanufacturing and

reuse. Europe is taking baby steps. The Swedish Foundation for Strategic

P a g e | 20

Environmental Research (Mistra) and the EU Horizon 2020 program published

their first call for circular-economy proposals in 2014. Since 2010, the Ellen

MacArthur Foundation, founded by the round-the-world yachtswoman, has been

boosting awareness of the idea in manufacturers and policymakers. And circular-

economy concepts have been successfully applied on small scales since the 1990s

in eco-industrial parks such as the Kalundborg Symbiosis in Denmark, and in

companies that include Xerox (selling modular goods as services), Caterpillar

(remanufacturing used diesel engines) and USM Modular Furniture. Selling

services rather than goods is familiar in hotels and in public transport; it needs to

become mainstream in the consumer realm. [4]

1.4 The circular economy of electronic and electrical

equipment

Definition of electronic waste

E-waste includes IT and telecommunications items including computers,

entertainment electronics, mobile phones and also supporting ancillary equipment

that is no longer of use to the original consumer. E-waste is a significant secondary

resource because of its suitability for direct reuse, refurbishment and material

recycling of its constituent raw materials (including computer chips, plastics, and

precious metals).

The EU (and emergent China) waste electronic and electrical equipment (WEEE)

directive defines the following IT and information systems related items and

ancillary equipment as potential e-waste (under Annex A of EU Directive

2002/96/EC): IT and telecommunications equipment; consumer equipment;

lighting equipment; toys, leisure and sports equipment; medical devices; and

monitoring and control instruments.

As an example, considering China as one of the main recycling destinations in the

world, it is noteworthy to illustrate the problems associated with electronic waste

in China: E-waste is a valuable source of secondary raw materials when treated

properly.

Improper treatment makes it a major source of toxins and carcinogens. Shortened

life cycles and low cost in the Chinese IT industry have resulted in a growing

problem which requires legal, technical, infrastructural and logistics systems.

These all fit within the CE framework discussed below.

P a g e | 21

Figure 1. 4 Illegal WEEE treatment and health concerns

China’s traditionally lower environmental standards (which have changed in the

past few years) reflects electronic waste being sent for processing – in many cases

illegally. Much of what occurs is due to informal networks of e-waste

organizations who practice uncontrolled burning, disassembly, and disposal.

These activities cause occupational and broader environmental, safety and health

problems because much of this waste is recycled typically using manual and low-

cost hand labor. While trade in waste is regulated and controlled by the Basel

Convention, the rules have been side-lined in China. E-waste is of concern largely

due to the toxicity and carcinogenicity of some of the substances if processed

improperly. These include lead, mercury, cadmium and polychlorinated biphenyls

(PCBs). The non-sustainability of discarding electronics and computer

technology also underscores the need to recycle – or perhaps more practically,

reuse – electronic waste. However, to be able to achieve this, more formalized

systems as recommended by CE are required. The major concerns are summarized

by Collins (2007) in a report by the United Nations University.

Inappropriate handling of systems may lead to:

Emissions of highly toxic dioxins, furans, and polycyclic aromatic

hydrocarbons caused by burning polyvinylchloride (PVC) plastic and wire

insulation.

Soil and water contamination from chemicals such as brominated flame

retardants (used in circuit boards and plastic computer cases, connectors,

and cables); PCBs (in transformers and capacitors); and lead, mercury,

cadmium, zinc, chromium, and other heavy metals (in monitors and other

devices). Studies show rapidly increasing concentrations of these heavy

metals in humans; in sufficient dosages, they can cause neuro-

developmental disorders and possibly cancer.

P a g e | 22

Waste of valuable resources that could be efficiently recovered for a new

product life cycle. [5]

In some processes, hazardous compounds such as dioxins and furans may form

during thermal degradation of waste PCBs. Consequently, environment as well

population surrounding the recycling sites are highly affected. For example, mean

blood lead level (BLL) (15.3±5.79 μg/dL, n=165) of children living in Guiyu, one

of the desired destinations of E-waste in China, is alarmingly higher than mean

BLL (9.94±4.05 μg/dL, n=61) of children living in a neighboring town without

E-waste processing [6]. Besides, soil, water, and air around E-waste processing

sites are 100 times more contaminated by polybrominated diphenyl ethers

(PBDEs), heavy metals and polycyclic aromatic hydrocarbons (PAHs) than other

places. [7], [8], [9]

PCB has an important part in WEEE circular economy

With the development of science and technology, the market demand for

production of electric and electronic equipment (EEE) is increasing rapidly. EEE

has become a necessary part of people’s life. Printed circuit boards (PCBs), an

essential part of almost all EEE, are widely subsistent in EEE. In recent years, the

average rate of world-wide PCBs’ production increases by 8.7%, and this number

is much higher in Southeast Asia (10.8%) and mainland China (14.4%). In

mainland China, the total production value of the PCBs manufacturing industry

has already reached more than $10.83 billion in 2005, only next to Japan, and

would reach more than $12 billion in 2006 by anticipating yields. Meanwhile,

both technological innovation and intense marketing continue to accelerate the

update rate of EEE and shorten the average lifespan of EEE.

As a result, the amounts of waste PCBs are dramatically increasing. The UN

Environment Program estimates that the world generates 20–50 million tons of

waste electric and electronic equipment (WEEE) each year and amounts are rising

three times faster than other forms of municipal waste. Waste PCBs are from all

kinds of electronic equipment as shown in Figure 1.5. [10]

P a g e | 23

Figure 1. 5 Waste PCBs from all kind of electronic equipment

As the environmental effects of waste electrical and electronic equipment

(WEEE) have been attracting the world’s attention for some time, the European

Union (EU) made the first consolidated regional effort to impose regulatory

control on WEEE, the level of awareness about the implications of this type of

waste has risen at an unprecedented pace throughout the world. [11]

A new environmental challenge is presented by waste PCBs, which contain plenty

of toxic substances, such as brominated flame retardants (BFR), PVC plastic and

heavy metals. They can cause serious environmental problems if not proper

disposal. If they are discarded randomly in the opening or landfilled simply, the

leachate may infiltrate into groundwater and soil. Uncontrollable incineration of

waste PCBs also produces potentially hazardous byproducts (including mainly

dioxins, furans, polybrominated organic pollutants and polycyclic aromatic

hydrocarbons) caused by burning BFR, epoxy resins and plastics. The materials

containing BFR are precursors to polybrominated dibenzo-p-dioxins and

dibenzofurans (PBDD/Fs). These are classified as persistent organic pollutants

(POPs) under the Stockholm Convention, a global treaty drawn up to protect

human health and the environment. Growing attention has been given to

hazardous components in waste PCBs, which pose a severe threat to human health

(inducing people’s nervous system diseases or immune system diseases) and the

sustainable economic growth as well. Recycling of waste PCBs is an important

subject not only from the treatment for waste but also from the recovery of

valuable materials. In general, waste PCBs contain approximately 30% metals

P a g e | 24

and 70% nonmetals. The typical metals in PCBs consist of copper (20%), iron

(8%), tin (4%), nickel (2%), lead (2%), zinc (1%), silver (0.2%), gold (0.1%), and

palladium (0.005%). The purity of precious metals in waste PCBs is more than 10

times higher than that of rich-content minerals. So, waste PCBs are considered as

an “urban mineral resources”. In addition, the nonmetal portions of PCBs consist

of thermoset resins and reinforcing materials. They also can be reused as fillers in

composite materials. Therefore, on the basis of the current situation of resources

in China, all materials in waste PCBs are a kind of resources and needed to be

recycled by a proper technology. Resource utilization of waste PCBs can protect

environment and alleviate bottleneck for economic development constrained by

resources shortage in China. [10]

1.5 Manufacturing and De-manufacturing in the

circular economy

However, a sustainable transition to Circular Economy businesses will need to be

supported by fundamental innovations, driven by the manufacturing industry, at

a systemic level, encompassing product design, value-chain integration, business

models, ultimately posing new challenges on the way de-manufacturing and

remanufacturing technologies and systems are conceived and implemented. De-

manufacturing and remanufacturing, briefly indicated as de and re-

manufacturing, are fundamental technical solutions for an efficient and systematic

implementation of the circular economy. More formally, de and re-manufacturing

includes the set of technologies and systems, tools, and knowledge-based methods

to systematically recover, reuse, and upgrade functions and materials from

industrial waste and post-consumer products, to support a sustainable

implementation of manufacturer-centric circular economy businesses.

At technical levels, different business options for “Circular Economy” have been

proposed to generate benefits by exploiting these value-creation mechanisms. For

example, Jawahir in its definition of “Sustainable Manufacturing” proposes the

so-called 6Rs model, where the traditional 3R model based on the Reduce, Reuse,

Recycle practices is enriched with three additional actions namely Recover,

Redesign, and Remanufacture. In practice, the following circular economy

options and the related business models are implemented in the industrial practice:

Reuse: a generic term covering all operations where a returned product is put back

into service, essentially in the same form, with or without repair or remediation.

P a g e | 25

Repair: the correction of specified faults in a product. Repair refers to actions

performed in order to return a product or component purely to a functioning

condition after a failure has been detected, either in service or after discard.

Remanufacturing for function restore: it returns a used product to at least its

original performance with a warranty that is equivalent or better than that of the

newly manufactured product. A remanufactured product fulfills a function similar

to the original part. It is remanufactured using a standardized industrial process,

in line with technical specifications.

Remanufacturing for function upgrade: the process of providing new

functionalities to products through remanufacturing.

Remanufacturing with upgrade aims to extend products’ value life enabling the

introduction of technological innovation into remanufactured products in order to

satisfy evolving customers’ preferences and, at the same time, preserving as much

as possible the physical resources employed in the process.

Closed-loop recycling: recycling of a material can be done indefinitely, without

properties degradation (upcycling). In closed-loop recycling, the inherent

properties of the recycled material are not considerably different from those of the

virgin material, thus substitution is possible.

Open-loop recycling: the conversion of material from one or more products into

a new product, involving degradation in the inherent material properties

(downcycling). In open-loop recycling, the inherent properties of the recycled

material different from those of the virgin material in a way that it is only usable

for other product applications, substituting other materials. These options entail

different value and material preservation levels. While reuse, repair and

remanufacturing constitute product functions and materials conservation

scenarios, recycling offers a material recovery scenario. Their implementation

calls for different de and re-manufacturing systems capabilities. [12]

1.6 The role of information in the circular economy

End of life product’s collection and information management is a critical aspect

of the implementation of the remanufacturing process chain. They are usually

gathered from three different sources, namely internal service centers,

independent service centers, or core dealers.

The development of inline inspection technologies, decision support systems and

disassembly planning tools and also specific solutions for PCB diagnostics,

failure mode analysis and automatic electronics remanufacturing, may contribute

to improve product regeneration rates and decrease remanufacturing costs, thus

P a g e | 26

boosting the competitiveness of industry in this sector. For the selected parts, the

geometry, as well as technical information, materials, and connections to other

parts, are gathered. This information is used to create a disassembly sequence.

Disassembly sequence planning includes a detailed scheduling of the disassembly

tasks and the shop floor control.

In fact, manufacturing and de and re-manufacturing objectives are clearly not

independent. The products that are manufactured and sold to the market today are

the products that will be collected and processed in the input at the de and re-

manufacturing system after the customer use phase. As a consequence, the rapid

introduction of new products and the increasing product variety in manufacturing

is reflected in the continuous evolution of post-use products collected from the

market. In addition, the use-phase introduces variability and uncertainty in the

conditions of these post-use products. Such variability needs to be smoothed by

the de and re-manufacturing system, in view of delivering a regenerated product

that is of comparable quality with respect to the new product. In the proposed

framework, this second aspect is represented by an increased variability of post-

use products collected from the market, with respect to pre-use products.

The information flow in Figure 1.6, is represented by interrupted lines between

the different elements, which creates and enriches the manufacturer knowledge

base, in a continuous learning perspective. Two major information flows are

relevant. The first flow makes the product design information available to the

manufacturing and de and re-manufacturing system. This information is useful

especially in the context of complex products, were managing product

information is a valuable asset for better targeting de and re-manufacturing

operations. For example, the widespread use of embedded software and system

integration in mechatronics makes product information availability as a strategic

resource for remanufacturers in the automotive market, due to the complexity of

diagnostics and testing. The second flow refers to the gathering of information

from the product use-phase back to the manufacturer. This information is useful

for better understanding the condition of post-use products and adapting de and

re-manufacturing decisions and operations accordingly. [12]

P a g e | 27

Figure 1. 6 Manufacturer-centric Circular Economy framework

In particular, when developing the circular economy, information is needed for

effective planning and management, including the creation of scenarios for

optimal reduction, reuse, and recycling.

Every corporate enterprise, from a small business to a large multinational

corporation, is part of a larger economic system or web. Companies are

interlinked via increasingly complex supply chains. Therefore, an information

system adopting a systems approach is required if decision-makers are to find

more environmentally and financially beneficial ways to plan and manage their

resources.

In most cases, accurate information is not available to decision-makers or is not

conveyed in a timely manner. Moreover, due to fragmented management

frameworks, different kinds of information often belong to different agencies. For

example, environmental protection agencies maintain control over emissions data

while economic development agencies usually collect, and control data related to

economic performance.

Critically, neither of these agencies is subordinate to the other, and cross-agency

collaboration is still rare, with the result that neither agency can play a leading

role nor collaborate in providing such information to the corporate world. [13]

P a g e | 28

In Figure 1.7 the relationship among information and de-manufacturing processes

is represented, showing most important flows.

Figure 1. 7 Information and good’s flow in the circular economy

In this sector data are a fundamental source of value creation, since information

about the market, must be assisted with information on collection process and

manufacturing processes to optimize the de-manufacturing system. Data about

consumers can be used to evaluate future incoming flows of products, but these

must be cross-analyzed with information on collection system, which follows a

process decoupled in time with consumptions. Typical problem is that disposal of

products can be delayed with respect to substitution of the product, cause people

still keep old parts without throwing them away. This usually happens in the

mobile phone sector where old phones are disposed after many years of their

actual substitution, for they are kept as possible back-ups by consumers. This

decoupling, happening in many sectors, brings a lot of uncertainty for de-

manufacturing companies, who work in a continuously delayed reference context.

In addition to this complexity, de-manufacturing processes are driven by products

characteristics, an external variable in this case, for the definition of technological

and operational parameters. Availability of information on contained parts,

assembling methods and material composition become a powerful source in the

process. Unluckily access to this asset is not always guaranteed, creating problems

for or even preventing from effective de-manufacturing. [2]

P a g e | 29

1.7 Problem statement and motivation of the thesis

By understanding the concept of circular economy, its effects, and advantages and

its vital role in economic and environmental aspects, it is noteworthy to find out

the related and needed factors that could pave the ground for implementing a CE

plan in each specific business area. One of the problems which make the de-

manufacturing process of WEEE (Waste Electrical and Electronic Equipment),

difficult or not optimized, is the shortage or lack of information about the products

which have reached their end of life stage and need to be processed to continue

their way on the closed loop of the circular economy path. Regarding the needed

information which may ease the optimization of the de-manufacturing process, it

should be considered that what type of information can be provided by

manufacturers and which type of information should be gathered and add to the

manufacture’s data to finally result in a more profitable, green and optimized de-

manufacturing process.

Regarding the EEE (Electrical and Electronic Equipment) CE, it can be notified

that there is a wide gap between the manufacturers and the recyclers and de and

re-manufacturing companies to optimize the usage of reverse flow of the WEEE.

At the current time, this lack of data is causing a huge economic loss in this

business. The economic loss is coming from two aspects, the material loss and

functionality loss from functioning parts of an EOL (End of Life) product.

Considering the WEEE, the focus of this study is mainly on the PCBs (Printed

Circuit Boards) used in different EEEs and also the expected needed data for the

recycler which can be provided by the manufacturer. Hence the motivation of

such a work is to respond and clarify the answers to these questions:

“What type of information would be useful in this way and why are they needed?”

Obviously, the first step of each way starts from the introduction and clarifying

the main needed factors and elements, hence in order to optimize the recycling

and de/re-manufacturing of PCBs, it is crucial to specify which kind of data can

be used and would be helpful to decrease the loss and generate more profit. To

answer this question, we had to think from the recyclers’ side and consider their

resource of profit. To be more precise, the resource of the recycler, de/re-

manufacturer comes from the product’s formative precious materials or the whole

product (which can be reused) or its parts and components (which are functioning

and reusable) respectively. The data sources are either market or manufacturer,

which provide different sets of data, and the scope of this work is to combine and

link them, in order to optimize de-manufacturing and specifically recycling.

In this work, we have precisely focused on the PCB manufacturer’s data that can

be provided. Hence among the wide variety of data and information which a

P a g e | 30

manufacturer has and could provide, those advantageous ones are those which are

related to material (in case of recycling) or the data about introducing a

functionality of the product and its consisting components, and also all the data

which helps the disassembly task in order not to damage the usable parts. In order

to answer the “Why they are needed?” it should be said that the PCB’s material

declaration data indicates the intrinsic material content of a board, and obviously

by combining this set of information with the data of material value in the market

and also considering the costs of needed treatments to achieve them, the potential

profit of treating and recycling it, could be calculated.

Regarding the PCB or its components functionality related data, it should be

noticed that the presence of used PCB and its components’ market obliged us to

think about matching the market needs and the returning PCBs from the users. As

a result, the knowledge about the board and its components’ functionality will get

a high level of importance in order to be able to answer the used board’s market

demand.

Who is the provider of the needed data for the recycler or de/re-manufacturer?

Regarding the source of data and information which are useful for the recycler,

de/re-manufacturer, it should be clarified, from whom, what type of data could be

achieved? An example could make it clearer: if the data about the exact amount

of precious materials of each present component on the board is needed to

optimize the recycling process, hence its provider also should be defined, whether

“board manufacturer”, “component manufacturer”, “market”, “de/re-

manufacturer” or “recycler”.

How do the different data can help the recycler or de/re-manufacturer?

The main question and the most controversial issue to be solved is HOW to benefit

from the data? HOW to link different information and use them? In this work, the

main focus would be on this concept. By considering a set of hypothesis and

assumptions it has been tried to make a path in order to link the data and make a

logical relation among them. And the more precise results will be achieved by

working on finding the exact numbers and values of the assumed measures.

1.8 Thesis outline

The framework of this thesis consists of 5 chapters which start from a 1. A

preliminary overview of the circular economy concept and its effects on the

environment and more precisely electronics and electric industry, then 2.

Focusing more on the printed circuit boards (PCB) and their structure,

manufacturing, and de-manufacturing processes in order to familiarize with PCB,

the target object which is the core of the work. After knowing the PCB in chapter

P a g e | 31

3. A “simulating” waste PCB treatment algorithm is proposed and based on

which, an economic model is defined and analyzed to optimize the PCB’s de-

manufacturing process. Furthermore, the validation of the model and sensitivity

analysis will be performed. As the model is defined, in chapter 4 the case study

of Italtel’s communication PCBs will be performed and the proposed economic

model will be applied to the case and the outcome and results will be analyzed at

the end. Regarding the last chapter 5, a critical sensitivity analysis of the whole

developed work is performed, reporting the conclusion of obtained results and

suggesting possible improvements.

P a g e | 32

2 State of the art in PCB manufacturing

and de-manufacturing

2.1 PCB Structure

Printed circuit boards are the heart of every electronic product and have been

therefore assembled in very big numbers. PCBs are an integral part of the majority

of electronic systems and are commonly found in consumer electronics, military

applications, and medical equipment etc. In this chapter after presenting PCBs

structure and production processes, current de-manufacturing technologies for

waste PCBs are reviewed and presented. [2], [14]

Figure 2. 1 A Printed Circuit Board with various high value components

A PCB is generally a mechanical support and electrically connects the electronic

or electrical components using some conductive tracks, pads and some other

features etched from one or more sheet layers made of copper which is laminated

onto and/or between the sheet layers of a non-conductive substrate. Generally, the

components are soldered onto the PCB to both electrically connect and

mechanically fasten them to it.

P a g e | 33

Classification of printed circuit boards

The physical property of the PCB can be either rigid, flexible, or a combination

of both. Irrespective of the both, PCB consists of three basic parts: 1) a non-

conducting substrate or laminate; 2) conductive circuits printed on or inside the

substrate, and 3) mounted components. Of the several types of substrates used in

PCBs, the most widely used one is the (Flame Retardant) FR4, which is made up

of glass fiber reinforced epoxy resin with a brominated flame retardant in the

epoxy matrix. Another widely used laminate is the FR2 laminate that is made up

of paper-reinforced phenolic resin with added flame-retardants. [14]

The printed circuit board can be considered a layered structure, usually with

multiple copper and insulating layers. The main portion is a nonconductive

material between 0.127 cm and 0.965 cm thick called substrate, often composed

of fiberglass and epoxy to separate copper layers. The conducting layers consist

of copper (Cu) foils etched away in specific areas where the user does not want

electrical connections. [2]

The first classification of PCBs can be based on the number of laminated

conductive foils: (a) single-layer PCB (b) double layer PCB, (c) multilayer

structure. [15]

(a) Single-Layer PCB

As shown in Figure 2.2, Single Sided PCBs contain only one layer of conductive

material and are best suited for low-density designs. Single-sided PCBs have been

around since the late 1950s and still dominate the world market in sheer piece

volume. Single-sided printed circuit boards are easily designed and quickly

manufactured. They serve as the most cost-effective platform in the industry. One

thin layer of thermally conductive but electrically insulating dielectric is

laminated with copper. Solder mask is usually applied on top of the copper.

Figure 2. 2 Single-layer PCB Diagram

(b) Double-layer PCB

Double Sided PCBs as presented in Figure 2.3, are the gateway to higher

technology applications. They allow for a closer (and perhaps more) routing traces

by alternating between a top and bottom layer using Vias. Today, double-sided

P a g e | 34

printed circuit board technology is perhaps the most popular type of PCB in the

industry.

Figure 2. 3 Double-layer PCB Diagram

(c) Multi-Layer PCB

Unlike a Double-Layer PCB which only has two conductive layers of material,

all multi-layer PCBs must have at least three layers of conductive material which

are buried in the center of the material and is shown in Figure 2.4.

Alternating layers of prepreg and core materials are laminated together under high

temperature and pressure to produce Multi-layer PCBs. This process ensures that

air isn't trapped between layers, conductors are completely encapsulated by resin,

and the adhesive that holds the layers together is properly melted and cured. The

range of material combinations is extensive from basic epoxy glass to exotic

ceramic or Teflon materials.

Figure 2. 4 Multi-layer PCB Diagram

The figure above illustrates the stack up of a 4-Layer/ multilayer PCB. Prepreg

and core are essentially the same material, but prepreg is not fully cured, making

it more malleable than the core. The alternating layers are then placed into a

lamination press. Extremely high temperatures and pressures are applied to the

stack up, causing the prepreg to "melt" and join the layers together. After cooling

off, the end result is a very hard and solid multilayer board. [16]

Another possible classification can be based on the type of substrates architecture,

where four different variances can be defined: (a) rigid boards, (b) flexible

circuits, (c) rigid-flex boards and finally (d) the high density interconnected, HDI,

boards.

P a g e | 35

Figure 2. 5 Different PCB Classifications

a) Rigid PCB

Rigid PCBs are printed circuit boards that are made out of a solid substrate

material that prevents the board from twisting. Possibly the most common

example of a rigid PCB is a computer motherboard. The motherboard is a

multilayer PCB designed to allocate electricity from the power supply while

simultaneously allowing communication between all of the many parts of the

computer, such as CPU, GPU, and RAM.

Rigid PCBs make up perhaps the largest number of PCBs manufactured. These

PCBs are used anywhere that there is a need for the PCB itself to be set up in one

shape and remain that way for the remainder of the device's lifespan. Rigid PCBs

can be anything from a simple single-layer PCB all the way up to an eight or ten-

layer multi-layer PCB.

b) Flexible PCB

Unlike rigid PCBs, which use unmoving materials such as fiberglass, flexible

PCBs are made of materials that can flex and move, such as plastic. Like rigid

P a g e | 36

PCBs, flexible PCBs come in single, double or multilayer formats. As they need

to be printed on a flexible material, they tend to cost more for fabrication.

Still, flexible PCBs offer many advantages over rigid PCBs. The most prominent

of these advantages is the fact that they are flexible. This means they can be folded

over edges and wrapped around corners. Their flexibility can lead to cost and

weight savings since a single flexible PCB can be used to cover areas that might

take multiple rigid PCBs.

Flexible PCBs can also be used in areas that might be subject to environmental

hazards. To do so, they are simply built using materials that might be waterproof,

shockproof, corrosion-resistant or resistant to high-temperature oils (an option

that traditional rigid PCBs may not have).

c) Rigid-flex

Flex-rigid PCBs combine the best of both worlds when it comes to the two most

important overarching types of PCB boards. Flex-rigid boards consist of multiple

layers of flexible PCBs attached to a number of rigid PCB layers.

Flex-rigid PCBs have many advantages over just using rigid or flexible PCBs for

certain applications. For one, rigid-flex boards have a lower parts count than

traditional rigid or flexible boards because the wiring options for both can be

combined into a single board. The combination of rigid and flexible boards into a

single rigid-flex board also allows for a more streamlined design, reducing the

overall board size and package weight.

Flex-rigid PCBs are most often found in applications where space or weight are

prime concerns, including Cell phones, Digital cameras, Pacemakers, and

Automobiles. [17]

d) Highly density interconnected

HDI boards have flexible or rigid substrates with very small conductive tracks,

increasing the density of assembled components. Unfortunately, manufacturing

costs are increased for difficulties in the operations.

P a g e | 37

2.2 Manufacturing of PCB

Figure 2. 6 A part of PCB Manufacturing line

Printed circuit board processing and assembly are done in an extremely clean

environment where the air and components can be kept free of contamination.

Most electronic manufacturers have their own proprietary processes, but the

following steps might typically be used to make a two-sided printed circuit board.

Making the substrate

Woven glass fiber is unwound from a roll and fed through a process station where

it is impregnated with epoxy resin either by dipping or by spraying. The

impregnated glass fiber then passes through rollers which roll the material to the

desired thickness for the finished substrate and also remove any excess resin.

(Figure 2.7)

P a g e | 38

Figure 2. 7 Making the Substrate

The substrate material passes through an oven where it is semi-cured. After the

oven, the material is cut into large panels.

The panels are stacked in layers, alternating with layers of adhesive-backed

copper foil. The stacks are placed in a press where they are subjected to

temperatures of about 340°F (170°C) and pressures of 1500 psi for an hour or

more. This fully cures the resin and tightly bonds the copper foil to the surface of

the substrate material.

Drilling and plating the holes

Several panels of the substrates, each large enough to make several printed circuit

boards, are stacked on top of each other and pinned together to keep them from

moving. The stacked panels are placed in a CNC machine, and the holes are

drilled according to the pattern determined when the boards were laid out. The

holes are deburred to remove any excess material clinging to the edges of the

holes.

P a g e | 39

Figure 2. 8 Component Assembly

The inside surfaces of the holes designed to provide a conductive circuit from one

side of the board to the other are plated with copper. Non-conducting holes are

plugged to keep them from being plated or are drilled after the individual boards

are cut from the larger panel.

Creating the printed circuit pattern on the substrate

The printed circuit pattern may be created by an "additive" process or a

"subtractive" process. In the additive process, copper is plated, or added, onto the

surface of the substrate in the desired pattern, leaving the rest of the surface

unplated. In the subtractive process, the entire surface of the substrate is first

plated, and then the areas that are not part of the desired pattern are etched away

or subtracted. We shall describe the additive process.

The foil surface of the substrate is degreased. The panels pass through a vacuum

chamber where a layer of positive photoresist material is pressed firmly onto the

entire surface of the foil. A positive photoresist material is a polymer that has the

property of becoming more soluble when exposed to ultraviolet light. The vacuum

ensures that no air bubbles are trapped between the foil and the photoresist. The

printed circuit pattern mask is laid on top of the photoresist and the panels are

P a g e | 40

exposed to an intense ultraviolet light. Because the mask is clear in the areas of

the printed circuit pattern, the photoresist in those areas is irradiated and becomes

very soluble.

The mask is removed, and the surface of the panels is sprayed with an alkaline

developer that dissolves the irradiated photoresist in the areas of the printed circuit

pattern, leaving the copper foil exposed on the surface of the substrate.

The panels are then electroplated with copper. The foil on the surface of the

substrate acts as the cathode in this process, and the copper is plated in the exposed

foil areas to a thickness of about 0.001-0.002 inches (0.025-0.050 mm). The areas

still covered with photoresist cannot act as a cathode and are not plated. Tin-lead

or another protective coating is plated on top of the copper plating to prevent the

copper from oxidizing and as a resist for the next manufacturing step.

The photoresist is stripped from the boards with a solvent to expose the substrate's

copper foil between the plated printed circuit patterns. The boards are sprayed

with an acid solution which eats away the copper foil. The copper plating on the

printed circuit pattern is protected by the tin-lead coating and is unaffected by the

acid.

Figure 2. 9 Inkjet PCB pattern printing on a substrate

P a g e | 41

Attaching the contact fingers

The contact fingers are attached to the edge of the substrate to connect with the

printed circuit. The contact fingers are masked off from the rest of the board and

then plated. Plating is done with three metals: first tin-lead, next nickel and then

gold.

Fusing the tin-lead coating

The tin-lead coating on the surface of the copper printed circuit pattern is very

porous and is easily oxidized. To protect it, the panels are passed through a

"reflow" oven or hot oil bath which causes the tin-lead to melt, or reflow, into a

shiny surface.

Sealing, stenciling, and cutting the panels

Each panel is sealed with epoxy to protect the circuits from being damaged while

components are being attached. Instructions and other markings are stenciled onto

the boards.

The panels are then cut into individual boards and the edges are smoothed. [18]

Components Assembly

Components can be assembled in two different ways: with the Through Hole

Technology (THT) or with the Surface Mount Technology (SMT). These two

technologies bring very different results in term of performances and costs.

Pin Through Hole (PTH) components, in Figure 2.10, have metal inserts soldered

into holes of the board. With respect to SMT components, PTH ones have strong

limitations in terms of density, because components can be assembled only on one

side of the board and the minimum distance between pins cannot be lower than 1

mm.

In the past only PTH components were available and for this reason, a lot of

components were developed for this technology. Some examples are electrolytic

capacitors, axial resistors, transistors in the so-called TO package and first

integrated circuits, put into the Dual Inline Package (DIP).

In order to increase the density of components on boards, a new mounting

technology has been developed: surface mounting devices, SMDs, are soldered

on small lands specifically coated on the substrate.

P a g e | 42

Figure 2. 10 Through Hole Technology & Surface Mount Technology

In this way some advantages are obtained: (a) leads are smaller than pins, (b)

higher density of packages, (c) smaller boards, (d) easier assembly operations,

particularly coating of lands is much easier with respect to drilling holes and

finally (e) possibility to mount components on both sides of the board. Besides

these advantages, the SMD packages bring problems for lower maneuverability

of components for their reduced dimensions, higher costs, and problems in the

inspection phase. [2]

Packaging

Unless the printed circuit boards are going to be used immediately, they are

individually packaged in protective plastic bags for storage or shipping. [18]

P a g e | 43

2.3 De-manufacturing framework and its implication on

EOL PCBs

Figure 2. 11 Repairing or Recycling the PCB

De-Manufacturing can be defined as the “break down of a product into its

individual parts with the goal of reusing and remanufacturing parts or recycling

the remainder of the components”. A De-Manufacturing strategy should find the

right mix between (i) product remanufacturing and re-use, (ii) sub-assemblies and

components re-use within manufacturing, (iii) material recycling and recovery,

(iv) incineration and (v) waste disposal in landfills to maximize the residual value

of end-of-life products and to minimize the environmental impact.

In order to push the establishment of structured De-Manufacturing operations,

different countries are pursuing this issue through regulations. For example, in

Europe the regulation fixes recycling targets for Waste Electrical and Electronic

Equipment (WEEE) and End of Life Vehicles (ELVs) – EU Directives

2012/19/EU and 2000/53/EC.

A De-Manufacturing System can be defined as the set of resources (human and

technological), organization, IT infrastructure and associated business model to

enable product De-Manufacturing (Figure 2.12).

P a g e | 44

Figure 2. 12 Products and materials Life Cycle

De-manufacturing processes typically include disassembly, remanufacturing,

recycling and recovery processes, which are going to be presented in detail in the

following section.

2.3.1 Disassembly in de-manufacturing

Material and product function liberation processes are usually the first stages of a

de and remanufacturing process-chain. They include disassembly and size-

reduction processes. Disassembly is performed with the scope of isolating (i)

hazardous components that should not enter the de and re-manufacturing flow,

(ii) re-usable parts with high residual value, (iii) parts that need to go through a

dedicated process-chain.

Disassembly makes it possible to achieve product function recovery, high

material-return rate, and pre-concentration of waste. However, it is usually an

expensive process due to the difficulty of the tasks and the demand for manual

labor. Disassembly processes have been classified into three categories, namely

destructive, semi-destructive and non-destructive. In non-destructive disassembly

all the output components remain undamaged. This is desired for maintenance,

reuse and remanufacturing. In semi-destructive disassembly, only connective

components are destroyed, e.g. via breaking, folding or cutting, leaving the main

components with little or no damage. Destructive disassembly deals with the

partial or complete destruction of obstructing components (e.g. welds). In general,

destructive processes, sometimes called dismantling, require reduced costs and

times with respect to non-destructive processes. The definition of the disassembly

P a g e | 45

mechanism to be adopted for a specific product depends on the (i) type of joint,

(ii) type of material, (iii) required performance and (iv) target automation level.

Disassembly is one of the key steps for efficient treatment of post-use products.

However, the process may become economically infeasible due to high labor

costs, especially in high-wage countries. Therefore, automation could be a

possible solution for cost-effective disassembly. Other reasons of interest for this

area are the attempt to increase the level of quality and standardization of the

process and the need of increasing the workers’ safety by reducing the risk of

contact with hazardous materials. However, uncertainties and variations in

returned post-use products are significant sources of problems for disassembly

process automation. [19]

PCB Disassembly and the state of the art PCB disassembly technologies

“The realization of electronic components' non-destructive unsoldering separation

from PCB scraps has ever been a tough challenge for electronic and electrical

products' recycling.” [20]

In order to mention some of the used methods in PCB’s components’ disassembly

process, firstly it is crucial to review the joints which are used to assemble the

electronic components to the board, and then talk about the techniques to break

the mentioned joint in order to dismount the components.

As mentioned before, in the manufacturing of PCB part, the mounted components

on the PCBs are mainly attached to the PCBs by either THT (Through Hole

Technology) method or SMT (Surface Mounted Technology) method. Hence the

needed technology to be executed for disassembly task depends on what type of

joint, the component has.

It should be noted, that the used disassembly technologies and methods can be

divided in different categories, based on the type of components’ joints or the

need for disassembling a precise component or all the components or being

destructive or non-destructive (using high temperature or low temperature

processes, using mechanical shearing forces or chemical solder solvents,…) that

finally considering the criticality and the aim of the disassembly task, the

disassembly method can be selected and employed to achieve the precise

components.

In the following a general view of the most common and also the state of the art

means of disassembly task, are presented:

P a g e | 46

Gas/air Heating

Gas/air heating is a widely chosen heating approach with the benefits of low cost,

cleanness and heating mildness. However, gas/air heating has two main

disadvantages: low heating efficiency and low controlling accuracy when heating

a large area. These are not big problems in stepwise disassembly because the area

needs to be heated is only slightly larger than the surface of a component. But for

simultaneous disassembly of PCB, the area is usually too large to exactly achieve

the expected spatial distribution and temporal variation of temperature. [21]

The thermal convection method uses air or nitrogen as the medium to transmit

heat. Waste PCBs are exposed to air or nitrogen, which is heated. When the

temperature is high enough to reach the melting point of the solder, the solder will

be melted. When gas is used as the medium to transmit heat, it is difficult to

control the temperature and thermal efficiency is poor.

The wide usage of gas heating method to disassemble the components is also due

to its flexibility with the different joint types (THT or SMT). By providing hot

gas with the temperature above the melting point of the solder alloy (the used

paste to attach components to the board), it would be possible to melt the solder

and break the joint of components and the board and taking the component apart

by means of a mechanical tool or force. [22]

In this case, there are various tools which use hot gas as the main factor in

breaking the joint among the components and the boards, such as 1. Hot Gun, by

different heads’ sizes, provides a concentrated heat zone on desired and selective

component and by melting the solder, the component can be easily dismounted

from the board 2. Hot Gas Chamber, in which, the PCBs are loaded and by

presence of hot air and rotation, all the components will be separated from the

boards, “When the disassembling temperature, rotating speed, and incubation

time are 265 ± 5 °C, 10 rpm, and 8 min, respectively, the solder can be completely

removed from the PCBs. [23]

Infrared Heating

It is a method to generate needed heat for melting the solder paste in order to let

the components be dismounted from the PCBs. As it is going to be explained,

the infrared technology is used both in manufacturing and also de-

manufacturing of PCBs, since it can be classified as flexible and controllable

technology which let us to choose its concentration level from a single

component to a wide range of assembled components on a wide area of the

board.

P a g e | 47

Figure 2. 13 Infrared heating application

Recently, there is a general trend for electronic equipment to be made increasingly

compact and lightweight. Accordingly, printed circuit boards having a large

number of electronic parts mounted in a limited area (hereunder referred to as

"high-density boards" or "high-density printed circuit boards") are widely used.

In the manufacturing of high-density boards, it is necessary to supply heat to a

narrow area between electronic parts on the high-density board in order to reflow

a paste solder or cure an adhesive resin when the electronic parts are connected to

the circuit board using a paste solder or a resinous bonding agent. As an industrial

heating apparatus for these purposes, a reflow furnace is used in which infrared

heaters are placed on the top and bottom walls of a tunnel-type heating zone. The

used infrared heater in the reflow furnace comprises a sheath heater and an

infrared radiation panel made of a surface-treated stainless-steel plate which is

disposed over the sheath heater. [24]

In a work done by (Park, Seungsoo), an apparatus is presented which consists of

3 disassembly modules, each one of which contains a pair of rotating feeding rods,

Infrared Radiation (IR) heaters, and steel brushes. The coupled rods have toothed

gears and are very close to each other. The rods roll in the opposite direction to

each other and a PCB is fed slowly into the module by placing between the feeding

rods. As the PCB moves down, a couple of IR heaters raise the temperature to a

value higher than the melting point of solder, so that the electronic components

are easily detached from the boards. The steel brushes which are placed similarly

as the feeders but roll faster than the feeders and impart a large abrasion force to

the surface of the PCB to scrape off the components from the boards. At this stage,

P a g e | 48

the gears on the feeding rods help to maintain a constant feeding speed as long as

the abrasion is in effect. [25]

The presented method has been used since the 1990s by the NEC Company for

production of PCB. This method is environmental friendly and efficient, but

electronic components near solders are also exposed to some infrared and may

suffer damages when using infrared heating. [22]

Chemical De-soldering

In this technology, environmental-friendly and widely used methanesulfonic acid

(MSA) electrolyte for Industrial lead-tin alloy electroplating is used together with

hydrogen peroxide (H2O2) as a novel chemical leachate to dissolve solder

selectively for dismantling electronic components from waste PCBs. As Zhang,

Xiaojiao in 2017 made the exploratory trials, the MSA solution without H2O2

was unable to achieve de-soldering separation of tin-lead alloy for dismantling of

ECs from waste PCBs dramatically. After the addition of H2O2 into the MSA

solution, the welded electronic components significantly fell off the PCBs. In the

MSA-H2O2 aqueous system, H+ and CH3SO3 are ionized by MSA, and the

differences of the standard electrode potential between hydrogen peroxide/tin

(1.912 V) and hydrogen peroxide/lead (1.902 V) are so close that they can be

considered to be stripped down easily from the PCBs with almost the same ratio

at room temperature. Furthermore, the thermodynamic calculations of using the

HSC Chemistry Version 5.0 software had demonstrated that the change of Gibbs

free energy (2181.5 kcal/mol at 208C) of the redox reaction was much lower than

the boundary of 210.1 kcal/mol. Thus, this phenomenon can occur spontaneously.

After the leaching process, the constituents of the leachate are similar to the

solution that was used to electroplate lead-tin alloy. Therefore, tin and lead in the

leachate may be directly electrolyzed to reclaim the solder rather than to undergo

a purification process. Moreover, H2O2 is relatively stable in the acidic medium,

and its final product is environmental-friendly H2O, hardly affecting the recycling

of the leachate. Therefore, the MSA-H2O2 aqueous system can be regarded as

environmental-friendly de-soldering separation of tin-lead alloy for dismantling

electronic components from waste PCBs. However, the Cu metallization in the

PCBs is covered by the lead-tin solder, with the dissolution of solder into the

leachate, the leaching of the copper layer may also occur partly at room

temperature, since the difference of standard electrode potential between

hydrogen peroxide and copper is 1.439 V. Therefore, it is critical to control the

leaching of copper to be least for the achievement of environmental-friendly de-

P a g e | 49

soldering separation of tin-lead alloy for dismantling electronic components from

waste PCBs by the MSA-H2O2 aqueous system.

The optimum concentrations of H2O2 and MSA are 0.5 and 3.5 mol/L,

respectively, and the reaction time is 45 min. Under the optimum condition, the

dissolution rates of lead and tin calculated with the equations are 99.68% and

102.91%, respectively. [26]

LASER PROCESS

Of all the methods of forming solder joints available to industry, laser soldering

is particularly close to hand soldering in that the heat energy is directed only at

the site of the joint to be formed, using the laser beam in a similar manner to a

soldering iron. The process therefore has several of the advantages associated with

hand soldering but with additional benefits resulting from the fact that it is a non-

contact method and one which fits readily into an automated electronics assembly

process.

The processes involved in forming laser soldered joints vary according to the

application but typically they include the following:

Place solder paste or solid solder at the joint location

Place parts to be joined

Focus laser energy on the joint with appropriate energy and for appropriate

duration

Repeat for subsequent joints

The benefits of the technique result from the focused nature of the applied energy.

This means that only the amount of energy required to form the joint is applied

and the remaining components are not heated. This confined heating eliminates

many of the problems associated with damage to heat sensitive components,

differential expansion of components and bridging between joints. In addition, it

has been shown that the rapid heating and cooling of the solder which laser

heating allows results in improved metallurgical properties, with finer grain

structure, than with the slower heating and cooling associated with conventional

methods. [27]

P a g e | 50

Liquid Heating

The liquid heating method uses some liquid as the heating medium, such as oils,

waxes, ionic solution, and molten salt. The molten solder is shaken down by the

vibrator after the solder melts. However, ECs will be contaminated by the heating

medium and thus need cleaning, drying, and other subsequent disposals, which

will generate waste water pollution. [22]

Generally, the used liquid should have excellent characteristics such as high

boiling point, low volatility and benign environmental performance. Based on the

experiment of paper’s sources work, a special liquid, methylphenyl silicone oil,

was used as the heating medium. Good disassembly results are obtained from the

experiment, and it indicates that the liquid heating disassembly approach has

many advantages. They had used a mixed liquid with the following

characteristics: the heating liquid needs many special characteristics. Firstly, its

boiling point must be higher than the melting point of the solder tin. Normally its

boiling point must be higher than 250°C, and it must have outstanding oxidation

resistance ability and thermostability at high temperature. Secondly, it must be

innoxious and non-poisonous to the environment and the human beings. Thirdly,

it must have well chemical inertia and no reaction with the material in the PCBs.

In the experiment, they had used methylphenyl silicone oil and

polydimethylsiloxane with thermo-oxidative stability additives, as the heating

medium and as they claim, the heating liquid had been under the top temperature

of 250 and the top heating time span of 6 hours.

Experiment equipment framework

Figure 2.14 is the structure schematic diagram of the experiment facility which

had been used in their work. The PCB is clamped by a resizable clamp and heating

in the liquid. The PCB immerges into the liquid horizontally with the components

mounted surface upwards. The liquid surface is just higher than the components

pins. The PCB which is clamped as is shown in the diagram, firstly would be

heated in the liquid and then the vertical mechanical vibration will be added on

the clamp by a dynamoelectric vibrator. And at the same time, the ultrasonic wave

added into the heating liquid by the ultrasonic vibrators on the bottom of the

vessel.

P a g e | 51

Figure 2. 14 Structure schematic diagram of the experiment facility

As regard as the vibrators and vibrating forces, there are two kinds of vibrations

added to the PCB in the disassembly process. One is the mechanical vibration that

is produced by the dynamoelectric vibrator. The mechanical vibration is added on

the clamp vertically in the heating process, it can help the soldering tin separate

from the PCB through the hole and drop into the heating liquid. The soldering tin

can be gathered from the liquid with high purity after the heating process. And

also, the mechanical vibration is added to the separation position. The components

mounted on the board drop onto the griddle. Because of the liquid heating effect,

the components simultaneous disassembly efficiency is very high.

The ultrasonic wave is also added to the liquid in the unsoldering process. Two

ultrasonic motors are installed on the bottom of the vessel. The experimental

ultrasonic wave is added to make the soldering tin into the liquid in melting state

through the effect of the cavitation and agitation of the ultrasonic wave. The

application of ultrasonic wave in the unsoldering process can enhance the

soldering separation efficiency, shorten the de-soldering time and simplify the

disassembly process. [28]

Various methods’ advantages and disadvantages overview

To summarize the research contents of the literature, Table 2.1 lists the properties

of de-soldering methods from disassembly efficiency, equipment cost, running

cost, components damage, energy utilization rate, and environmental harm.

P a g e | 52

Table 2. 1 Comparisons between different heating methods for de-soldering

From Table 2.1 we can see, the main defects of thermal convection are low energy

efficiency, damage to the electronic components, and environmental pollution.

The incremental innovation should be adopted to increase benefits and decrease

harms. Thus, its next step is to invent an improved thermal convection method

that can increase the energy utilization and decrease damages. While the liquid

heating has higher energy efficiency, it cannot avoid damaging components and

the environment. So, in order to promote this method by increasing its ideality,

the ameliorative liquid heating method should adopt an environmentally friendly

heating liquid medium that does not damage the electronic components.

The infrared and laser technologies maturity is below the current average stage of

the de-soldering domain, which is now in a growth stage. Thus, infrared and laser

heating methods are still in the germination stage. The main defect of infrared

heating is high running cost, and the main defects of laser heating are high running

cost and low de-soldering efficacy, as is shown in Table 2.1. [22]

2.3.2 Remanufacturing in de-manufacturing

Remanufacturing entails renovating used parts or components so that they can

perform their function similar to new ones. Remanufacturing generally consists

in dismantling the used units into their components, inspecting these components,

repairing defective components or replacing them with new ones, reassembling

the units, readjusting as necessary and submitting them to the final quality test,

that usually the same used for newly manufactured parts.

Re-manufacturing is recognized to be the most beneficial end-of-life treatment. It

preserves more or less 85% of the initial value, while recycling preserves almost

7.5% of initial value.

Re-manufacturing processes use 20-25% of the energy needed to manufacture the

same product. The cost of remanufacturing can be between 45% and 65% less

than the manufacturing cost. Remanufactured components are traditionally not

P a g e | 53

reused in the manufacturing process to be assembled in new products, although

their quality could support this model.

Remanufactured components are typically sold in the aftermarket as spare parts.

For example, in the automotive industry, of the total gross profit of the car

manufacturer, the new car sells contribute to the 18%, the service generates, 14%

while the spare parts market generates 39%. [29]

Generally speaking, remanufacturing of the EOL products, has a range of various

steps which is going to be explained:

Inspection: Inspection processes are applied in de and re-manufacturing

systems at different stages, for several purposes. Different types of

applications are found in material characterization during recovery

processes and in product functionality assessment during remanufacturing

processes. Concerning applications in recovery processes, the goal of

inspection is to gather information about the composition, the particle

shape and size of the mixture under treatment. Applications are usually

off-line and require sample extraction, preparation and measurement

execution. The collected off-line information is useful for mixture value

and quality assessment but cannot be exploited for process control. With

the objective to provide intelligence to de and re-manufacturing systems

for material recovery and enable their adaptability to different post-

consumer products, the application of in-line material characterization

technologies would provide capabilities for (i) a full materials data storage

and traceability, (ii) a remote monitoring and control of the processes, and

(iii) an easy reconfiguration of the system. Concerning applications in re-

manufacturing processes, inspection is mainly applied at three different

stages of the process chain, namely post-use product acceptance, part

inspection, and final product testing. These stages of inspection have

different objectives and technologies. The objective of core inspection is

to remove parts that will be uneconomic or impossible to remanufacture.

P a g e | 54

Figure 2. 15 PCB Inspection

This stage is usually performed by visual inspection done by skilled

operators or by simple data gathering methods, such as barcode reading.

The second stage of inspection is carried out once the parts have been

disassembled. It aims to identify and remove damaged, non-reusable

components from the product. This stage can be performed by a

combination of visual, dimensional measurements and non-destructive

tests (NDT), such as magnetic particle inspection. Functional tests are also

adopted in case of electronics. The third stage has the goal to assess and

validate the quality of the regenerated product, before delivering it to the

customer. Visual, dimensional, geometric and functional tests are adopted.

Cleaning: Cleaning is among the most demanding steps in

remanufacturing and it is an essential process because the quality of the

surface cleanliness directly affects the capability to perform surface

inspection, reconditioning, reassembly and painting. Re-manufacturing

cleaning is distinct from cleaning in maintenance or cleaning in new

production. Cleaning in maintenance concentrates on the surroundings of

the damage before repairing. Cleaning in remanufacturing is done for the

whole part in order to meet the quality requirements after

remanufacturing. Cleaning in re-manufacturing is performed on parts that

are of high variability in sizes, materials, shapes and surface conditions.

Cleaning in re-manufacturing has the objective to reduce contaminants

present in or on a component until specified cleanliness levels have been

reached. The objectives of re-manufacturing cleaning in each stage of the

P a g e | 55

re-manufacturing process chain are different. Before the disassembly

process, the goal is mainly to reduce the level of contaminations outside

the products. The goal of cleaning after the disassembly is to detect surface

abrasion, micro cracks or other failures in order to assess the re-

manufacturability of the product and to adjust the needed process stages

according to the part conditions. The cleaning process after regeneration

and before reassembly and painting is performed to remove surface

machining contaminants and to prepare the surface for additional

treatments. Therefore, each phase requires a specific analysis of the best

techniques and methods to be applied.

Figure 2. 16 PCB cleaning with degreaser spray

Reconditioning: Reconditioning processes are applied in de and re-

manufacturing process-chains in order to restore the functionality of

components after disassembly, cleaning, and inspection. Depending on the

part features to be reconditioned and the specific defect type, different

reconditioning processes may be applied. In any case, common

reconditioning processes include the following four steps: (i) surface and

shape defects removal, (ii) material addition and deposition, (iii) material

properties restoration, (iv) surface finishing. In the first step, cracks,

scratches, nicks, and burrs, burnt or corroded regions, and inclusions are

removed by subtractive machining processes such as turning, milling,

drilling, and grinding. Surface finish and tolerances are not of top priority

but rather the removal of all stress raisers. Shape defects, such as bends,

P a g e | 56

warps, and dimples, are also removed if technically feasible. In the second

step, part with cavities and material lacks can be restored to the intended

shape through additive processes, such as welding, powder deposition,

laser cladding. Due to the application of high temperature, pre-heating is

required to avoid cracking. In the third step, desired material properties

are restored through heat treatments, which either remove unwanted

residual conditions or prepare the part to be more resistant to its operating

condition. High surface quality can be achieved by grinding, reaming,

honing, and hard turning processes. In other cases, painting, coating by

spraying processes, and polishing can be applied. Great potential for

modern part reconditioning processes in remanufacturing is provided by

Additive Manufacturing (AM). AM combines the advantage of flexibility

in processing free-form geometries, typical of damaged parts, and the

ability to feed different mixes of materials, thus controlling the

composition of the added layer and making functionally graded materials.

Therefore, AM provides the opportunity to support regeneration and

upgrading of part functionalities, for example adding a wear-resistance

layer on the processed surface. [12]

Figure 2. 17 Reconditioning the PCB to be reused

Remanufacturing concept and PCB remanufacturing

Current product’s life-cycle analysis (LCA) demonstrates that the disposal phase

contributes substantially to the environmental impacts of waste electrical and

electronic equipment (WEEE), (EEC Council Directive on hazardous waste 1991,

P a g e | 57

Hawken 1993, EEC Council Amending Directive hazardous waste 1994),

particularly in products containing toxic materials (ECTEL 1997a), scarce or

valuable materials, or materials with a high energy content.

Within WEEE, there is the combination of all these situations, including, for

example, batteries, quality plastics, precious metals and toxic solder. The figures

of the EU in 1998 for WEEE were in the region of 6.5–7.5 million tons per year.

This represents less than 1% of the total EU solid waste stream but accounts for

at least at base case scenario, 50% of metals in general and almost 100% of heavy

metal in leachate. The ECTEL report indicates that this is because PCBs contain

almost all of the hazardous waste in WEEE. Thus, reusing a PCB means avoiding

this fraction and thus, according to all LCA studies, the heavy metal portions

representing 2% of landfill and 80– 90% of heavy metal in leachate. Thus, it is

critical to find cost-effective methods of addressing WEEE reuse because their

quantity is rapidly increasing.

The significance here is that WEEEs are the EU’s fastest growing waste stream,

but small-sized WEEEs are typically not profitable to re-manufacture as their

volatile technological pace makes their disassembly by conventional means

overly expensive.

Manual disassembly has been used in recycling plants to varying levels of success

since the early 1990s. However, in high turn-over, life-cycle products such as

those in the technology sector (e.g. telecommunication and other hand-held

electronic products), manual disassembly hinders profitable disassembly. For

profitable re-manufacturing of such products, significant economic efficiencies

are required in their disassembly. Active disassembly (AD is an alternative to

conventional dismantling techniques that enable the non-destructive, self-

disassembly of a wide variety of consumer electronics on the same generic

dismantling line, thus reducing disassembly cost aims include hierarchical,

controlled, non-destructive and specific product component release to optimize

reuse from macro to micro dismantling. [30]

2.3.3 Recycling and recovery processes

Recycling and recovery processes and PCB recycling

Recycling is mainly performed by mechanical processes or by thermal processes.

Mechanical recycling systems are multistage systems including size-reduction

and separation technologies. The scope of size-reduction technologies is to (i)

reduce the size of the particles in favor of downstream separation processes, (ii)

liberate inhomogeneous particles thus increasing the quality of the separation

P a g e | 58

process. The scope of separation stages is to split a mixed input stream into two

or more output streams in which the concentration of target materials is greater

than in the input stream. In recycling industry, mechanical separation is performed

by a series of separation stages involving different separation technologies that

classify materials on the basis of their properties. A separation process creates an

environment in which particles with a high value of the property move differently

from those with a low value of the property. Under ideal separation conditions,

the inspection will be perfectly accurate, and the material flow will be correctly

classified.

However, in the real world, random disturbances cause “contamination” of the

output flows, where materials are wrongly classified. Thermal recycling processes

are typically based on pyrolysis. This process enables to separate organic from

non-organic fractions. However, although improvements have been proposed in

the last years, pyrolysis generates serious environmental concerns due to the

energy required for the process and the need for emissions treatment plants. For

some materials, recycling processes enable to reach target grades that are

comparable to market requirements (copper, aluminum). Otherwise, recovery

processes need to be implemented.

Recovery processes make use of chemical reactions to separate the target

materials at very high-grade levels. They are typically performed in batches. One

of the most advanced technology in this field is hydro-metallurgical processes.

These processes have recently been proved to be more sustainable than traditional

metallurgical recovery processes. Nowadays, there are several limitations that

prevent companies from implementing a successful de-manufacturing system.

First, the inefficiency of disassembly processes was pointed out as one of the

major limitations. The low productivity and the high labor cost of manual

disassembly in high-wage countries are not compatible with the residual value of

many disassembled products. An extreme disassembly process efficiency increase

(65%-80%) would be required to make automatic disassembly a convenient

strategy for products such as mobile phones and personal computers. On the other

hand, the implementation of automated disassembly is made technically complex

by the high variability of input materials and their quality conditions. This would

require flexible systems able to detect the type and conditions of parts to be

disassembled and to adapt to their shape and wear conditions, both at macro-level,

for the mechanical parts, and micro-level, for electronic components. [29]

P a g e | 59

Considering the PCB recycling procedure, the mentioned main recycling steps as

1. Size reduction processes and 2. Separation processes will be explained more

by a general overlook over the main activities and methods which are done in this

field.

Size reduction processes

Different mineral processing unit operations such as shredding, crushing and

grinding can be used to liberate metals from cladding materials such as resin,

fiberglass and plastics. Various types of hammer crushers, rotary crushers, disc

crushers, shredders, cutters equipped with a bottom sieve, are used for liberation.

Ball milling and disc milling are also reported for pulverizing the PCBs after

cutting into small sizes. As the PCBs are made of reinforced resin, copper wires

and glass fibers (multilayer), the conventional crushers may not achieve good

liberation. In contrast, shredding or cutting, which works on the principle of

shearing, is found to be more useful. Unlike mineral ores, PCBs do not have a

particular size fraction for liberation; instead, different types of elements are

liberated at different size fractions. Based on the studies on the liberation

characteristics of PCBs and the effect of shape and size on liberation. It is claimed

that below 6 mm size, ferromagnetic and copper are completely liberated and at

the same time, aluminum is found to be liberated in much coarser fraction. Also,

it is reported that glass fiber-reinforced epoxy resin undergoes brittle fracturing

more readily than metallic materials and concentrates in the finer fraction during

impact milling of PCB scraps. Regarding the papers sources, it is also reported

that after milling to below 150 mm size, no interlocking of metallic and non-

metallic particles is observed. Nevertheless, one of the major challenges in

crushing and grinding of waste PCBs is the generation of fine dust that may be

very difficult to handle during subsequent processes.

Figure 2. 18 Single shaft shear shredder, and Chamber of a cutting mill

P a g e | 60

Separation processes (Physical and Chemical processes)

Physical separation process

The density-based separators are widely used to separate lighter fraction from

heavier ones based on the density difference. The metallic fractions can be

separated from the plastics using heavy density liquids such as Tetrabromoethane.

However, efficiency of the separation is poor, as the particle shape and size play

a crucial role in the separation. Peng performed the density-based separation of

waste PCB fines (50e300 mm) in an inclined separation trough and achieved more

than 95% recovery of metallic materials. Air classification method, in which

separation is based on the settling velocity of the particles in the air, has also been

reported for the separation of plastics from the metals.

Figure 2. 19 density-based material separation

Electrostatic separation is another promising technology for separating non-

conducting materials from conducting ones, owing to its advantages of less

environmental hazards, low energy consumption and easy operation.

P a g e | 61

Figure 2. 20 Electrostatic material separation

In electrostatic separation of a mixture of copper, silicon and woven glass

reinforced resin and it is proposed that multiple-stage separation is necessary for

effective separation of conductors, semiconductors and non-conductors. Both

fundamental and practical aspects concerning design of suitable electrostatic

separator for industrial applications have been reported by a number of

researchers. The differences in density and electrical conductivity between

plastics, metals and ceramics provide an excellent condition for application of

corona electrostatic separator. Regarding the employing of corona discharging

electrostatic separator for PCBs recycling and it was realized that corona

electrostatic separator is suitable for particles with size ranges between 0.6 and

1.2 mm, but productivity decreases with finer size fraction.

Eddy current based electrostatic separator is also successfully employed to

separate plastic particles from metal/ plastic mixture as well as non-ferrous metals

from ferrous metals. The separation efficiency depends on the different

trajectories of particle movements due to eddy current (induced in the non-ferrous

particles) and external magnetic field, which deflects the ferrous particles to a

higher degree. Low-intensity drum magnetic separators are generally used to

recover ferrous materials from the non-magnetic fraction. Researchers at Daimler-

Benz in Ulm, Germany, have developed a mechanical treatment approach in

which magnetic separators are used to remove ferrous elements before fine

grinding. The magnetic separators have been used for enrichment of valuable

P a g e | 62

metallic fraction before electrostatic separation of PCB fines. Magnetic separators

are not very much useful for crushed PCBs because the particles are agglomerated

during the process and the non-magnetic materials escape with ferrous one.

Nevertheless, it can be used as groundwork before electrostatic separation for easy

handling of non-conducting materials.

Figure 2. 21 Eddy current material separation

In spite of having diverse compositions, the main driving force for recycling of

PCBs is the recovery of metals. Nevertheless, the recovery of each metal may not

be feasible due to economic reason and technological limitations. The

recyclability of a metal can be determined by the “contribution score” of the

individual metal that is related to weight content, environmental hazards

associated with the metal, energy consumption, natural resources depletion, etc.

The most widely used assessment index is the resource recovery efficiency (RRE)

regarding the paper’s source which compares different metals based on their

weight content, recycling efficiency and world reserves and is expressed as:

Where, E is the recovery percentage, F is the amount of resource/ton of scrap, P

and C are the annual production and consumption of the primary resource

respectively, R is the world reserve of the resource, and i counts the type of the

resources in the scrap. Another approach by the paper’s source, calculates Quotes

for environmentally Weighted RecyclabiliTY (QWERTY) score to determine the

environmental performance by the following equation:

P a g e | 63

Where, EVWactual,i is actual environmental impact for the weight of material i.

EVWmax,i maximum environmental impact for the weight of material i.

EVWmin and EVWmax are total defined minimum and maximum environmental

impact for the complete product, respectively. Based on the above two

approaches, a developed Model for Evaluating Metal Recycling Efficiency from

Complex Scraps (MEMRECS) for prioritizing the selection of target metals is

presented.

The main advantage of this approach is that it includes the environmental impact

as well as natural resource conservation aspects for calculating recyclability.

According to this model, the metal recovery priority should be on the precious

metals such as Au, Ag, and Pd along with some base metals such as Cu, Sn, and

Ni. Nevertheless, the actual score may vary depending upon the weight content

of metals in waste PCBs. [31]

Chemical processes

Many recycling processes of waste PCBs have been tested on a laboratory scale.

For instance, the pyrolysis process was employed to obtain high pure metals.

Uncertainty and potential pollution have declined the process to expand into a

field scale. In this type of recycling, the PCBs are depolymerized into smaller

useful molecules by several techniques, such as pyrolysis, gasification or

application of supercritical fluids. The obtained products (fuels, gases, and tar)

are refined by conventional approaches and the metallurgical approaches are

employed for the treatment of the metallic fraction. A major shortcoming is the

presence of the significant amount of dioxin precursors in pyrolysis oils, which

can possibly be reduced by adding CaCO3, Fe2O3 during pyrolysis. In recent

years, supercritical fluids have been an effective medium for the destruction of

the epoxy adhesive layer.

Pyrometallurgical processes (SMELTING)

Pyrometallurgy, energy intensive and high-cost process, is the traditional

approach for metal recovery from the waste PCBs but selective recovery of

individual metals can hardly be done by this route. Pyrometallurgical techniques

include incineration, smelting in plasma arc furnace, blast furnace or copper

smelter, high-temperature roasting in presence of selective gases to recover

mainly non-ferrous metals. Currently, more than 70% of waste PCBs is treated in

smelters rather than through mechanical processing. The main advantage of

pyrometallurgical treatment is its ability to accept any forms of scrap. Hence,

electronic scrap can be used as a part of raw materials in the smelters for recovery

P a g e | 64

of copper along with Au and Ag. A recycling method, developed by Technical

University Berlin in 1997, turned waste PCBs into a Cu-Ni-Si alloy, a mixed

oxide (mainly Pb and Zn) and environmentally agreeable slag by a top blown

reactor. At Boliden Ltd. Ronnskar smelter, Sweden, waste PCBs are directly fed

into Cu converter to recover Cu, Ag, Au, Pd, Ni, Se, and Zn while the dust

containing Pb, Sb, In and Cd is processed separately for metal recovery. At

Umicore's integrated metal smelter and refinery, electronic scraps are first treated

in IsaSmelt furnace to recover precious metals along with Cu in the form of Cu-

bullion. Cu is first recovered from this bullion through Cu leaching and

electrowinning followed by precious metals recovery from Cu-leached residue in

the precious metal refinery. A popular method to recover the Cu and precious

metals is primary smelting. Waste PCBs can be used to produce a Cu-Ni-Si alloy,

a mixed oxide (mainly Pb and Zn) and slag by a top blown reactor. Vacuum

metallurgy separation (VMS) is suitable for Bi, Sb, Pb and other heavy metals

with high vapor pressure.

HYDROMETALLURGICAL PROCESSES (LEACHING)

The hydrometallurgical route is more selective towards metal recovery from

waste PCBs or pretreated PCBs, easier to control over reaction and creates less

environmental hazards than pyrometallurgical approach. The base metals

recovery has a substantial impact on the economics of the process due to the larger

available amount of waste PCBs. Moreover, recovery of base metals also ensures

the enrichment of precious metals in the solid residue, making it easier to leach

out subsequently. Low capital cost hydrometallurgical processes are mainly used

for recycling of the metallic-ferrous fraction where the extraction of the metal

content is profitable. Depending on the substrate material (ceramic, glass, or

polymer) there are different hydrometallurgical processes used. [32]

2.4 The flow of needed information in de-manufacturing

The recovery of both toxic and non-toxic materials from billions of end-of-life

electronics calls for efficient processes and exploration of opportunities for

computer integrated de-manufacturing for materials recovery. To date, recycling

automation for selective disassembly has been limited by the proliferation of

product designs, the difficulty of acquiring product feature and material content

information, and the lack of integration of collection and de-manufacturing

processing. [33]

P a g e | 65

Also, considering the “industry 4.0” paradigm the concept of information sharing

(computer integration, digitalization) among different industrial actors will

become an important issue.

Making product structure and material composition information from a product

design profile available will support planning models, Petri net algorithms, and

control models for de-manufacturing, as well as integrated manufacturing and de-

manufacturing. Advances in imaging and materials identification techniques, as

well as more flexible technologies to separate materials, may provide new

opportunities for expert Petri net approaches for selective robotic disassembly.

Computer integrated manufacturing seeks to coordinate design, planning,

scheduling and control of manufacturing operations to improve price,

customization, product development, and delivery lead-times. Extending this

concept to electronics de-manufacturing operations may reduce costs, enhance the

development and processing of components, improve various material grades

from billions of end-of-life electronics, and decrease environmental impact. The

current product flow for electronics, illustrated in figure 2.22, shows that

electronics may be disposed or de-manufactured. Disposal may include

landfilling or incineration. Electronics recycling companies may de-manufacture

thousands of different incoming products to remanufacture them for resale,

recover components for resale and/or separate materials for recycling.

Materials recovery is critical to recapturing nonrenewable resources. As the

number of discarded electronics increases, recycling legislation is enacted, reuse

options face obsolescence and price limits, data security concerns rise, and

hazardous materials recovery is promoted, the potential demand for materials

separation of billions of electronics warrants examination of the potential for

computer integrated de-manufacturing for materials recycling.

Figure 2. 22 Current product flow

P a g e | 66

Currently, third-party de-manufacturers do not receive product design

information from other entities in the product supply chain illustrated in figure

2.22. A product environmental profile may include product structure, fasteners,

material content and environmental attributes. Product design information to

support computer integrated de-manufacturing may be centralized in a company

database as indicated in figure 2.23 or decentralized in a product lifecycle unit

that acquires and stores information throughout the product lifecycle as shown in

figure 2.24.

A significant challenge in computer integrated de-manufacturing is obtaining

information about the product design and the current state of the recycling

network. Product disposition decisions in computer integrated de-manufacturing

require assessment of the product structure and materials, the current location and

quantity of inventory, and the status of operating capacity of the de-manufacturing

facilities. Product knowledge may be acquired through a product information

system, a barcode, a radio frequency identification (RFID) tag, sensor, symbols

or prior experience with the product or experimental tests.

The challenge with a product information system is that the forward supply chain

for electronics is a complex web of many different subcontractors. The product

material information approach requires capturing and coordinating information

from many different component manufacturers and preserving the link between

this information and the product in a way that is accessible to third-party recyclers

many years later. A further complicating factor is that the product may undergo

upgrades and repairs throughout its useful product lifecycle that change the

product structure and material content information.

Figure 2. 23 Future centralized information flow for product environmental profile

P a g e | 67

Figure 2. 24 Future decentralized information flow for product environmental profile in which information

is embedded in the product

Furthermore, random product conditions may occur including damage, rust, dust,

oil, missing or added components which also cause difficulty in designing robust

automated disassembly systems.

As billions of end-of-life electronics are retired, the need for efficient component

and materials recovery is escalating. When the end-of-life electronics are

organized by common materials and/or structures, there is potential for automated

processing. Product design information management may be centralized or

decentralized. In either case, up-to-date product design and material content

information need to be easily accessible to de-manufacturers. For example, even

if barcodes and product tags are attached to products at the time of manufacture,

their information may require revision during the product lifetime and the de-

manufacturer must be able to access and read them. Demand-driven selective

disassembly planning will require advances in reserve supply chain information

sharing. In addition to developing and implementing product information

systems, computer integrated de-manufacturing processes will require

information from the collection networks, product receipts inventory and

forecasts, work-in-process inventory, operating capacity, finished goods

accumulation to fill orders and part and material demand forecasts. [33]

Regarding the information flow which is needed for an efficient way of product’s

manufacturing and de-manufacturing, there are various arguments about using

automated identification and data capture (AIDC) technologies to ease the way of

transferring the data among different parties. The emergence of automated

identification and data capture (AIDC) technologies such as RFID tags and sensor

networks, specifically when it is linked to other sensor technologies and

networked databases, improves the availability and quality of information

associated with the product to which it is attached, and has the potential to

improve the effectiveness of business decisions that are made at the various

P a g e | 68

phases along the product lifecycle. Management of end-of-life (EOL) products

has been gaining importance in the past few years due to environmental concerns

as well as the widespread recognition of the financial benefits that accompany

efficient EOL product recovery. With the increasing focus of manufacturing

companies on providing products that are environmentally sound, the need for

information enabling the best decisions is becoming increasingly essential.

However, loss of information associated with a product after the point-of-sale is

one of the major obstacles for recovering value from EOL products efficiently.

Figure 2. 25 Product Lifecycle

The flow of products through their life can be represented broadly by the five

basic steps, which is then classified into three key phases of the product lifecycle

as shown in Figure 2.25 The EOL phase of a product begins when the user returns

or disposes products due to a variety of reasons ranging from malfunction to the

arrival of new and better products in the market. These products are collected

either by municipal collection facilities or through manufacturers’ take-back

programs. After collection, the products are sent either to recoverers, who execute

recovery activities which may be product, part or material recovery.

Recoverers carry out significant value-adding activities, converting the discarded

EOL product into product, parts, and materials that can be placed for sale in the

secondary market. At this stage, several options such as reuse, refurbish, recycle,

etc. are available for recovering value from these products.

However, throughout the literature, academics and industrialists have lamented

the lack of information available for product recovery decision making. It is noted

that it is often difficult for remanufacturers and recyclers to obtain product

information from manufacturers. It is claimed that the major handicap in product

recovery operations to be the lack of product data. In their study, they found that

most product recovery facilities, even in-house operations, have limited

knowledge about the product’s design and its material composition, let alone its

functional condition at the time of return. This point was reiterated by White [39]

who while studying the recovery process of computers, note that the data needed

to perform market forecasts and economic analysis of the viability of various

P a g e | 69

recovery options are often not readily available, even for OEMs recovering their

own products. To summarize, products today are not recovered at the end of their

life in an efficient and environment friendly manner because of lack of

infrastructure, excessive costs and lack of necessary information. Hence as a

matter of attaching the needed data to the products, which can be used in different

phases of product’s lifecycle, the available technologies are presented in the

following part with a brief introduction of each one and a final comparison among

them, to show their appropriate condition of usage for different products.

Three key technologies are currently, commercially, available and in use for

identifying products: barcodes, contact memory buttons and Radio Frequency

Identification (RFID).

1. Barcode Technology

Barcode labels can be classified into two categories: one-dimensional (1D)

and two-dimensional (2D) barcodes. 1D bar codes store data in the widths and

spaces of printed parallel lines (left side of Figure 2.26) and can be used to

store from six digits up to around 50 characters or 30 bytes. On the other hand,

2D bar codes store data in patterns of dots or concentric circles and can

provide an efficient medium for storing large amounts of data in a small space

(right side of Figure 2.26). The storage capacity for 2D barcodes is related to

the size of the matrix, but normally has a capacity to hold up to 3,116 numeric

digits, 2,334 alphanumeric characters or 1,556 8-bit ASCII characters.

Figure 2. 26 1D and 2D Barcode Labels

2. Contact memory buttons

Contact memory buttons are passive read/write electronic devices designed to

work in extreme operating environments. They are coin style devices housed

in rugged metal cases and come in several diameters from 8 to 30 mm (Figure

P a g e | 70

2.27). Depending on the style and size of the device, they can hold up to 8 MB

of data. The devices can support security password mechanisms and data

correction algorithms that claim to ensure accuracy of data over the 100-year

life of the device.

Figure 2. 27 Contact Memory Buttons

3. Radio frequency identification (RFID)

RFID is a technology, whereby a small transponder device (which we will call

a ‘tag’) communicates the identity or/and other information contained within

the tag via radio waves to a reader, usually by causing a unique characteristic

modulation of the carrier signal, which when demodulated by the reader,

reveals the identity of the tag which was in the field of the reader. RFID tags

can also offer additional functionality, such as on-board sensors or memory

for storing additional product data on the tag itself.

Figure 2. 28 RFID Label

P a g e | 71

It is widely acknowledged that the quality of decisions is highly dependent on the

quality of available needed information to make those decisions. In the case of

product recovery decisions, sources of information associated with the product

are not perfect. In the following part there is a review on various information

systems that support product recovery decisions, for example, product databases

or data monitors and sensors. The effectiveness of product recovery decisions

depends on the quality of product information (as provided by these information

systems and technologies) available to make the decisions.

There are different characteristics of the needed information for being beneficial

for the de-manufacturer, such as completeness, accuracy, accessibility,

interpretability, and uniqueness.

Considering the completeness of the data and information system as one of the

most important factors, the ideal information database, should be capable of

capturing the key events associated with the product along all the stages of the

product’s lifecycle (e.g. repair, usage, environment). This means that the product

information system should have the ability to rewrite the information associated

with the product as it occurs in the real world.

Table 2. 2 Comparison of ID Technologies

P a g e | 72

As it is shown in the table, different AIDC technologies are compared based on

their capabilities and exploitation limits. As a matter of fact, the key advantage

of barcode technology is that the production of barcode labels is relatively cheap.

Moreover, data standards are mature and starting to merge across application

boundaries. Due to the optical nature of this technology, it is not susceptible to

performance degradation arising from the materials that they are attached to.

Considering the contact memory buttons, it is shown that they are physically

robust devices that can be attached to products or application areas as required.

Their performance is not dependent on the material that the label is being attached

to. They have a large read/write memory storage capability, supporting security

protocols and have an operating life time of 100 years. However, contact memory

buttons are contact devices and by their nature are only suitable for semi-

automated applications.

RFID has several advantages over manual scanning using optical barcodes or

memory buttons, since many tagged items (or embedded components of a

composite product) could be simultaneously identified in an automated manner,

very quickly and without the need for line-of-sight to each item. RFID systems

are subject to several limitations that are inherent to the physical properties of RF

communications, as well as legal stipulations around their operation in different

countries. These limitations can include factors such as tag interrogation and

programming speeds, read ranges that are possible, detuning effects due to

materials such as water and metals within the vicinity of the tag or readers,

product/reader orientation limitations and interference due to electrical and

wireless communication devices in the local area. Many of the limitations listed

above have been solved and many successful RFID installations have been

achieved.

The use of RFID in improving product recovery has also been investigated by

other researchers. Zikopoulos and Tagaras [34] examined remanufacturing

operations, and they found that quick sorting enabled by such technologies is

profitable for low-quality products when the costs associated with sorting,

disposal and transportation (between the sorting center and remanufacturing

center) are low and the disassembly costs and sorting accuracy are high. [34]

P a g e | 73

3 Defining and modeling an economic

model in optimizing PCB De-

manufacturing

This is the starting point for the definition of a structure which takes the

information and data as an input and consequently provides the optimal treatment

procedure for recovery the EOL products. It is vital to analyze the market of each

EOL product’s possible outcome, in order to understand what exactly would be

needed from the market as the output of the EOL products treatments or recovery

procedures. As a result of such an analysis, it would be understood whether the

EOL product, can be repaired, reconditioned and be reused again or it cannot be

used as a second-hand product, but its components and subassemblies can be

reused and hence be sold in the market. Recycling instead of reuse is the most

inefficient way to close the circular economic loop. As we will precisely focus on

the treatment of EOL PCBs, it is crucial to consider the market of achievable

element from the PCBs. As far as the PCBs are considered as the “Urban Mineral

Resources” and contain a high level of precious metals and rare-earth materials,

the market analysis regarding the reusable PCBs and components and materials

can be considered as necessary. Thus, in the following part, we will present a brief

market analysis which can be a supportive reason for our work methodology and

why we assumed different possible outcomes for the EOL products treatment.

3.1 Market analysis

In order to cater the demand, resulted from the colossal development and

innovations in electronic industries, there is a tremendous growth of PCB

manufacturing industries across the globe. Figure. 3.1 shows the trends of the

investments in PCB industries and futuristic growth worldwide. From this trend,

it can be assumed that in the coming years, there will be a steep increase in PCB

productions and eventually it will lead to enormous waste PCBs generation. The

figure is clearly illustrated that the trend of increasing the production and because

of that, the demand for PCB, triggers to find a solution for this enormous amount

of worth in the near future. [31]

P a g e | 74

Figure 3. 1 Trends of global investment in PCB production

Figure. 3.2 shows the production trend of PCBs in different countries and regions

for the last thirty-five years. China, which shares nearly half of the total PCB

industries’ market in 2015, is the fastest growing nation in the PCBs production.

Currently, waste PCBs account for ~3% of the total E-waste generated globally.

In China alone, considering the total E-waste that is generated and imported, more

than 500,000 tons of waste PCBs needs to be treated in a single year and the

amount is growing each year due to reducing average lifetime of electronic goods.

Figure 3. 2 PCB Production in various countries

P a g e | 75

According to the fact that China has a huge amount of printed circuit boards`

production share, it would be interesting to investigate the price and cost

improvement for the last decade. The data is collected from the People's Republic

of China from 2007 to 2017, though the analysis is applicable for many non-

standardized electronic components from China. China is expected to be the main

manufacturing country of PCB for the next five years, so analyzing the printed

circuit board manufacturing industry of this country would be a good estimation

for the future production and demand of the PCB market.

The two main cost drivers of PCB, namely labor and material costs will be

addressed. An analysis of how these drivers has affected the profit margin of PCB

manufacturers in China from 2007 to 2017 is presented.

Labor cost

The central government of China has steadily increased the level of minimum

wage from 2007 to 2017 and the increase has also been higher in financial

pressure areas as e.g. Shenzhen and Shanghai. The increase in minimum wage

affected the labor cost for trades and industries who require skilled labor at a

higher extent. The wage level for experienced and educated operators in the PCB

industry, has therefore been affected above average. The majority of the

manufacturers are located in financial pressure areas and a significant part of their

workforce are trained and educated operators, so the total labor cost increase

effect is significant.

Material cost

The commodity price index depicts a price increase over the last 24 years,

however, this data does not encompass the substantial increase in commodities in

4 quarter 2016. The question is then, is this a temporary fluctuation and will this

affect the price of finished goods? The key manufacturing entities in the market

has reported that they have significantly been pressed on price over the last 5

years, to the extent that they are forced to pass the increased costs of material to

the product owners in 2017. It is expected or at this stage, it is evident that the

price of copper foil due to material shortage will be the first material in PCB to

increase in price and resin will most likely follow shortly thereafter.

Copper foil is used to produce both Copper-Clad Laminate (CCL) and Lithium

batteries. The global copper foil productivity is limited, but with the increasing

demand from electrical vehicles, more and more copper foil are used to produce

Lithium batteries and less is allocated for the CCL and the PCB industry as the

quality demand is higher and the profit margins are lower in this industry. In

P a g e | 76

recent years, the copper foil industry has also competed fiercely, this has resulted

in diminishing return on investment and some copper foil manufacturer has been

closed down, which results in a further decrease of supply and CCL manufacturers

can now not keep up with the increasing demand.

The global copper foil manufacturing capacity is app. 480,000 tons per year (t/y),

which is less than the current total demand 516,000 t/y, including CCL and PCB

which is 25% of the total capacity. Some experts predict a sharp increase in the

demand for electric vehicles due to pollution and technological breakthrough and

this will further stress the supply situation for copper foil. By 2018, the demand

of Lithium battery copper foil is estimated by some to increase by 200% - 300%

and the copper foil industry is presently expanding output and productivity by 4%

- 5% per year. It is evident that if this scenario becomes a reality the demand and

price situation will be quite challenging.

Figure 3. 3 PCB Production and Market price index 2007-2017

Conclusion:

Labor cost is increasing, automation is no longer providing the necessary

productivity gains, material is scarce, material demand is growing, and material

supply is not increasing at an equivalent rate (figure 3.3).

The profit gap (PCB market price - PCB Production Cost) for manufacturers in

China for PCB technology with 2-4-6 layers, volume production with standard

material and stack up, solder mask and 35um copper thickness has decreased

substantially over the last two years, as labor and material costs has increased.

The expected result is hence that prices will increase in 2017 and it is not evident

when and at what price level the new equilibrium will be found. [35]

P a g e | 77

Second-hand market for PCB:

Existence of a market for used printed circuit boards would be considered as one

of the thesis objectives. According to the data gathered from some reliable

websites in the field of online shopping, the market price of new brand and

second-hand boards for different types of printed circuit boards, which are used

in various electronic applications, are provided. It should be noted that all the

second-hand boards presented by the websites are refurbished and guaranteed by

the seller to have the same quality of the new one for the costumer. It was easily

predictable that the price of the refurbished products offered by the market would

be at least 50 percent lower than the new one. The following tables 3.1, and 3.2

figure out the data related to mobile phone’s boards, and motherboards,

respectively from famous brands of computer motherboards and phone producers.

Considering these information, it would be possible to consider the availability of

an active second-hand market for computer motherboard or phone boards.

Model Price of New

(€)

Reference

Seller

Price of Used

(€)

Reference

seller

Samsung

Galaxy S7 169 eBay 87 Alibaba

iPhone 5S discontinued - 48 AliExpress

LG G4 108 eBay 62 eBay

Sony Z3 discontinued - 57 eBay

Table 3. 1 Mobile phone PCB price (New and Used PCB)

Model Price of New

(€)

Reference

seller

Price of

Used (€)

Reference

seller

Gigabyte

GA-H61M-

S2PV

102.5 eBay 40.5 Alibaba

ASUS

B75M-PLUS 127.5 eBay 56 AliExpress

ASUS

P5P43TD 83 eBay 50.5 AliExpress

ASROCK

G31M-S discontinued - 13 Alibaba

P a g e | 78

MSIX99A

GAMING

PRO

574 amazon 319 AliExpress

Table 3. 2 Computer motherboard PCB price (New and Used PCB)

Component

Type

Functionality Price of

New (€)

Price of Used

(€)

Reference

seller

DIP14

Parity generator 0.52

- -

Trigger Inverter 0.24

Flip-flop 0.31

SI/PO SHR 1.04

Voltage

regulator 0.22

SOIC20

Buffer 0.35

- -

Flip-flop 0.37

Transient

voltage 1.77

Bus transceiver 0.33

QFP 132 Multiprocessor

6 - - Microprocessor

QFP 32 SRAM 1.13 - -

Table 3. 3 Electronic Package price

As it was argued and is visible in the results of market analysis, in case of precious

PCBs, such as motherboards, telecommunication, and phone PCBs, it is possible

to think of a source of revenue with a good profitability level which could be in a

much higher level compared to that of recycling. Regarding the second-hand

market for the mounted components on the PCBs, it should be said that there were

no data regarding the existence of such a market, but assuming the future

technological improvements in disassembly, test and quality assurance for the

components, also it would be possible to consider the components as demanded

from the market and structure the economic model based on the hypothetic idea.

P a g e | 79

3.2 General PCB treating work description in de-

manufacturing

To find out how to exploit the data and information (either from manufacturer or

market) and help the recycling or de-manufacturing processes to move forward to

be more efficient and optimized, we had to assume a logical treatment procedure

and based on which, define the possible needed data for each step.

Firstly, it is noteworthy to clarify the state of the assumed EOL products which

are assumed as the objects of various recovery treatments. An EOL product can

be provided either with an AIDC label which provides all the needed data to the

de-manufacturer, or without any information provider label. The aim of defining

this work methodology is to define the needed data which should be provided in

an informative label for optimizing the EOL products’ de-manufacturing, hence

we assumed that there is no label or any provided data, attached to the PCBs.

Hence with this procedure, also it would be possible to think about the present

products in their mid of life cycle and those products which are already in their

end of life cycle. The importance of thinking about the presence or absence of an

AIDC label on the products is that they will influence the design of the set of

treatment activities on the PCBs because in case of presence of an informative

label, some cognitive steps e.g. “Image analyzing” to recognize the product can

be ignored.

After defining the possible treatment structure with the specified working and

processing stations, it could be possible to think about each station’s needed data.

Afterwards, by knowing what would be beneficial for each step, it could be easier

to link the data and formulate an objective function to maximize the de-

manufacturing profit and get a final optimized result.

We set and arranged a general work methodology to base our economic model on

it. The underlying reasons of the main structure of the mentioned methodology is

based on the literature review and market analysis. Based on the literature review,

it is claimed that thinking about reusing the functioning electronic components

should be investigated, because most of them are useable and intact at the time of

their disposal. Also, regarding the results of the market analysis the increasing

demand for used functioning PCBs in computer or mobile phone industry is

visible. The mentioned reasons were convincing enough to make a very general

structure or platform on which it could be possible to execute every possible

scenario for every possible industry that the PCB is coming from. The work

methodology should have been drafted in a way that could help us, in defining

what data would be needed and which player in the circular economy loop

P a g e | 80

(manufacturer or market or …), could provide it. As regard as the complexity of

the work and wide variety of possible scenarios in treating a PCB based on its

condition, various assumptions and hypothesis have been made to make it simpler

and easier to be executed.

To better illustrate the work methodology, we have decided to break it into sub-

divided sections which are split by decision making steps in the work description.

The main paths which a board can pass through are mainly four and they are

selected based on the board’s status and condition which defines its following

rout. The main four paths are explained more in detail in the following:

1. A PCB can be intact, and its reuse is considered as the most profitable

choice

2. A PCB can be defective, and its repair and reusing is considered as the

most profitable choice

3. A PCB can be either intact or defective but reusing its functioning

components is the most profitable choice

4. A PCB can be either broken or intact or defective and its recycling profit

prevails over itself or its components’ reuse

Regarding the high cost of test, repair and selective disassembly, among the four

possibilities that a board could have, the first three scenarios are less common for

the most of the EOL PCBs which are returning to de-manufacturing and recycling

companies, but the growth of technology and the demand for used functioning

boards and components, are convincing enough to have a general map covering

all the possible scenarios that could happen in order to step towards the most

optimized circular economic plan for PCBs. Regarding the fourth path, which is

the most common treating way for the EOL PCBs, it should be noted that the lack

of knowledge and economic treating technologies have resulted in a not optimized

way of treatment which neglects all the possible profits coming from reusing or

remanufacturing the boards or its components.

Before going more into detail for illustrating the possible routes of treating a PCB,

it would be better to explain the main points and steps used in the work

description. To be more precise and accurate in finding the most optimized

decision that could be made for a PCB, we have assigned several steps to get data

from a PCB, analyze the achieved data and finally decide what to do with it based

on its potential value, condition, and status. In the following, we have explained

each specific working or processing station which is available in the work

description. It should be reminded that different treatment paths of PCBs could

be composed of repetitive steps, for this reason, we have first presented the steps

P a g e | 81

without any specific order and afterward, we have introduced different possible

treatment paths, by just mentioning the name of different steps. Furthermore, there

would be the table of needed data for each step which could be helpful and

essential to perform either operational or processing steps. It should be said that

all the mentioned data are used in either economic model for the aim of de-

manufacturing optimization, or to perform the operations more efficiently. The

following table provides the abbreviation of each operative or decision-making

step which is depicted in the presented de-manufacturing algorithm.

Abbreviation Operational or Processing Step

BI Board Inspection

BGC Board General Cleaning

IP Image Process

1st BT First Board Test

2nd BT Second Board Test

1st CA First Cost Analysis

2nd CA Second Cost Analysis

BFC Board Final Cleaning

BP Board Packaging

SDIC Selective Disassembly of Intact Component

SDDHCMC Selective Destructive Disassembly of High Concentrated Material

Component

SD Simultaneous Disassembly

CQC Component Quality Control

CP Component Packaging

CR Component Recycling

BBR Bare Board Recycling

SDDC Selective Disassembly of Defective Component

BPC Board Partial Cleaning

ANC Assembly of New Components

Table 3. 4 List of operating and decision-making steps’ abbreviation

P a g e | 82

Figure 3. 4 The work description of PCB treatment’s possible streams

P a g e | 83

a) Board inspection (BI)

The starting point would be the very basic data provider and decision-making

step for our work, hence the inspection step task would be the identification of

any deviation from the state of a not-damaged PCB. So, any cracked, broken

or damaged board will be identified as a not reusable board, but it could be

used as the source of reusable components. As a result, the Board Inspection

step as a processing step, makes the decision and based on the severity of

damage, the inspector decides whether it should be sent directly to recycling

steps, or cleaning step. Another task which is assigned to the inspection step is

to provide two needed data for the final cleaning step. These needed data are

the a. Contamination type and b. Contamination level, in order to properly

perform the “Final Cleaning” task. Considering the needed data to perform this

task, in the following table, the data coming from different players

(Manufacturer, Market, and De-manufacturer) and also the output data of the

step is provided.

Step Manufacturer

data

Market

data

De-

manufacturer

data

Output data of operational

or processing step

BI - - -

Damaged PCBs

Not Damaged PCBs

Contamination Type

Contamination Level

Table 3. 5 The needed data for the “Board Inspection” step

b) Board general cleaning (BGC)

Board General Cleaning step as an operational step and considering a board

which has passed the “Inspection Step”, a general cleaning activity would be

needed to prepare the board for the following step which captures the image of

the board and processes it with the archived images of boards that the

manufacturer has provided the de-manufacturer. In order to do so, a general

cleaning would be needed to blow away the available dust on the boards and

to make it easier for the “Image Process” step to recognize the boards and their

components. In case of doing this task, there would be no need for any data

from none of the players in the loop.

c) Image process (IP)

The recognition of a PCB, using an image process system, has several steps

which starts from the very first classification of a PCB to identify which

classification of PCBs, can be assigned to the processed PCB. The following

P a g e | 84

level of processing the PCB’s image is to recognize just the type of the defined

packages on the PCB (SOIC, QFP…). For the first level of PCB’s

classification, the PCB will be analyzed by the system and based on its physical

characteristics like dimension and weight, it will be classified as phone’s PCB,

computer’s motherboard…). The first step is based on a knowledge-based

algorithm. The output would be the classification of the PCBs inside big

classes: Motherboard, PC, Phone…. The reason of doing such a classification

is that we can have a first approximate idea of the value of the PCB. As the

improved level of PCB recognition system, the processing system can also

provide the information regarding the PCB’s packages, by which it would be

possible to know the quantity of different types of available components on the

board. As a novel idea in case of image processing, we have assumed that while

the system is fed with the data from manufacturer including the image of its

PCBs, the image processing system could be able to compare the acquired

image with the archived ones and as it matches with one of them, the data

regarding the archived image, also could be used for the PCB which is under

process.

The objective of acquiring the PCB’s image and subsequently process and

compare it with the archived ones (Provided by the manufacturer), is to answer

these questions:

a) If the PCB is among the market demanded PCBs or not?

b) If the PCB has the market demanded components or not?

c) What are the missing components? (Considering the need to repair a

PCB)

d) What are the high concentrated material components (HCMC)?

(Considering the need to be disassembled and recycled)

The flow of the inspected and generally cleaned boards, enter the “Image

Acquisition” step and after capturing each PCB’s image, the image will be

processed and compared with an archived set of PCBs’ images that are

provided by the manufacturer. Based on the image process, three possible

scenarios can happen, 1. The PCB is identified as the market demanded PCB,

2. The PCB is identified as not demanded by the market but has the market

demanded components, 3. The PCB is neither demanded nor has the demanded

components. Considering the occurrence of first two scenarios, the PCB would

be processed by the following “Cost Analysis” step to understand whether its

treatment would be profitable or not. In case of the occurrence of the third

scenario, the PCB will be sent to the recycling steps.

P a g e | 85

Regarding this concept, a project in laboratory scale in STIIMA-CNR is

executed by using an optical vision system which is able to recognize different

electronic packages from a 2D image. Importance of this system is that it can

analyze the PCBs in-line and give information about the grade and assembled

components. In this way it is possible to decide what processes to use for

different type of boards, improving the effectiveness of recycling.

Notably, this important step is still missing in the process flow because

available technologies do not allow effective industrial applications or require

unbearable costs. [2]

Considering the hypothesis and assumptions that we have made in the image

acquisition step, it is considered that the accuracy of the image comparison and

process is high enough to give us the data about the presence or absence of

defined components and also the number and type of present packages on the

board. In the following table 3.6, the data which should be provided for the

“Image Process” step is mentioned.

Step Manufacture

r data

Market

data

De-

manufacture

r data

Output data of

operational and

processing step

IP

PCB's image

Demanded

PCB's

standard

code

-

Demand status of PCB (the PCB

is demanded or Not)

Electronic

components

functionalities

Demanded

component'

s standard

Codes

The Complete Set of packages'

data, in case of matching the

archived PCB images

PCB's code - The present quantity of demanded

components

Packages' types - The quantity of HCM packages

Packages'

standard codes

-

The estimated material content of

PCB's packages

- The type of packages, (when the

PCB is not in the archived PCBs)

-

The classification of the PCB

The estimated placement of HCM

components, in case of not

matched with archived images

Table 3. 6 The needed data for the “Image Process” step

P a g e | 86

d) 1st Cost Analysis (1st CA)

Considering the PCBs that are considered as demanded by the market or having

the market demanded components, the first cost analysis step, would be needed

to calculate their related costs and potential revenues. In order to do such an

analysis, we will need the output data of the “Image Process” step (missing

components, present HCM components …), as the input data. By knowing the

cost of each specific PCB’s needed process (testing, component disassembly,

component assembly …), it would be accurately possible to estimate the

treatment cost for each PCB and finally make the decision for sending it to

either the “Testing” or “Recycling” steps. Hence it is not known whether the

board is functioning or not, its components work properly or not and etc., the

main assumption should be made. The assumption that we made in analyzing

the cost of treating a PCB is that from the 4 main possible streams of the

boards’ treatment, 2 main streams can be neglected since their treatment profit

would be between the other two streams. As a result of which, we would have

the other two main streams which have the extremum profits. Hence to make

it clearer, we will set different names for the different streams’ profits, and

show it in a simpler way:

First category of PCB treatment (Reusing aimed treatment)

Profit of 1st Main Stream (Reusing the Intact PCB): P1

Profit of 2nd Main Stream (Reusing the repaired PCB): P2

Profit of 3rd Main Stream (Reusing the Intact Components of PCB): P3

Second category of PCB treatment (Recycling aimed treatment)

Profit of 4th Main Stream (Recycling the PCB): P4

As it is shown, the 4 main streams are divided in two main categories.

Regarding the first category, its 1st main stream’s profit (P1) would prevail the

profit of other two streams (P2 and P3), since they bear some additional costs

(disassembly, assembly, testing …) which the 1st stream does not have. Hence

based on our assumption, P1 is greater than P2 and P3. Thus, in the cost

analysis, we should compare the P1 and P4 to decide about sending the PCB

to testing step or recycling. Consequently, if P1 will be greater than P4, the

PCB should be tested in “Testing” step and based on its functionality status, it

would be analyzed in the second cost analysis step. On the other hand, if P4

will be greater than P1, for sure it is also greater than the other profits (P2 and

P3) and will be sent to recycling process. Furthermore, it has to be mentioned

P a g e | 87

that for calculating the profit of the 1st main stream, it is assumed that the PCB

is functioning and can be sold in the second-hand PCB market.

Step Manufacturer

data

Market

data

De-

manufacturer

data

Output data of

processing step

1st

CA

Bare Board's

mass (Kg)

Price of second

hand PCB PCB's testing cost

The most profitable

stream among reusing

or recycling streams

Components

mass (Kg)

Quantity of each

type of

demanded PCB

PCB's final cleaning cost

-

-

Price of second

hand component PCB's packaging cost

Quantity of each

type of

demanded

component

The Estimated Material

Content of PCB's Packages

(IP step output)

-

The Quantity of HCM

packages (IP output)

Unitary cost of selective

destructive disassembly of

HCM component

(€/component)

Unitary cost of

simultaneous disassembly

of component (€/PCB)

Unitary cost of component

recycling (€/Kg)

Unitary cost of bare board

recycling (€/Kg)

Table 3. 7 The needed data for the “1st Cost Analysis” step

e) Board Testing (BT)

The boards which have passed the cost analysis and reached the “Testing” step

are the demanded boards or those ones with demanded components that

treating them is profitable even if after test step, they will be identified as not

functioning and should be recycled. The main assumption of the test process

is the ability of this step to define 1. State of the functionality of the board, 2.

State of the functionality of each demanded component which is present on the

board. Also, due to electronic issues sometimes the components’ test should

be done after their disassembly. Hence, the output of this step will update the

data and give the precise condition of the PCB, in order to replace the assumed

assumption which, we made in the previous cost analysis step, with the

accurate data. Hence the result of testing the PCB would be 1. Functioning

board, 2. Not functioning board with the specific functioning components (the

components which will be tested, are those ones which are considered as

P a g e | 88

demanded components). Considering the requirement of this step, the testing

procedure of both the PCB itself and also its components should be provided

by the manufacturer, in order to make it possible for the de-manufacturer to

test the PCBs based on it and adjust the testing variables as the precise

measures suggested by the manufacturer.

Figure 3. 5 An Automatic PCB Testing machine in STIIMA-CNR

As far as the data is concerned, the following table 3.8, is showing the crucial

needed data which should be provided for PCB’s test activity. The test of a

PCB should be performed based on a procedure which is known by its

manufacturer. Also, the adjustable testing variables and the precise measures

should be mentioned by the manufacturer in order to let the de-manufacturer

to perform the testing task respecting the needed steps.

Step Manufacturer

data

Market

data

De-

manufacturer

data

Output data of operational

and processing step

BT

PCB's testing

procedure

- -

The status of demanded PCB's

functionality (Does the PCB work

properly or not?)

PCB's testing

variables (voltage,

amperage)

The status of demanded component's

functionality (Does the demanded

component work properly or not?)

Component's

testing procedure

Indicate the place of defective

components, in case of not

functioning the demanded PCB

Component's

testing variables

(voltage,

amperage,

capacitance,

resistance)

The quantity of defective

components

Table 3. 8 The needed data for the “PCB Testing” step

P a g e | 89

f) 2nd Cost Analysis (2nd CA)

As regard as the second cost analysis, it is noteworthy to mention, that the

analysis will be done for both streams of the tested PCBs (functioning and not

functioning), in this way, two possible routes could be considered for treating

the PCB:

A. If the PCB passes the testing step with the result of Functioning PCB,

the cost analysis will analyze the profit of treating such a board

considering three possible alternatives which are: 1. Final cleaning and

packaging (profitable to reuse the board), 2. Selective disassembly of its

functioning components and recycle the rest of the PCB (profitable to

reuse its components), 3. Selective destructive disassembly of its HCM

components and recycle them separately from the rest of the PCB

(profitable for recycling)

B. If the PCB passes the testing step with the result of failed PCB, the cost

analysis will analyze the profit of treating such a board considering three

possible alternatives which are: 1. Selective disassembly of defective

components and prepare the board for repair (profitable to repair and reuse

the board), 2. Selective disassembly of its functioning components and

recycle the rest of the PCB (profitable to reuse its components), and 3.

Selective destructive disassembly of HCM components and recycle them

separately from the rest of the PCB (profitable for recycling)

Step Manufacturer

data

Market

data

De-

manufacturer

data

Output data of

processing step

2nd

CA

Quantity of each

type of package

Price of second

hand PCB PCB's testing cost

The most profitable stream

among reusing intact PCB,

repairing the PCB and reuse

it, reusing the components or

recycling streams (send to

"Test" step or not?)

Bare Board's mass

(Kg)

Quantity of each

type of

demanded PCB

PCB's final cleaning

cost

The revenue of each treatment

stream

Components mass

(Kg)

Price of second

hand component

PCB's packaging

cost

The cost of each treatment

stream

Component's

material content

Quantity of each

type of

demanded

component

The Estimated

Material Content of

PCB's Packages (IP

step output)

The profit of each treatment

stream

-

The price of

new

components to

be assembled

The Quantity of

HCM packages (IP

step output) -

- Unitary cost of

selective destructive

P a g e | 90

disassembly of

HCM component

(€/component)

Unitary cost of

simultaneous

disassembly of

component (€/PCB)

Unitary cost of

component

recycling (€/Kg)

Unitary cost of bare

board recycling

(€/Kg)

Unitary cost of

selective

disassembly of

intact component

(€/component)

Cost of component

quality check

(€/component)

Cost of component

packaging

(€/component)

Unitary cost of

cleaning a

component's place

(€/component)

Unitary cost of

assembling new

component

(€/component)

Table 3. 9 The needed data for the “2nd Cost Analysis” step

g) Board Final Cleaning (BFC)

The boards which had been identified as demanded (after image process) and

had passed the testing step with the satisfactory result and finally had been

selected as profitable to be reused, will enter final cleaning step. The

preparation for packaging the PCB regarding its cleanness should be

performed in this step. As regard as the cleaning task, and the cleaning time, it

is assumed that the cleaning task is performed based on the type and level of

PCB’s contamination. Hence the data about these two concepts, which are the

outputs of “Inspection” step, can be used to decide on how to perform the

cleaning activity (which cleaning material should be used for specific

contamination type and level?). Regarding the contamination types and levels

which is possible to be observed on the PCBs, based on the literature review,

the most common types could be either dust (or other dry, free flowing solids)

or grime (dust/dirt mixed with something tacky - oil, wax, etc.). To remove

P a g e | 91

dust or other light debris from a printed circuit board it is generally advised to

simply brush the substance with a delicate brush and use a vacuum cleaner or

air blower, to blow them away from the PCB. Grime or greasy dust is

particularly common on monitor chassis boards where high temperatures cause

outgassing of components coated with wax. The wax forms a sticky film on

the board which traps dirt easily and since the high-voltage potential of the

monitor anode will attract airborne particles, monitor boards can become

particularly filthy. Other boards will sometimes become encrusted with dirt

and gunk that simply won't brush away, so more aggressive cleaning can be

become necessary. Grime that appears to be mixed with some form of oil or

wax, will need some form of detergent applied followed by mild scrubbing to

expunge the offending contaminants.

Figure 3. 6 Before and after PCB cleaning

As regard as the required data for this step, the need for data would be satisfied

by the output data of the former step “Board Inspection” which is a part of de-

manufacturer’s PCB treatment line.

Step Manufacturer

data

Market

data

De-

manufacturer

data

Output of the operational

step

BFC - -

Contamination type

(BI step output) -

Contamination level

(BI step output)

Table 3. 10 The needed data for the “Board Final Cleaning” step

P a g e | 92

h) Selective Disassembly of Intact Components (SDIC)

The disassembly of the demanded functioning components should be

performed in this step. Two main flows of PCBs are assumed to enter this step:

1. The PCBs, which have been recognized as containing demanded

components (by image process) and have passed the “Board Testing” step with

a satisfactory result, regarding the functionality of the PCB (hence its

components also are intact), 2. The PCBs which have been recognized as

containing demanded components and have passed the “Board Testing” step

with the NOT satisfactory result, regarding the functionality of the PCB but

satisfactory result regarding the functionality of its demanded components.

Regarding the objective of disassembly activity, it should be noted that the

components which will be disassembled, their quality will be controlled in the

quality control step to be sure that they are in a good condition after

disassembly, in order to enter the market. Hence the disassembly task should

be done selectively and accurately. As mentioned before, different components

have different types of joint types to the PCB, hence the joint type of each

component which identifies its disassembly technology should be provided as

one of the input data for this step. Regarding automatic disassembly of

components, also the precise coordination of the components’ place on the

board and their dimensional characteristics and maximum tolerable

temperature would be needed as input data.

Figure 3. 7 Automatic PCB component disassembly station at STIIMA-CNR

The needed data (Table 3.11) for performing the “Selective Disassembly of

Intact Components” could be provided by the PCB’s manufacturer, and as it

was mentioned before, the “Components Place Coordination” would be

beneficial for the disassembly task, just in case of having automatized

disassembly system. By which it means, both the de-soldering and picking of

P a g e | 93

the component could be performed by an automated CNC system which can

use the coordination.

Step Manufacturer

data

Market

data

De-

manufacturer

data

Output data of the

operational step

SDIC

Component's placement

code

- - -

Component's placement

coordination

Component's

temperature resistance

Component's

connection type

Table 3. 11 The needed data for the “Selective Disassembly of Intact Components” step

Another method for de-soldering the components which do not harm neither the

component nor the board, is using a low temperature soldering material to help

the de-soldering procedure. The mentioned technology uses a hot air nozzle to

preheat the board and the low temperature solder wire would be melted on the

package’s leads and as it unifies the base component’s solder, it can be removed

by a suction probe. The manufacturer claims that the temperature will not exceed

150 Celsius degrees and as an average, the package can be removed in less than

180 seconds. [36]

i) Selective Destructive Disassembly of HCM Components (SDDHCMC)

Regarding nominating this disassembly step as “Selective destructive

disassembly”, it should be said that we nominated this disassembly step as

“Selective” since the specific components (e.g. big integrated circuits) which

have been identified as HCM Components (the high value components which

their material value is higher than their disassembly cost and are recognizable

by image process step) are going to be disassembled and not all the components

in the same disassembly step. Furthermore, as the objective of this disassembly

task is to collect HCMCs to increase the grade of recycling materials (their

precious material content is high, so, there would be a greater amount of

precious materials in a specific amount of shredded HCMCs comparing the

similar amount of shredded normal components). We also called it

“Destructive”, since the state of intactness of the component after disassembly

would not be important as it will be recycled and not reused again. Two main

flows of PCBs are assumed to enter this step: 1. The PCBs, which have been

recognized as demanded boards or containing demanded components (by

image process) but have passed the “Board Testing” step with the NOT

P a g e | 94

satisfactory result, hence the only profitable task in treating them, would be

using their HCM components for recycling and increasing the material grade,

2. All the other streams of PCBs which do not have any profitability in case of

reusability, thus they have to be recycled.

Regarding the needed data to perform this task more effectively, the

connection type, coordination of the place and also the dimensional

characteristics of the components can be useful to use the proper application

for de-soldering or picking the components. As it is mentioned in the table 3.8,

if the under-processing PCB has not been introduced to the processing system,

hence its precise data regarding the components’ places coordination and

connection type are missing, thus the de-manufacturer could use the output

data of its “Image Processing” station in order to provide the approximate place

and connection type of the target component which is going to be

disassembled.

Step Manufacturer

data

Market

data De-manufacturer data

Output data of the

operational step

SDD

HCM

C

Component's

place code

-

The estimated placement of

packages, if the PCB’s image

is not matched with archived

images (IP step output) -

Component's

place coordination -

Component's

connection type

Component’s

dimensional

characteristic

- - -

Table 3. 12 The needed data for the “Selective Destructive Disassembly of HCM Components” step

j) Simultaneous Disassembly (SD)

Simultaneous disassembly considers all the PCBs that are identified as either

1. Not demanded or not containing demanded components, or 2. PCBs that

although are demanded or contain demanded components, are not profitable to

be treated in various steps, and also have passed the “Selective destructive

disassembly” step. Hence the complementary treatment for the objective of

increasing the grade of recycling material, would be the simultaneous

disassembly of all the remaining components on the PCB.

P a g e | 95

Figure 3. 8 Before and after simultaneous disassembly of the PCB

Since for the simultaneous disassembly of the components from the PCB, there

is no need to be cautious about damaging the components or precisely

disassembling a particular component, hence just the dimensional

characteristic of the board is important to adjust the disassembler machine’s

feeder. The PCB’s components simultaneous disassembler machine which is

commonly used nowadays, consists of two rolling blades which with a

sheering force, separates all the components from the board in a mechanical

way, thus its feeder should be adjusted for different thicknesses, width and

lengths of boards.

Step Manufacturer

data

Market

data

De-manufacturer

data

Output data of the

operational step

SD PCB's dimension - - -

Table 3. 13 The needed data for the “Simultaneous Disassembly” step

k) Component Quality Control (CQC)

The aim of this step would be the final check of the intactness of the demanded

components which have been selectively disassembled. So, the incoming flow

of this step would be the output of the “Selective disassembly of intact

components”. Hence, as a result of the testing process, the component could

be either functioning and working properly or defective and not appropriate to

be sent to the market. Based on the quality check result, the component could

be sent to the final “Packaging” step or to the “Recycling” step. As regard as

the data which could be needed for this quality check point, it is noteworthy to

mention that the test of the component to understand whether it works or not,

should be done based on a procedure which is standard and is in hand of its

manufacturer, hence the “Component Quality Control” step based on the

P a g e | 96

provided testing procedure could properly perform the test task and give the

final result as the output data (Table 3.14).

Step Manufacturer

data

Market

data

De-manufacturer

data

Output data of the

operational step

CQC Component's

testing procedure

- - The status of demanded

component's functionality

(Does the component work

properly or not?) Each component's

testing variables

Table 3. 14 The needed data for the “Component Quality Control” step

l) Selective Disassembly of Defective Components (SDDC)

Considering the PCBs that should be repaired, one of the main activities that

should be performed, is to precisely disassemble the defective components

which cause the problem and prepare the PCB for the following steps of

reconditioning. This step needs a high level of accuracy and attention since the

disassembly task can cause other defects in the PCB. As regard as the needed

data for this step, it should be noted that the disassembly task is exactly as

precise as the “Selective disassembly of intact components” which was

explained before. Hence the needed data are the same as needed data for

disassembling the intact components.

Step Manufacturer

data

Market

data

De-

manufacturer

data

Output data of the

operational step

SDDC

Component's place

code

- - -

Component's place

coordination

Component's

temperature resistance

Component's

connection type

Table 3. 15 The needed data for the “Selective Disassembly of Defective Components” step

m) Board’s Partial Cleaning (BPC) (Component assembly preparation)

The flow of those PCBs that had been recognized as repairable, after passing

the “Selective disassembly of defective components” step, will enter this

station to be partially cleaned and prepared for the assembly task. To explain

more in detail, the place of those disassembled components which should have

been replaced with the new ones will be cleaned and prepared for “Component

assembly” step. Regarding the needed data for this step, it is noteworthy that

component’s place code can be helpful to easily find where to prepare for

component assembly. Regarding the number of the places that the cleaning

P a g e | 97

activity should be performed on, the “Testing” step should provide the proper

data.

Step Manufacturer

data

Market

data

De-manufacturer

data

Output data of the

operational step

BPC Component's place

code -

The number of

defective components

(from BT step)

-

Table 3. 16 The needed data for the “Board’s Partial Cleaning” step

Figure 3. 9 Board’s partial cleaning and preparation for assembly

n) Assembly of the New Components on PCB (ANC)

The PCBs which are going to be repaired, after passing the disassembly and

cleaning steps, then should enter the “Assembly of new components” step. The

most important thing regarding the assembly of a new component is to

assemble a new component with the same functionality, in the place of the

previous disassembled or missing one. Hence the knowledge and data about

the functionality of that specific component and also its coordination and

tolerable soldering temperature should be provided as the needed data, by the

manufacturer. The place and the quantity of defective components can be

provided by the “Board Testing” step, which has identified the problem of a

PCB.

P a g e | 98

Step Manufacturer

data

Market

data

De-manufacturer

data

Output of the

operational step

ANC

Component's place

code

-

The place of defective

components (BT)

-

Component's place

coordination

The quantity of defective

components (BT)

Component's

temperature

resistance

- Component's

connection type

Component’s

functionality

Table 3. 17 The needed data from different players and the output data from the

“Board’s Partial Cleaning” step

3.2.1 Main PCB’s treatment flows in the work description

As the main steps of the workflow have been explained, it is possible to illustrate

the general possible routes which pass these steps. Considering the main four

routes that have been mentioned before, we are going to clarify each one and

explain a sample PCB’s assessments to show the decisive causes which put the

PCB in a specific treatment way.

Figure 3. 10 The main assumed flows of the waste PCBs treatment

1. A PCB can be intact, and its reuse is considered as the most profitable

choice

The simplest scenario which could happen is that a PCB which has been

requested by the market as used or second-hand PCB, is available in the

returned EOL PCBs arrived at the de-manufacturer with a functioning

P a g e | 99

condition. In this case, the PCB would pass all these following decision-

making steps with the satisfactory result: (1.Board Inspection, 3. Image

Process, 4.First Cost Analysis, 6.Second Cost Analysis related to the

functioning PCBs), thus the route which these boards pass, would be as

follows: the PCB enters the inspection step and will be assessed and its general

condition will be ok since there is not any crack or broken part in it, hence, it

will be qualified as “Not damaged PCB” and enters the second stage in which,

it will be almost clean and without any sever dirtiness. As the PCB has passed

the “Cleaning Step”, it will be processed in the “Image Process” station and it

will be identified as one of the demanded PCBs. After that it is conceived that

the board is demanded, it will be evaluated by the “First Cost Analysis” step

with its assumed assumptions, to be understood if it would be profitable to treat

this board in testing, cleaning, and packaging steps, or not. In this scenario, the

PCB will be identified as “Profitable to be Reused”. As there wasn’t any test

or functionality checker with definitive result till now, so the PCB will be sent

to the “Testing Step”. In this case, the board is functioning and can continue

on its way. Also, in the “Second Cost Analysis” which works with definitive

data gotten from “Test” step, it will be identified as “Profitable to be Reused”.

Hence after a final thoroughly cleaning, the PCB will be packaged and sent to

the market.

2. A PCB can be defective, and its repair and reusing is considered as

the most profitable choice

The second main route would be for those PCBs which are in the same market

demand position as the first scenario’s PCBs are, but regarding their

functionality condition, they have been delivered to the de-manufacturer with

NOT functioning condition. Hence as far as the passed decision-making steps

are concerned, they will pass: (1.Board Inspection, 3. Image Process, 4.First

Cost Analysis, 6.Second Cost Analysis related to NOT functioning PCBs),

considering these PCBs, their passing route would be: The PCB would pass

the “Inspection” step with the “Not Damaged PCB” result, after being cleaned

and processed by the “Image Process” and identified as demanded PCB, it

should be evaluated by the “First Cost Analysis” with its assumed assumptions,

and in this case, the result would be “Profitable to be Reused”. After that the

PCB is identified as profitable to be reused, it should be tested and regarding

in this case, the PCB would be recognized as not-functioning but repairable. In

the following “Second Cost Analysis” step, it will be evaluated if the revenue

of selling the refurbished and repaired PCB would be high enough to cover all

P a g e | 100

the Testing, Cleaning, Disassembly, Assembly, Packaging costs or not.

Therefore, the PCB will be sent to “Selective Disassembly of Defective

Components” step and while the defective parts are removed, it will enter the

“Partial cleaning” step to be prepared for repairing and assembling tasks,

which is mounting the new components on the PCB. After all these steps, the

PCB should be tested again in the “Testing Step” to pass the market’s needed

quality and as it passes the test step with the satisfactory result, it will be

cleaned thoroughly and packaged in the last steps.

3. A PCB can be either intact or defective but reusing its functioning

components is the most profitable choice

According to the third mainstream of PCBs, it should be said that the incoming

PCB itself, does not have the market demand but some of its components are

requested by the market. Considering a PCB which is both demanded by the

market and also contains some demanded components, the decision to send it

for PCB market or Component market, will be made in the cost analysis step

which evaluates the profit of selling the whole board or disassembled

demanded components and sell them separately. To name the decision making

steps that the PCBs pass: (1.Board Inspection, 3. Image Process, 4.First Cost

Analysis, 6.Second Cost Analysis related to functioning PCBs or 6.Second

Cost Analysis related to NOT functioning PCBs), in this case the PCB which

is inspected as not damaged, will be cleaned generally and enter the “Image

Process” step, the result of the “Image Process” step would assure the presence

of demanded components on the PCB, hence in this case the following “First

Cost Analysis” step which evaluates the profitability of different treatment

choices, will indicate the component reuse, as the most profitable choice to be

selected. Thus, the PCB will be sent to the “Testing” step, to find out the

definite status of components condition, which have been marked as

demanded. As the test step provides the definitive result about the components’

conditions, the “Second Cost Analysis” step will reevaluate the profitability of

alternative choices, so regarding this case, the most profitable choice would be

treating the PCB to reuse its components. The PCB should be sent to the

disassembly station to disassemble its intact components. As the components

are disassembled, the following “Quality control” step will take care of testing

the component to assure its functionality. The components which pass the

quality check with the satisfactory result, will be sent to packaging and finally

to the market. And those quality failed components will be sent to recycling.

P a g e | 101

4. A PCB can be either broken or intact or defective and its recycling

profit prevails over itself or its components’ reuse

As the recycling concept is concerned, there are several streams that end with

the recycling steps. Thus, the possible scenarios that PCBs will go for being

recycled are 1. The boards which are broken or burnet, and are rejected by the

first inspection step, hence will be sent to the recycling steps, 2. The boards

which after being processed by Image process step, are identified as neither

demanded nor containing demanded components, 3. The boards which after

passing Inspection step and Image process (are considered as demanded PCBs

or containing demanded Components) but in the first cost analysis step, they

are recognized as not beneficial to be sent forward, hence they follow the

recycling procedure, 4. Those PCBs which have been identified as demanded

PCB or containing demanded components and also passed the test step with

the satisfactory result but after analyzing the expenses that should be paid, the

recycling would be selected as the most profitable choice, and 5. Those PCBs

which have been recognized as demanded or containing demanded

components but have passed the test step as defective boards, and finally their

recycling is selected as the most profitable choice from the second cost analysis

step.

3.2.2 The need for a collected database about each PCB

Considering the assumed treatment flows of the returning EOL PCBs, it is

obvious that a wide range of data and information are needed to efficiently

perform the needed tasks in both working and processing stations. In this part, we

would focus on the PCBs data gathering and the reason for their presence in our

suggested database, which the manufacturer could provide.

Considering PCB’s circular economy and its main role players (manufacturer,

EOL products collector, de-manufacturer and recycler), the presence of

information, its availability and advantage for the subsequent player can make

PCB’s circular economy more and more optimized, efficient and profitable, and

hence, give better economic, environmental and financial results.

Nowadays there is not such a set of collected data related to EOL PCBs which

could serve the recycler to understand the PCBs’ potential financial value, to help

system optimization and individual operational tasks. Thus, a common way to

evaluate the potential value of a board from de-manufacturing and especially

recycling point of view will be done just by taking an overall look over the board

and grading them as low, medium or high grade. As it can be understood, this way

P a g e | 102

of board’s grading and assessment is not based on a precise measurement or

knowledge-based method, so it can affect the recycling process and make it as an

inefficient work.

3.2.3 The structure of the database and its different sections

As it was discussed at the first part of this chapter, the data that is needed to be

used in the de-manufacturing process, comes from different parties, thus it is

needed to be understood that this need of information is something different from

the data which a manufacturer holds and can provide. As an example, the material

content of each component on the PCB is not known by the manufacturer of the

board, but it is mentioned as the needed data. So, it should be noted that the

following proposed set of data does not consider whether the manufacturer holds

the data and easily can provide it or should ask from its upstream suppliers. As a

matter of fact, the set of proposed data which is essential for optimizing the de-

manufacturing processes, cannot be acquired for free for the manufacturer, but

considering the beneficial result of the efficient de-manufacturing process and the

margin which could be generated, it can be considered as an investment. The

PCB’s proposed collected database contains the information which can be

beneficial for different aspects of de-manufacturing or particularly recycling

processes, such as being beneficial in economic, operational, financial and

environmental point of views. In this case, the structure of the database and the

need for such an information base will be presented. it is vital to remind that some

of the proposing data are needed to help with the recognition of the EOL product

which is going to be processed, such as the products picture, weight, and size that

are used in the “Image Processing” step and are considered as the source of

comparison of the under-process product and its original one before entering the

consumption phase.

a) PCB’s Code and picture

The basic information for identifying and recognizing a PCB can be its image

and more specifically its code number. In order to understand which board the

collected data are referring to, especially in the case of process automatization.

By using automatic recognition system for the boards, a reference image is

needed to be used for matching the other same type of incoming boards with

the reference one, and as a result, identify it as the same board to the one in the

database. Meanwhile, the database should be completed by another data which

shows the “Market Demand”, regarding the PCB or each individual electronic

component. Hence the boards can be identified, and their demand status can be

P a g e | 103

understood. Furthermore, in order to give a better view of the board and its

components’ types, a picture can give a better understanding by giving a visual

sense instead of the numerical data. Regarding the PCB's code, it is obvious

that the code of a PCB can play a vital role to help fit the market data and de-

manufacturers gathered data (the photo that will be acquired from the board, is

recognized as de-manufacturers gathered data). The code addresses a specific

type of board among a wide range of coded PCBs with completely different

parts and information.

b) PCB’s Physical characteristics

As mentioned before, each one of the collected data can affect different aspects

of the de-manufacturing or recycling processes. In this case, physical

characteristics of a board such as dimensional measures and weight can play a

crucial role in designing an automated treatment line’s feeding mechanism and

also the design of an automated disassembly line’s fixtures. To go more into

detail, the board’s dimension is a primary data for each one of the processes

like cleaning, visioning, assembly or disassembly stations and also packaging.

As it can be understood, the dimension of the boards can help to design an

automatic treating line. Considering the need for the dimension of a board, it

should be mentioned that for the classification of the PCB which is under

process in image process step, it would be needed and helps to classify a board

as a phone board, motherboard, telecommunication board or ….

c) PCB’s Production related data

As regard as the importance of a product’s lifespan from the de-manufacturing

or recycling point of view, it would be interesting to know the quantity and

also the status of a product in its service life, in order to estimate its returning

time to the recycling phase and plan for the incoming load of EOL PCBs.

Considering the average life span of a product such as PCB, there is a wide

variety of different factors (such as working temperature, vibration, working

environment pollution, the number of components and the possibility of each

component’s failure in functionality) that each one can play a significant role

in defining the life span of a PCB, so it cannot be a precise number but based

on the number of returned PCBs for repair and the manufacturers estimation

of their life, more or less it can be estimated and considered as the PCB’s

lifespan estimation. The database can provide the Production date, Produced

Quantity, and Estimated lifespan as the helpful data in this concept.

P a g e | 104

d) Component’s type, quantity, code and functionality information

Regarding the market’s need for second-hand electronic components, it would

be obvious that the information related to the mounted components on the

boards will be important from the dismantler or recycler’s point of view in

order to be able to link the data from market’s need and their available

components in hand. Hence, general information such as the components’ type

(package, capacitator, relay…), package type (QFP, SOIC, DIP…) and their

functionalities make the planning process easier and feasible to systemize the

recycling process and cope with the market demand. Furthermore, as each

component is distinguished by a specific code, the availability of components’

codes can make it easier to exactly search for a specific component with a

unique code on a range of different boards with different codes. The

importance of providing a set of components’ codes becomes more logical

while thinking of a fully trained visioning system, which is responsible to take

the board’s photo and match it with a reference board’s photo to give a list of

its useful components and their available quantity.

e) Component’s needed data related to de-soldering and picking tasks

Data regarding the physical characteristics of each component is needed from

an operational point of view in assembly, disassembly, cleaning, packaging

and even recycling processes. As regard as the means and tools which are used

to assemble and disassemble or dismount the components from the board, the

physical and dimensional information will be essential in order to use the

proper heating tool and gripping method. Some of the dimensional

characteristics can be used to define the needed tool and equipment to heat or

grip the components but also, they would help to define the strength of a

component against heat and decide about the heating technology or

temperature, which is explained more in the following section that is talking

about soldering profile and the recommended heating profile for each

component. As regard as components connection type (SMD or PTH), the

related data help to choose among the present disassembly technologies and

decide on which method the dismantler should carry out to get a better result

in case of part’s intactness, quality and overall efficiency of the system.

Considering the assumption of returning and reusing the used electronic

components in the market, the components’ intactness and sensitive selective

disassembly gets a high level of importance, hence based on the parts’

connection type and the dimensional features, the recommended de-soldering

temperature can be presented, in order not to affect the functionality of the

P a g e | 105

components. Based on a manufacturer’s document (Freescale semiconductor

Ltd) regarding the temperature limit of a package, it is claimed that the

tolerable temperature for a component for maintaining its functionality, is a

function of its thickness and total volume (Table 3.18, 3.19), thus regarding

the importance of not damaging the PCB or its components in case of being

repaired and reused, the dimensional data about the components will help to

dedicate a specific soldering or de-soldering temperature, during assembly or

disassembly tasks respectively.

Package thickness Volume mm3 < 350 Volume mm3 ≥ 350

< 2.5mm 240+0/-5°C 225+0/5°C

≥ 2.5mm 225+0/-5°C 225+0/5°C

Table 3. 18 Package Peak Temperature for SnPb Solder Materials [37]

Package

thickness Volume mm3 < 350 Volume mm3 350-2000 Volume mm3 > 2000

< 1.6mm 260+0°C 260+0°C 260+0°C

1.6mm-2.5mm 260+0°C 250+0°C 245+0°C

≥ 2.5mm 250+0°C 245+0°C 245+0°C

Table 3. 19 Package Peak Temperature for Pb-free Solder Materials [37]

f) Component’s formative precious materials information

To evaluate the intrinsic value of a PCB (from recycling perspective), it would

be a good estimation to evaluate its value, based on its consisting precious

metals (Gold, Silver, Copper…) and their used quantity. Hence the present

content of each precious material in the components is presented and make it

possible to assess the intrinsic value of the board just based on the components’

material value. As a matter of fact, the value of each component (in case of

recycling) comes from its precious metals, hence for calculating this value and

use it as one of the needed data for cost analysis, the quantity of each used

element in the components will be so important.

g) Component’s Place code on the board and its coordination

Thinking about automatization of a boards’ components’ disassembly or

assembly line, the coordination of each component can be used by a computer-

aided heating and gripping system, in order to de-solder and dismount the

P a g e | 106

component from the board, and also solder new components on board (in case

of repairing the board). So, based on this hypothesis and for trying to pave the

ground and make it easier to implement an automated system to disassemble

the components from the boards, the place coordination of each component is

presented and in order to just make it easier to find a specific component by

just viewing the board, the place code can also be provided.

It is noteworthy to remark that, the collected data could be supportive in

different aspects of de-manufacturing procedure, e.g. the coordination of each

component would be used if an automated assembly or disassembly

application could be used, otherwise there would be no need to provide such a

database for the de-manufacturer or the number of each component would be

useful for comparing the output of image process (the number of each present

component on the PCB which is under process) with the number of

components which should be present on the same board type.

In the following parts of this work, we have tried to implement these introduced

data in an optimization program with an objective function of operating cost

minimization, thus to maximize the potential profit, out of EOL PCBs.

h) PCB and the components’ testing data

As far as the testing process is considered as a crucial recognition step to get

the status of a board or even its components, we must consider its needed

information in the PCB’s database. The data regarding the test of a PCB would

be clustered as highly technical information. The testing Voltage and Current

would be considered as the main data and for the components, the testing

Voltage, Current, Resistance, and capacitance are among the most important

and main ones.

3.3 Economic model

As it was claimed in the first part of this work, the main objective of defining such

a work methodology and also defining every needed data type for the waste PCB’s

treatment steps, was to benefit the PCB’s de-manufacturing procedure and help

to optimize it.

The economic model which is presented in this section is going to link the data,

which come from different parties like the manufacturer and the market and also

is achieved by the de-manufacturing different activities (PCB Inspection, Image

process, Testing …) on the PCBs. The economic model will be used as the basis

P a g e | 107

of the cost analysis step which takes the data from the manufacturer’s provided

database and also could be updated with the new sets of data coming from related

processing points, simultaneously as the PCB is under process. The economic

model will support the decision-making step in giving the most profitable path for

treating the PCBs.

In order to present the model, it would be better to break the concept of “Cost”

down into its sub-sections and afterward assign different types of expenses related

to different types of PCB’s treatment activities, to each one of the cost sub-

sections.

As far as the cost of PCB’s de-manufacturing is concerned, it is noteworthy to

mention that the cost is a sum of two types of costs which are either “Fixed” or

“Variable”.

Cost = Fixed costs + Variable costs

Regarding the fixed part of the costs, which contain the Investment, Rent,

Equipment’s indirect energy costs, and Administrative costs, they will not appear

in our model, but they will be used in order to define the break-even point of de-

manufacturing activity. Thus, it would define the payback time. The part of

variable costs would be the defining and decisive part in our economic model

since it comprises all the variable cost elements that depend on the PCB’s needed

treatment steps. In order to make it clearer that how the variable costs of various

processing steps would help the decision-making task to find the most profitable

PCB’s treating stream, it would be needed to break it down into different sub-

sections and analyze them one by one and define the variable costs of different

processing steps.

In the following part the variable costs will be presented more in detail:

Variable costs = Direct labor cost + Direct material cost + Direct

energy cost

“Direct labor cost” is defined as a variable cost in estimating the PCB’s overall

cost. Since the cost analysis step compares the revenue and the costs of treating

one unit of PCB, and also because the different PCBs have different streams of

treatment, hence it is not logical to divide the labor salary over the number of

treated PCBs. And it is needed to assign each PCB, its own related costs. In order

to define the labor cost of each PCB, the efficient time which is spent over it in

different steps should be calculated and be summed up. Hence as it is shown in

P a g e | 108

the formula, the overall direct labor cost would be the sum of the direct labor cost

of all the passed steps by the treated boards.

Total direct labor cost (DLC) of a PCB =∑ 𝐷𝐿𝐶(𝑖)𝐼

𝑖=1

DLC = Wage Rate (€/h) * Processing Time (h)

Operational step (i = the index of operational step)

The assumed number of operational steps (I = 17, Based on our work description)

Regarding the second part of the cost formula, the “Direct material cost” of the

PCB’s treatment should be calculated. It covers all the costs of the used materials

which are needed for a PCB’s treatment in each one of the passed operational

steps. As an example, in case of using cleaning substances in cleaning step or

using new components and solder material in the assembly step, their costs and

material expenses which directly is spent on a specific board should be taken into

account and be summed up as a total material cost of a specific PCB.

Total direct material cost (DMC) of a PCB = ∑ ∑ 𝐷𝑀𝐶(𝑖, 𝑗)𝐽𝑗=1

𝐼𝑖=1

DMC (j) = ∑ 𝑃𝑟𝑖𝑐𝑒(𝑗) ∗𝐽𝑗=1 𝑈𝑠𝑒𝑑 𝑞𝑢𝑎𝑛𝑡𝑖𝑡𝑦(𝑗)

DMC: Price of Used Material (€/Component or €/Unit of Used Material) *

Quantity of components or Units of used material

Operational step (i = the index of operational step)

The assumed number of operational steps (I = 17, Based on our work description)

Material type (j = the index of used material type)

The assumed number of used materials in each step (“J” depends on each step,

e.g. in assembly step, if we have 3 different types of components to assemble on

the board and also the soldering material, then “J” would be equal to 4)

Considering the third part of the costs, it refers to the “direct energy cost” which

exactly has been spent for a PCB by the different equipment.

Some costs can be referred to as mixed costs. An example could be electricity, the

electricity usage may increase with production but if nothing is produced a factory

still may require a certain amount of power just to maintain itself. Based on

literature “Overheads are the expenditure which cannot be conveniently traced to

or identified with any particular cost unit, unlike operating expenses such as raw

material and labor. Therefore, overheads cannot be immediately associated with

the products or services being offered, thus do not directly generate profits.” [38]

P a g e | 109

Hence it could be considered as an overhead cost which has no relation to the

production. In the following part, based on the mentioned work description and

also the explained steps of treating or processing the PCBs, the table of related

costs are provided.

Total direct energy cost (DEC) of a PCB = ∑ 𝐷𝐸𝐶(𝑖)𝐼𝑖=1

Operational step (i = the index of operational step)

Direct Energy Cost: Equipment energy consumption rate (kW) * Energy cost

(€/(kW.h)) * Processing time for each PCB (h)

The assumed number of operational steps (I = 17, Based on our work description)

The assumed number of processed PCBs (Y=2000, Based on our processed

hypothetic PCBs)

Hence, as a result, the overall variable treatment cost of a PCB could be

summarized in the following formula which is the sum of all the variable costs of

“direct labor cost”, “direct material cost”, and “direct energy cost”.

Variable costs

∑ (∑ 𝑫𝑳𝑪(𝒊, 𝒚)

𝑰

𝒊=𝟏

+ ∑ ∑ 𝑫𝑴𝑪(𝒊, 𝒋, 𝒚)𝑱

𝒋=𝟏

𝑰

𝒊=𝟏 + ∑ 𝑫𝑬𝑪(𝒊, 𝒚)

𝑰

𝒊=𝟏)

𝒀

𝒚=𝟏

Considering the potential achievable revenue from de-manufacturing the EOL

products, it would be possible to formulate it in the following way. Obviously, the

output of the de-manufacturing process depends on the various factors such as the

market demand and also the condition of the under-process product, thus the main

parts of the revenue model, are the coefficients of the decision-making steps and

the price of possible outputs (PCB, component, and material). (Figure 3.11),

(Table 3.20)

Revenue of treating a PCB

∑ (𝑷𝑶𝑩 ∗ 𝟐𝒏𝒅𝑪𝑨𝑹𝑩(𝒚) + 𝟐𝒏𝒅𝑩𝑻(𝒚) ∗ 𝑷𝑶𝑩 + 𝟐𝒏𝒅𝑪𝑨𝑹𝑪(𝒚)𝒀

𝒚=𝟏

∗ [𝑷𝑶𝑪 ∗ 𝑵𝑸𝑪𝒑𝒂𝒔𝒔𝒆𝒅𝑪(𝒚) + 𝑪𝑴𝑽 ∗ 𝑵𝑸𝑪𝒇𝒂𝒊𝒍𝒆𝒅𝑪(𝒚)]

+ [(𝟏 − 𝑩𝑰𝑹(𝒚)) + (𝟏 − 𝑩𝑫𝑺(𝒚)) + 𝟏𝒔𝒕𝑪𝑨𝑹𝑹(𝒚) + 𝟐𝒏𝒅𝑪𝑨𝑹𝑹

+ (𝟏 − 𝟐𝒏𝒅𝑩𝑻(𝒚))] ∗ [𝑵𝑪 ∗ 𝑪𝑴𝑽 + 𝑩𝑩𝑽 + 𝑯𝑪𝑴𝑪𝑽 ∗ 𝑵𝑯𝑪𝑴𝑪(𝒚)])

P a g e | 110

Figure 3. 11 Representation of the coefficients of the main assumed flows of the waste PCBs treatment

P a g e | 111

Damaged Board Not Damaged Board

BIR Board Inspection Result 0 1

Not Demanded Board Demanded Board

BDS Board Demand Status 0 1

The Quantity of Demanded Components

NPDC Number of Present

Demanded Components ≥ 0

Profitable to Reuse the Board

Profitable to Reuse the Components

Profitable to Recycle the Board

1st CARB

1st Cost Analysis Result for Board reuse

1 0 0

1st CARC

1st Cost Analysis Result for Component reuse

0 1 0

1st CARR

1st Cost Analysis Result for Recycling

0 0 1

Not Functioning Board Functioning Board

1st BT 1st Board Test 0 1

Profitable to Reuse the Board

Profitable to Reuse the Components

Profitable to Recycle the Board

2nd CARB

2nd Cost Analysis Result for Board reuse

1 0 0

2nd CARC

2nd Cost Analysis Result for Component reuse

0 1 0

2nd CARR

2nd Cost Analysis Result for Recycling

0 0 1

Not Functioning Board Functioning Board

P a g e | 112

2nd BT

2nd Board Test 0 1

POB Price of Board

MP Material Price

M Material Type

MCC Material Content of Component

Table 3. 20 Abbreviation clarification and the meaning of coefficient values

To clarify the abbreviations, and make the formula easier to understand:

BDS: Board demand status (1 for demanded PCB, else 0)

2nd CARB: Second cost analysis result for reusing the board (1 for profitable to

reuse the PCB, else 0)

1st and 2nd BT: First and second board test results (1 for passed, 0 for failed)

POB: Price of Board

2nd CARC: Second cost analysis result to reuse the components (1 for profitable

to reuse the Components, else 0)

NQCpassedC: number of quality control passed components

POC: Price of Component

M: Material type

MCC: Material Content of Components

MP: Material Price

Consequently, by having the cost and revenue of each PCB’s treatment, it is

possible to get its relevant profit and compare it to the other possible treatments

such as recycling or even its disposal.

P a g e | 113

4 Case study and application of the

model on the case

In this chapter, we are keen to specifically focus on Italtel case study and applying

the explained economic model to this case. Firstly, the company will be briefly

described and secondly, its target boards on which the work has been done will

be introduced. Furthermore, the related and needed gathered data, application of

the presented economic model on the case and finally, the results of such a model

application will be presented.

4.1 Italtel and Italtel’s boards

Italtel case study refers to an Italian PCB manufacturer which produces

telecommunication and network boards. Regarding the fast growth of technology,

while some of the boards, which have been used in the telecommunication

industry, are still functioning, others have reached their end of life stage and

should be replaced by the new ones. Hence, they need to be treated and returned

to the circular economy loop. As it was illustrated in the previous chapters, the

PCBs are considered as urban mineral resources of precious materials, and also

containing valuable functioning components which can be reused as backup

items. So, this stream of returning end of life PCBs can be considered as an

opportunity for the manufacturing company. As a result, the company has shown

an interest in understanding the potential value of these boards and also how could

ease the de-manufacturing task by providing the de-manufacturers or recycler’s

needed data. The first part was the subject of previous work which was done by

Paolo Citterio [2] and the second part, concerning “How to use the data?” is the

subject of this work.

As regard to the wide range of produced PCBs by the company, the work had

been started concentrating on the most produced ones which consist of fourteen

“User-Interface” and six “Power-Supply” PCBs. For the first part of the study, the

preliminary analysis will be executed on the Italtel’s PCBs sample and the needed

data which is illustrated in the third chapter will be gathered for the Italtel case.

In the following part, we would apply the proposed economic model on the case

and try to get the results considering the various possible scenarios.

P a g e | 114

Figure 4. 1 The sample of Italtel’s User-interface and Power-supply PCBs

4.2 Preliminary analysis and the review of the work

which had been done on the data gathering of the

PCBs

The starting point of the work, started with the data provided from the former

study which had been done considering the PCBs’ components, types, present

precious materials and their amounts. Considering the previous work, it has been

mentioned that the PCB’s image acquisition is the vital step and based on its

result, it could be decided what grade the PCB is or which components it contains.

The data which had been gathered about the PCBs had supported the detailed

material estimation model for estimating each specific component’s amount of

precious materials. Hence as the final result, each PCB’s value was estimated

based on its main precious metals. [2]

Regarding the PCBs’ possible useful data which could be used in recycling or de-

manufacturing procedure, different main aspects were evaluated. The main

needed data and information, are economic, environmental and operational ones.

Considering the previous study, we faced some lack of information which

P a g e | 115

considering operational or environmental issues could be seen as needed data. The

former study was mainly focused on the economic aspect and the PCB’s value

regarding its precious materials.

As the “Image Process” provides crucial data in the PCB’s treatment decision

making, it is noteworthy to explain more what has been done and what could be

achieved as an output from this workstation. The visioning system of PCBs which

is performed by Dott. Baiguera in STIIMA-CNR laboratory provides the

opportunity to capture the PCB’s image and based on the color differentiation,

understands the presence of a variety of components on the boards, especially ICs

(integrated circuits). In the following figure 4.2, the captured image from a PCB

is shown and the different packages are highlighted as the identified components.

In the image processing workstation in STIIMA-CNR the work is under

improvement and till now the level of PCB recognition is quite good at perceiving

the type of different packages and, since the system is provided with the data that

each package type consists of which materials and in what quantity, hence the

total material estimation would be feasible. In the following Figure 4.2, it is

visible that the different boards have been analyzed and their packages are

recognized. The occurrence of some errors regarding both the partition algorithm

and the classification one could happen, but generally, this doesn’t have a negative

impact on the final material estimation.

P a g e | 116

Figure 4. 2 PCB’s analyzed image by the software

The image acquisition system performance would be as follows. The boards are

analyzed using the visioning system which consists of a LED light to enlighten

the boards and to avoid shadows, a linear camera to take high-quality pictures,

which will be reconstructed in a single image as in figure 4.2, and a transportation

belt to transport the boards. Once the pictures have been taken, they are elaborated

by a specific software which is developed by the researcher to recognize the

different components on the board. In figure 4.2, the different recognized

components are colored and categorized, as the large QFP package in the center

and many SOICs and D2PKs. This system, together with a classification

algorithm, based on weight and dimensions, has to be considered as PCBs sorting

instruments. This system is ready to be transfer to a real industrial facility. [2]

As a matter of the fact, the general work methodology that has been presented in

the third chapter covers all the possible scenarios, which a returning PCB could

have. For the application of the board, we have focused on Italtel PCBs, which

based on company’s representative’s belief, there is no demand in the market for

PCBs, but its components, and hence it is comprehended that the possible

scenarios for the PCBs, would be reusing the components and recycling. Although

it is conceived that the PCBs of Italtel will not have the scenario of reusing the

boards, just to verify the program which is based on the work description and

economic model, their provided database contains those data which is needed for

other scenarios and it is assumed that some of the PCBs have market demand.

P a g e | 117

4.3 Data gathering and data structure

The need for a collected group of data and information about the four Italtel’s

boards which their data was provided by Italtel will be explained in this chapter

and also the reason of the presence of each kind of data in the database will be

presented. The studied PCBs are three User-Interface and a Power-Supply PCB

among the twenty PCBs which had been studied in the previous work done by

Paolo Citterio. [2]

In the case of Italtel’s EOL PCBs, there is a wide time gap, starting from their

production until their final disposal to the third-party collectors or recyclers. In

this case the de-manufacturer and recycler who want to treat the boards need a set

of information and data by which they could optimize their decision-making

procedure about purchasing decision and operational decisions.

The structure of the Italtel’s PCBs’ databases and their different sections

I should be noted that the Italtel Company provides all the needed data. The

establishing and organizing of a profitable database for the recycler could be an

interesting objective in our work. Therefore is important to create a reasonable

structure for the useful data, which will be the main aim of the project.

The PCB’s collected database contains the information, which can be beneficial

for a different aspect of de-manufacturing or particularly recycling processes,

such as being beneficial in economic, operational, financial and environmental

point of views. In this case, the sample form of the structure of the database and

the need for such an information base will be presented. As the sample case, the

information about the PCB code S7338-L6030-U2-A2 is presented.

PCB code: S7338-L6030-U2-A2

This specific PCB is recognized as one of the User-Interface PCBs with a

“Medium” level of grade (regarding the former study on the material value

estimation), as it is shown in Figure 4.4, the PCB is mostly containing just SOICs

and QFPs which are considered as the high-value components regarding the level

of gold content. More precisely speaking, regarding its interesting components

from the recycling point of view, there are 90 notable different components from

which, the number of different types of packages are 60 SOICs, 16 QFPs, 12

Tantalum Capacitor, 1 DIP and 1 DPAK.

P a g e | 118

Figure 4. 3 Italtel PCB S7338-L6030

Regarding the table of the database, and specifically considering the economic

point of view, firstly, the most valuable packages are selected as the most

important components for which the mentioned database should have been

completed. Hence as the first column, the name of the packages are listed.

Secondly, in the next column, we have the number of packages as a useful type

of data in order to multiply their unitary bearing costs or revenues. In the third

place, the code of each component will be presented, as it would be needed to

match the market request of components with the available components on EOL

PCBs. In the fourth column, also the functionality of the component is available

as the supportive data for matching the market demand with available

components, in case of reusing them after disassembly (regarding the reuse of the

intact components). As the fifth place, the packages’ formative precious materials

and their present amount are provided (regarding recycling choice). The amount

of precious metals is estimated based on statistical approaches. [2]

As regard as the operational point of view, in the next columns, we have foreseen

the need of the data related to the “Maximum Temperature Resistance”,

“Connection Type”, “Place Code”, and “Place Coordination” just to ease the

disassembly or assembly of the components. The place code and coordination of

the components (in case of automated assembly or disassembly) will get a high

level of importance. As the last data, the dimension of each component also is

P a g e | 119

needed for the gripping task which is another operational activity in assembly or

disassembly.

Component Quantity Code Functionality Materials(mg) Temperature

Resistance(°c)

Connection

Type

Place

Code

Coordination

Au Ag Cu X Y

SOIC 20 1 74abt240 Buffer/Line

driver 0.63 1.7 219.7

260 SMD U2 33528000 25400000

2 74ac273 Flip-flop U51 27432000 1684000

U62 21336000 9398000

QFP 68 1 hd64180 Microprocessor 1.81 2.68 1429

245 SMD U37 17907000 14224000

DIP 32 1 27c1001 UV EPROM 0.65 1.15 710.8

245 PTH U55 16002000 11176000

… … … … … … … … … … … …

Table 4. 1 Sample of the database regarding the PCB S7338-L6030

The detailed sub-sections of the database is explained precisely in the following:

a) PCB’s Code and picture

PCBs are identified by a wide range of codes which each one specifies a kind of

PCB with a specific functionality and range of characteristics. The Italtel PCBs

comprise a barcode and a written code which is shown in Figure 4.4 In the

database, the picture and code are inserted, to be useful as a source of picture

comparison in the “Image Process” step and also be used in “Demand Analysis”

respectively.

Figure 4. 4 Front and rear of an Italtel PCB code S7338-L6-30-A2-A1

b) PCB’s Physical characteristics

General information such as dimensions or the weight of the PCB which can be

helpful to give an overall vision and comprehension of the PCB to the recycler or

P a g e | 120

de-manufacturer and also can be used as a basic and essential data to design any

treating machine will be provided in the database. Regarding this specific model

of PCB, we have the following table of data.

Board Dimension (mm) Board weight (gr)

415 × 248 460

Table 4. 2 Dimension and weight of the Italtel PCB S7338-L6030

c) PCB’s Production related data

As regard as the production data of Italtel’s PCBs, the precise data except for the

total quantity of production, couldn’t be provided by the manufacturer, hence the

range of the years of production is presented in the database. As the lifespan of

the PCB is concerned, it was estimated based on the assumed lifespan that the

manufacturer had expected at first of production and also the duration of their

service time which has been exceeded from that of manufacturer’s expected value.

The production-related data is provided in the following table 4.3.

Board Production

Start date 1990

Stop date 2001

Produced Quantity 131770

Average lifetime (year) 30

Table 4. 3 The Production related data of the Italtel PCB S7338-L6030

d) Component’s type, quantity, code and functionality information

Regarding the presence of various types of components, to provide a complete

database which could be useful to give the detailed components’ information, the

type, code and functionality of the components are provided. As it was mentioned

before, the reason for their presence is giving more data in order to make the

searching for the market’s demanded components among the available

components on waste PCBs easier and possible. As an example, just to make it

clearer, if the demand is about a component with different “Component Code”,

then the other types of data like the type of component and its functionality, could

make it possible to match them, even the codes are not the same (Table 4.4). This

deviation of codes happens when the production of a specific component has been

P a g e | 121

stopped and the other manufacturer has produced the same component with the

same characteristics and functionality, with another code number.

Data Type Market Demand Availability in WPCB

Component Code AS6C1008 m5m51008

Component Package Type SOIC 32 SOIC 32

Component Functionality SRAM SRAM

Table 4. 4 Deviation of the component codes for the same component

A part of the structure of the explained data of the Italtel PCB S7338-L6030-U2-

A2 is shown in the following Table 4.5.

Component Total Quantity Type Quantity Component

Code Functionality

SOIC 16 12

1 rs232 Dual Transmitter / Receiver

3 74act109 Flip-flop

1 74act175 Flip-flop with master reset

1 ds34c87m Differential Line Driver

2 74LS123 Monostable multivibrator

1 ds34c86m Differential Line Receiver

2 74act161 Binary Counter

1 74act138 Decoder / Demultiplexer

SOIC 8 1 1 tl7705a Supply Voltage Supervisor

QFP 68 1 1 hd64180 Microprocessor

QFP 44 1 1 z85c3010 Communication Controller

DIP 32 1 1 27c1001 UV EPROM

Table 4. 5 Sample of Italtel PCB S7338-L6030 database, as the component introduction data

P a g e | 122

e) Component’s needed data related to de-soldering and picking tasks

A part of the data which is provided in the gathered database, considering the

operational tasks like de-soldering or picking of the components, for the Italtel

PCB S7338-L6030-U2-A2 is presented in Table 4.6. As it is visible, concerning

the de-soldering task we have provided dimensional and the connection type of

the components, since the dimensional data, can define the highest tolerable

temperature for de-soldering and the proper dimension of the component picking

gripper. Whereas the connection type can define the de-soldering technology that

should be used in disassembly activity. In the database, the component’s highest

tolerable temperature is written under one of two present clusters of de-soldering

(Reflow and Wave), depending on the component’s connection type. In this case

for the PTH components which are assembled on the board and can be de-soldered

by both wave and reflow de-soldering methods, the component’s highest tolerable

temperature is written under the both methods, while for the SMD components,

that the “Wave de-soldering” is not possible, the component’s highest tolerable

temperature is written just under the “Reflow de-soldering”.

Component Type

Quantity

Thickness

(mm)

Dimension (mm) Connection

Type

Package

Volume

(mm3)

Temperature resistance (C)

Length Width Diameter Reflow de-

soldering

Wave de-

soldering

SOIC 16

1 1.9 10.2 5.3 -

SMD

102.714 260

-

3

1.6 9.9 3.9 - 61.776

260

1

1

2

1

2

1

SOIC 8 1 1.65 4.9 3.9 - SMD 31.5315 260 -

QFP 68 1 3.9 24.23 24.23 - SMD 2289.66231 245 -

QFP 44 1 3.3 16.6 16.6 - SMD 909.348 245 -

DIP 32 1 4.2 41.9 13.2 - PTH 2322.9 245 245

Table 4. 6 The data regarding the De-soldering and Picking tasks

P a g e | 123

f) Component’s formative precious materials information

As regard as the recycling point of view, the source of revenue comes from the

PCB’s material content. The major part of the precious materials of a board comes

from its components, hence the data regarding the formative materials of the

Italtel’s PCBs had to be considered in the database as an important part of the

database to serve the recycler’s need for defining the most valuable components

which considering their recycling costs, their treatment would be beneficial. To

be more precise in explaining how to define the “Valuable components”, it should

be said that the variable costs of recycling for each recycler would be something

unique because the type of used recycling equipment could be from a wide range

of different types, hence different prices. As a result, the investment cost for

different recyclers would be different from the other ones. As a matter of fact, the

threshold of the components’ value can help to define the point in which the

component’s destructive disassembly and recycling costs would be equal to its

revenue coming from its formative materials. Hence the threshold of the

component’s value would define the break-even point in labeling a component as

a valuable component (“High Concentrated Material (HCM) Component”) or an

ordinary component. As it is explained before, the selective destructive

disassembly of the HCM components will be performed to increase the final grade

of the output powder of shredding machines. Hence the decision regarding the

choice of selecting the treatment path for a component as an HCM component’s

treatment path will also be made by getting the help from the following material

content database Table 4.7.

Component Type

Quantity

Materials of one component (mg)

Au Ag Cu Ni Ta Fe

QFP 68 1 1.81424 2.68 1428.8 - - -

QFP 28 1 0.74704 2.68 643.8165 - - -

DPAK 1 0.002 0.019 200.844 - - -

Tantalum

Capacitor 12 - 4.176 - 6.96 59.504 9.168

Table 4. 7 The formative material content of the components

g) Component’s Place code on the PCB and its coordination

As the final part of the PCB’s database, the place of the component is going to be

presented as a crucial data for finding the exact component on the board.

Considering the market demand which requests a precise component with a

specific type, code, and functionality, if the waste PCB is going to be treated in a

way which its component will be reused, so the exact component should be easily

P a g e | 124

found. Hence the place code and coordination would be needed to recognize the

right component, about which the database is giving the information. The sample

of Italtel’s PCB’s component information regarding the place code and

coordination is presented in Table 4.6.

Component Type

Quantity

Place Code Coordinate on Board

X Y

SOIC 16 1 U79 15240000 4826000

3 U19 18288000 20574000

U44 24384000 13716000

U63 18288000 9144000

1 U69 21336000 6858000

1 U32 33528000 16002000

2 U3 24384000 25146000

U24 30480000 18288000

1 U23 33528000 18288000

2 U18 21336000 20574000

U27 21336000 18288000

1 U59 30480000 9144000

SOIC 8 1 U10 27305000 22606000

QFP 68 1 U37 17907000 14224000

QFP 44 1 U71 14732000 7112000

DIP 32 1 U55 16002000 11176000

Table 4. 8 A part of the database regarding the components’ places

4.4 Economic model application

The application of the model on the case will demonstrate the difference between

the profit of EOL PCB’s information-provided de-manufacturing and the profit

which comes from the basic recycling treatment of EOL PCBs without the

exploitation of any type of data and information. Before going more into detail, it

is essential to highlight the assumed assumptions and the main hypothesis which

is made in this application.

Considering the Italtel’s case, the 4 different types of PCBs (3 User-interface and

1 Power supply) were studied and after gathering their related databases which

shows their actual characteristics and quantity of components, the different

assumptions about their conditions were assumed. Also, as a sample of our

P a g e | 125

economic model application, we will apply the model on one of the Italtel’s User-

interface boards (PCB S7338-L6030-U2-A2) with the actual and assumed data

about its hypothetic condition.

Consequently, the assumptions regarding the different steps, are presented

respectively. It is assumed that the EOL PCBs which are going to be processed in

the “Board Inspection” step, are 20 percent damaged, which it means they are

considered as partially broken, burnt or generally not feasible to be processed in

the following “Image Process”.

Similarly, regarding the boards which can pass the first steps of “Board

Inspection” and “Image Process”, it is assumed that among which 20 percent are

intact and functioning and the rest are considered as not functioning. Also, another

main assumption is related to the second board test step, that as it is placed after

the repairing steps (disassembly and assembly of defective components), it can be

assumed that with a 90 percent of probability, the outcoming boards are

functioning properly. With providing some random numbers, for each conditional

data (quantity of demanded components, the quantity of intact components, the

quantity of present high-value components, …) of each kind of PCB, we could

make a simulation of processing a wide range of the same type of board with

different conditions. As it is obvious, different conditions of each board, limit the

possible treatment activities to which is applicable on the board, e.g. if a PCB

does not contain any demanded component, practically those streams which are

designed for component reuse will be ignored. So, by applying the model to the

boards, as the result, it will give the most profitable and feasible stream for

treating the PCB. Considering the maximum profits of all the suggested streams

which fit the conditions of the PCBs, the mean of generated profits could be

extracted and be compared with the profit of the basic scenario of recycling the

board.

In order to clarify how the economic model, based on the various PCB’s

conditions, will provide the optimum work stream with the highest profit, we will

explain the condition, revenue, and cost evaluations more in detail.

The evaluation of a PCB’s treatment cost and the achievable revenue are

dependent on its condition after being used and transferred to the de-

manufacturing plant and also to its market situation. Thus, the model with the help

of some coefficients (for each decision-making step), would be able to assign a

group of possible applicable treatment scenarios for a specific board with its

unique condition. As regard as the de-manufacturing treatment assignment for

each board condition, the model starts with the “Board Inspection Result” (BIR)

coefficient which qualifies the undamaged PCB (BIR=1) or damaged one

P a g e | 126

(BIR=0) for the rest of the treatments. After the inspection step, the 10th stream

will be assigned to the damaged PCBs which will bear all the recycling related

costs, while delivering the revenue of selling the PCB’s materials content. Hence

the profit of 10th treatment stream would be calculated as the difference of

mentioned costs and revenue. Considering the undamaged PCBs, they will pursue

their identification process in the “Image Process” step and from which, an

important coefficient “Board Demand Status” (BDS) and two needed data,

“Number of Present Demanded Components” (NPDC) and “Number of High

Concentrated Material Components” (NHCMC) can be achieved. Where BDS for

demanded PCB would be equal to 1 and for not demanded PCB equal to 0,

similarly the NPDC would be equal to the quantity of demanded components

which can be disassembled and be sold to the market and the NHCMC would be

equal to the quantity of those components which, their disassembly cost is lower

than its material value.

By getting these data, it could be possible to separate the 9th treatment stream with

(BDS=0 and NPDC=0), which guide the PCBs toward the recycling treatment.

obviously, the 9th stream’s costs would be the summation of “Board Inspection”

and the “Image Processing” steps, and its revenue would be all the recycling

related costs. As it was explained before in the work methodology part in the 3rd

chapter and also is illustrated in figure 4.5, the PCBs which are considered as

demanded ones or contain demanded components, will be analyzed in the “Cost

Analysis” step and with a hypothetic assumption about their intactness and

functionality status (20% intact and 80% defective), their different treatment

streams’ profitability would be evaluated. So, based on which and also the

generated random values regarding the studied PCB’s conditions, the boards

which are even demanded or contain demanded (BDS = 1/0 and NPDC ≥ 0)

components, could be sent to the 8th stream for being recycled, as their evaluated

profitability in the other streams were lower than that of its recycling. After

analyzing the boards’ de-manufacturing profitability in the “Cost Analysis” step,

they will be tested and the needed coefficient regarding the “Board’s Intactness

Status” (BINT) will be generated, where the coefficient for the intact boards will

be equal to 0 and for the defective ones would be equal to 1.

As regard as our sample PCB, we will evaluate 10 same type PCBs with 10

different conditions to show all the possible treatment streams that a board can be

treated. For showing the 10th stream, it is assumed that the first assessed board is

damaged, thus (BIR = 0).

P a g e | 127

BIR

0

Table 4. 9 The conditional coefficient and assumed data for the 1st condition of PCB S7338-L6030

As regard as the second hypothetic condition, to show the 9th stream, the assumed

condition is that the board is not damaged (BIR = 1) and neither demanded (BDS

= 0) nor containing any demanded components (NPDC = 0) (as the demand is

dependent on the market, hence in this case it is assumed that the demand

condition is as explained). Also, as it has passed the image process step, it is

identified as (S7338-L6030-U2-A2 PCB) and based on the cost equation of

selective destructive disassembly of HCMC and the material value of each one of

present components on the board, it is realized that it does not have any HCMC

(NHCMC = 0).

BIR BDS NPDC NHCMC

1 0 0 0

Table 4. 10 The conditional coefficient and assumed data for the 2nd condition of PCB S7338-L6030

For the third hypothetic condition, to show the 8th stream, the assumed condition

of the board is that the board is not damaged (BIR = 1), and it is demanded and

contains demanded components (as the demand is dependent on the market, hence

in this case it is assumed that the board is demanded and containing demanded

components). based on the last explanation about the HCMC in this PCB, there is

no HCMC (NHCMC = 0).

BIR BDS NPDC NHCMC

1 1 ≥ 0 0

Table 4. 11 The conditional coefficient and assumed data for the 3nd condition of PCB S7338-L6030

To calculate the related costs, revenues, and profits of the 10th, 9th, and 8th streams,

the related equations are presented respectively in the following part. Also, the

cost related assumed factors of various steps are provided in table 4.12, the

board’s assumed revenue related factors in table 4.13, and the boards cost, and

revenue influential factors are presented in table 4.14.

P a g e | 128

LW (Labor Wage) (€/h) 8

BI (Board Inspection) (sec/board) 10

BGC (Board General Cleaning) (sec/board) 10

IP (Image Process) (sec/board) 30

SDDHCMC (Selective Destructive Disassembly of HCMC) (sec/component) 60

CR (Component Recycling) (sec/gr) 0.3

SD (Simultaneous Disassembly) (sec/board) 70

BBR (Bare Board Recycling) (sec/board) 0.3

Table 4. 12 Assumed Constant Cost Related Factors

MEANCMV (Mean Component Material Value) (€/component) 0.02109

BBV (Bare Board Value) (€/board) 0.000197

HCMCV (HCMC Value) (€/component) 0.1678124

Table 4. 13 Assumed Revenue Related Factors for PCB S7338-L6030

mass PCB (gr) 460

mass HCMC (gr) 1

mass BB (gr) 292

Table 4. 14 Cost and Revenue Influential Factors for PCB S7338-L6030

Cost of 8th stream

= LW ∗ (BI + BGC + IP + SDDHCMC ∗ NHCMC + CR ∗ massHCMC + SD

+ CR ∗ (massPCB − massHCMC − massBB) + BBR ∗ massBB)

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 8𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚

= (MEANCMV ∗ (massPCB − massBB − massHCMC) + massBB ∗ BBV+ HCMCV ∗ massHCMC)

Cost of 9th stream

= LW ∗ (BI + BGC + IP + SDDHCMC ∗ NHCMC + CR ∗ massHCMC + SD+ CR ∗ (massPCB − massHCMC − massBB) + BBR ∗ massBB)

P a g e | 129

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 9𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚= (MEANCMV ∗ (massPCB − massBB − massHCMC) + massBB ∗ BBV+ HCMCV ∗ massHCMC)

Cost of 10th stream = LW ∗ (BI + SD + CR ∗ (massPCB − massBB) + BBR ∗ massBB)

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 10𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚 = (MEANCMV ∗ (massPCB − massBB) + massBB ∗ BBV)

The overall result of treating such a board, is presented in the following table 4.15,

based on the various conditions which the boards have.

Treatment Stream Revenue

(€/board)

Cost

(€/board)

Profit

(€/board)

10th (1st PCB Condition) 1.8983 0.4844 1.4138

9th (2nd PCB Condition) 1.8983 0.5733 1.3249

8th (3rd PCB Condition) 1.8983 0.5733 1.3249

Table 4. 15 The 8th, 9th, and 10th treatment streams’ related revenue, cost, and profit comparison

P a g e | 130

Figure 4. 5 The PCB’s stream allocation, regarding the 8th, 9th, and 10th streams

P a g e | 131

Hence considering the de-manufacturing algorithm, after separating the streams

of intact and defective boards by the “Board Testing” step, each one of which will

be divided into various scenarios. The intact board’s various treatment scenarios

(1st, 2nd, and 3rd streams), will be evaluated in the cost analysis and their related

profit will be calculated as their related costs i.e. cost of board’s refurbishment,

components reuse, and recycling, and also their achievable revenue from the PCB,

component, and materials have been defined previously. The influential steps in

separating the streams such as decision-making or operational steps are illustrated

in figure 4.6. The fourth hypothetic condition of our sample PCB will be presented

to show the costs and revenues related to the 1st treatment stream (PCB reuse).

Thus, the 1st stream is designed for the PCBs which have the following condition:

BIR = 1, BDS = 1, (1st Board Test Result) BINT = 0.

BIR BDS NPDC NHCMC BINT

1 1 ≥ 0 0 0

Table 4. 16 The conditional coefficient and assumed data for the 4th condition of PCB S7338-L6030

The fifth hypothetic condition of the PCB is like the previous one, with a slight

change that, for the 1st treatment stream, the board must be demanded but for the

2nd stream (component reuse), as it is designed for component reuse, so its BDS

can be either 1 or 0, but the board should contain some demanded components

(NDC > 0). Regarding the related coefficients and data to the fifth condition, table

4.17, depicts that the PCB is not demanded but it has 18 demanded components

(based on the hypothetic output data of Image Process step, the number 18 out of

78 present packages on the board is presumed). BINT equal to 0 depicts the

intactness of the board which is the requirement of the boards which enter the

upper side of the algorithm (1st, 2nd, and 3rd streams).

BIR BDS NPDC NHCMC BINT

1 1 or 0 18 0 0

Table 4. 17 The conditional coefficient and assumed data for the 5th condition of PCB S7338-L6030

The 6th hypothetic condition of the PCB is again similar to the previous 4th and 5th

conditions, with this difference that for the 3rd stream (PCB recycling), the BDS

can be either 1 or 0, the NPDC can be either 0 or more, but it should be reminded

that there should be a market demand for either the board or its components to

reach these (1st, 2nd, and 3rd) streams. (Table 4.18)

P a g e | 132

BIR BDS NPDC NHCMC BINT

1 1 ≥ 0 0 0

1 0 > 0 0 0

Table 4. 18 The conditional coefficient and assumed data for the 6th condition of PCB S7338-L6030

The assumed cost-related factors for 1st, 2nd, and 3rd streams, plus those ones that

were provided for the 8th, 9th, and 10th streams (table 4.12, 4.13, and 4.14) are

provided in table 4.19, furthermore, the assumed revenue-related factors in table

4.20, and cost and revenue influential factors are presented in table 4.21. Although

the parameters of selective disassembly of intact component and component

quality control are functions of type and dimension of component, it is

approximated to be constant for all the components.

1st BT (First Board Test) (sec/board) 120

BFC (Board Final Cleaning) (sec/board) 180

BP (Board Packing) (sec/board) 90

IP (Image Process) (sec/board) 30

SDIC (Selective Disassembly of Intact Component) (sec/component) 200

CQC (Component Quality Control) (sec/component) 8

Table 4. 19 Assumed constant cost related factors

POB (Price of Board) (€/board) 4

POC (Price of Component) (€/component) 0.5

Table 4. 20 Assumed revenue relate factors for PCB S7338-L6030

NIC (Number of Intact Component) (component) 17

NQCpassedC (Number of Quality Control Passed Components) (component) 14

massQCfailedC (Mass of Quality Control Failed Components) (gr) 3

Table 4. 21 Cost and revenue influential factors for PCB S7338-L6030

P a g e | 133

𝐶𝑜𝑠𝑡 𝑜𝑓 1𝑠𝑡 𝑠𝑡𝑟𝑒𝑎𝑚 = 𝐿𝑊 ∗ (𝐵𝐼 + 𝐵𝐺𝐶 + 𝐼𝑃 + 1𝑠𝑡𝐵𝑇 + 𝐵𝐹𝐶 + 𝐵𝑃)

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 1𝑠𝑡 𝑠𝑡𝑟𝑒𝑎𝑚 = POB

𝐶𝑜𝑠𝑡 𝑜𝑓 2𝑛𝑑 𝑠𝑡𝑟𝑒𝑎𝑚

= 𝐿𝑊 ∗ (𝐵𝐼 + 𝐵𝐺𝐶 + 𝐼𝑃 + 1𝑠𝑡𝐵𝑇 + 𝑆𝐷𝐼𝐶 ∗ 𝑁𝐼𝐶 + 𝑆𝐷𝐷𝐻𝐶𝑀𝐶 ∗ 𝑁𝐻𝐶𝑀𝐶

+ 𝐶𝑅 ∗ 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 + 𝑆𝐷 + 𝐶𝑅 ∗ (𝑚𝑎𝑠𝑠𝑃𝐶𝐵 − 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 − 𝑚𝑎𝑠𝑠𝐵𝐵)

+ 𝐵𝐵𝑅 ∗ 𝑚𝑎𝑠𝑠𝐵𝐵 + 𝐶𝑄𝐶 ∗ 𝑁𝑃𝐷𝐶 + 𝐶𝑃 ∗ 𝑁𝑄𝐶𝑝𝑎𝑠𝑠𝑒𝑑𝐶 + 𝐶𝑅

∗ 𝑚𝑎𝑠𝑠𝑄𝐶𝑓𝑎𝑖𝑙𝑒𝑑𝐶)

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 2𝑛𝑑 𝑠𝑡𝑟𝑒𝑎𝑚

= POC ∗ NQCpassedC + MEANCMV ∗ massQCfailedC + MEANCMV

∗ (massPCB − massBB − massHCMC − NIC ∗ 1.5) + massBB ∗ BBV

+ massHCMC ∗ HCMCV

𝐶𝑜𝑠𝑡 𝑜𝑓 3𝑟𝑑 𝑠𝑡𝑟𝑒𝑎𝑚

= 𝐿𝑊 ∗ (𝐵𝐼 + 𝐵𝐺𝐶 + 𝐼𝑃 + 1𝑠𝑡𝐵𝑇 + 𝑆𝐷𝐷𝐻𝐶𝑀𝐶 ∗ 𝑁𝐻𝐶𝑀𝐶 + 𝐶𝑅

∗ 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 + 𝑆𝐷 + 𝐶𝑅 ∗ (𝑚𝑎𝑠𝑠𝑃𝐶𝐵 − 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 − 𝑚𝑎𝑠𝑠𝐵𝐵) + 𝐵𝐵𝑅

∗ 𝑚𝑎𝑠𝑠𝐵𝐵)

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 3𝑟𝑑 𝑠𝑡𝑟𝑒𝑎𝑚

= (MEANCMV ∗ (massPCB − massBB − massHCMC) + massBB ∗ BBV

+ HCMCV ∗ massHCMC)

The overall result of treating such a board, is presented in the following table 4.22,

based on the various conditions which the boards have.

Treatment Stream Revenue Cost Profit

1st (4th PCB Condition) 4 0.9778 3.0222

2nd (5th PCB Condition) 3.003 8.8731 -5.8701

3rd (6th PCB Condition) 1.8983 0.84 1.0583

Table 4. 22 The 1st, 2nd, and 3rd treatment streams’ related revenue, cost, and profit comparison

P a g e | 134

Figure 4. 6 The PCB’s Stream Allocation, Regarding the 1st, 2nd, and 3rd Streams

P a g e | 135

Considering the defective and non-functioning PCBs, the various scenarios

(streams 4, 5, 6, and 7), will be economically evaluated as well. As it is shown in

the following figure 4.7, the PCB which has passed the primary steps as physically

undamaged (BIR = 1), demanded (BDS = 1) or containing the demanded

components (NDC > 0); but it is recognized as defective board in the “Board Test”

step (BINT = 1), its treatment profitability considering its condition-related

scenarios will be evaluated and the most profitable stream would be selected as

its appropriate treating stream. The 4th and 5th streams will be divided from each

other, as the “Board test” step (after the board’s repair) indicates the “Board’s

Repair Result” (BRR) coefficient. It is assumed that 90 percent of the PCBs after

being repaired, will be qualified as functioning and proper to be packed and sold

in the market, thus their BRR would be equal to 1 and for the other 10 percent of

the boards, which could not pass the board test step, the BRR would be 0.

Considering our sample PCB, its 7th condition will guide the PCB to be treated in

the 4th treatment stream. Thus, to depict its condition the following table 4.23 is

presented. As it is shown, the PCB which is guided in the 4th treatment stream

(PCB repair and reuse), should be a demanded board from the market and its first

board test result, should be negative, in order to be repaired in the 4th stream. The

other coefficient which is added to the former coefficient tables, is the result of

the second board test step and it is expressed as the “Board Repair Result” (BRR).

A board which passes the second board test step (after being repaired) with a

satisfactory result (BRR = 1) can be packed in the following step and be sold.

Hence the influential values regarding its cost and revenue calculation will be

expressed in tables 4.24, 4.25, 4,26 plus the former data which were depicted in

tables 4.12, 4.13, 4.14, 4.19, 4.20, 4.21.

BIR BDS NPDC NHCMC BINT

1 1 ≥ 0 0 1

Table 4. 23 The conditional coefficient and assumed data for the 7th condition of PCB S7338-L6030

The next hypothetic condition of our sample board is the 8th condition which is

guiding the board to the 5th treatment stream (PCB repair and reuse). The 5th

stream is designed for those boards which are not damaged (BIR =1) and are

demanded (BDS = 1), but as they fail the first board test step, their BINT

coefficient is equal to 1, and also, after being repaired they will fail the second

board test (BRR = 0). (Table 4.24). Although this scenario is the worst case in

the model and is not the source of the profit for our board, the probability of

P a g e | 136

occurrence for the fifth stream would be too low and because of that it would like

to suppress this scenario.

BIR BDS NPDC NHCMC BINT BRR

1 1 ≥ 0 0 1 0

Table 4. 24 The conditional coefficient and assumed data for the 8th condition of PCB S7338-L6030

The next condition of our sample PCB would be the ninth condition and based on

the PCB’s specific condition which is shown in table 4.25, the board will be

guided to the 6th treatment stream (component reuse). Obviously as the aim of the

6th treatment stream is the use of components, hence, there should be some

demanded components on the PCB while the board could be either demanded or

not. The 2nd and 6th treatment streams are the same, but the 2nd stream is dedicated

to the boards which could pass the first board test step while the 6th stream is for

those boards which have failed the first test step.

BIR BDS NPDC NHCMC BINT

1 1 or 0 18 0 0

Table 4. 25 The conditional coefficient and assumed data for the 9th condition of PCB S7338-L6030

As the last hypothetic condition of the sample PCB, the 10th condition would

guide the PCB to the 7th treatment stream (PCB recycling). As it is provided in

table 4.26, the board’s condition coefficients show that the board has been

identified as the same as the previous 8th and 9th condition with a slight difference

that the board which is guided to the 7th should not necessarily have demanded

component or be demanded board, but as it was noted before for the 3rd stream

also, the board should either be demanded or have demanded components to be

reached the lower side of the algorithm (4th, 5th, 6th, and 7th streams).

BIR BDS NPDC NHCMC BINT

1 1 ≥ 0 0 1

1 0 > 0 0 1

Table 4. 26 The conditional coefficient and assumed data for the 10th condition of PCB S7338-L6030

P a g e | 137

As it is explained, it is obvious that all the 10 various streams which depend on

both the board’s condition and the board or components’ market demand status,

could be divided from each other by the defined coefficients. Consequently, by

applying the related coefficients for each stream’s cost and revenue drivers, the

profit of each stream could be calculated and be compared.

ANC (Assembly of New Component) (sec/component) 300

2nd Board Test (Second Board Test) (sec/board) 15

Table 4. 27 Assumed Constant Cost Related Factors

NDC (Number of Defective Component) (component) 1

massDDC (Mass of Disassembled Defective Components) (gr) 1

Table 4. 28 Cost and Revenue Influential Factors for PCB S7338-L6030

𝐶𝑜𝑠𝑡 𝑜𝑓 4𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚

= 𝐿𝑊 ∗ (𝐵𝐼 + 𝐵𝐺𝐶 + 𝐼𝑃 + 1𝑠𝑡𝐵𝑇 + 𝑆𝐷𝐷𝐶 + 𝐵𝑃𝐶 + (𝐴𝑁𝐶 + 2 ∗ 𝑃𝑂𝐶)

∗ 𝑁𝐷𝐶 + 2𝑛𝑑𝐵𝑇 + 𝐶𝑅 ∗ 𝑚𝑎𝑠𝑠𝐷𝐷𝐶 + 𝐵𝐹𝐶 + 𝐵𝑃)

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 4𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚 = massDDC ∗ MEANCMV + POB

𝐶𝑜𝑠𝑡 𝑜𝑓 5𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚

= 𝐿𝑊 ∗ (𝐵𝐼 + 𝐵𝐺𝐶 + 𝐼𝑃 + 1𝑠𝑡𝐵𝑇 + 𝑆𝐷𝐷𝐶 + 𝐵𝑃𝐶 + (𝐴𝑁𝐶 + 2 ∗ 𝑃𝑂𝐶)

∗ 𝑁𝐷𝐶 + 2𝑛𝑑𝐵𝑇 + 𝐶𝑅 ∗ 𝑚𝑎𝑠𝑠𝐷𝐷𝐶 + 𝑆𝐷𝐷𝐻𝐶𝑀𝐶 ∗ 𝑁𝐻𝐶𝑀𝐶 + 𝐶𝑅

∗ 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 + 𝑆𝐷 + 𝐶𝑅 ∗ (𝑚𝑎𝑠𝑠𝑃𝐶𝐵 − 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 − 𝑚𝑎𝑠𝑠𝐵𝐵) + 𝐵𝐵𝑅

∗ 𝑚𝑎𝑠𝑠𝐵𝐵)

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 5𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚

= (massDDC ∗ MEANCMV + MEANCMV ∗ (massPCB − massBB

− massHCMC) + massBB ∗ BBV + HCMCV ∗ massHCMC)

𝐶𝑜𝑠𝑡 𝑜𝑓 6𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚

= 𝐿𝑊 ∗ (𝐵𝐼 + 𝐵𝐺𝐶 + 𝐼𝑃 + 1𝑠𝑡𝐵𝑇 + 𝑆𝐷𝐼𝐶 ∗ 𝑁𝐼𝐶 + (𝑆𝐷𝐷𝐻𝐶𝑀𝐶 ∗ 𝑁𝐻𝐶𝑀𝐶

+ 𝐶𝑅 ∗ 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 + 𝑆𝐷 + 𝐶𝑅 ∗ (𝑚𝑎𝑠𝑠𝑃𝐶𝐵 − 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 − 𝑚𝑎𝑠𝑠𝐵𝐵)

+ 𝐵𝐵𝑅 ∗ 𝑚𝑎𝑠𝑠𝐵𝐵) + 𝐶𝑄𝐶 ∗ 𝑁𝑃𝐷𝐶 + 𝐶𝑃 ∗ 𝑁𝑄𝐶𝑝𝑎𝑠𝑠𝑒𝑑𝐶 + 𝐶𝑅

∗ 𝑚𝑎𝑠𝑠𝑄𝐶𝑓𝑎𝑖𝑙𝑒𝑑𝐶)

P a g e | 138

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 6𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚

= POC ∗ NQCpassedC + MEANCMV ∗ massQCfailedC + MEANCMV

∗ (massPCB − massBB − massHCMC − NIC ∗ 1.5) + massBB ∗ BBV

+ massHCMC ∗ HCMCV 𝐶𝑜𝑠𝑡 𝑜𝑓 7𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚

= 𝐿𝑊 ∗ (𝐵𝐼 + 𝐵𝐺𝐶 + 𝐼𝑃 + 1𝑠𝑡𝐵𝑇 + 𝑆𝐷𝐷𝐻𝐶𝑀𝐶 ∗ 𝑁𝐻𝐶𝑀𝐶 + 𝐶𝑅

∗ 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 + 𝑆𝐷 + 𝐶𝑅 ∗ (𝑚𝑎𝑠𝑠𝑃𝐶𝐵 − 𝑚𝑎𝑠𝑠𝐻𝐶𝑀𝐶 − 𝑚𝑎𝑠𝑠𝐵𝐵) + 𝐵𝐵𝑅

∗ 𝑚𝑎𝑠𝑠𝐵𝐵)

𝑅𝑒𝑣𝑒𝑛𝑢𝑒 𝑜𝑓 7𝑡ℎ 𝑠𝑡𝑟𝑒𝑎𝑚

= (MEANCMV ∗ (massPCB − massBB − massHCMC) + massBB ∗ BBV

+ HCMCV ∗ massHCMC)

The overall result of treating such a board, is presented in the following table 4.29,

based on the various conditions which the boards have.

Treatment Stream Revenue Cost Profit

4th (7th PCB Condition) 4.0211 2.1567 1.8644

5th (8th PCB Condition) 1.8983 1.3511 0.5472

6th (9th PCB Condition) 2.2928 4.7731 -2.4803

7th (10th PCB Condition) 1.8983 0.84 1.0583

Table 4. 29 The 4th, 5th, 6th, and 7th treatment streams’ related revenue, cost, and profit comparison

P a g e | 139

Figure 4. 7 The PCB’s Stream Allocation, Regarding the 4th, 5th, 6th, and 7th Streams

P a g e | 140

4.5 Results

After presenting the model application on the sample PCB (S7338-L6030) with

10 hypothetic conditions, now it is the time to show the results of evaluating a

wide range of possible conditions for all the 4 Italtel’s studied PCBs. In this part,

regarding each one of the PCB models, there would be 2 graphs showing the main

needed treatment streams and the relative achievable profit respectively.

Before starting the presentation of the results, it is appropriate to explain about

the numbers, conditions, and the main assumptions which were assumed for

evaluating the PCBs’ treatment cost, revenue, and profit.

The model application has been performed on 2000 different hypothetic

conditions of each one of the models of PCBs, considering the different

characteristics and properties of each model of the PCBs. Regarding the main

concept and aim of our work methodology which is the profitability assessment

of data implementation in de-manufacturing processes, we made some

assumptions which turns on the probability of reusing the EOL boards which are

applicable to Italtel and other kinds of PCBs like computer mother-board or

phones’ boards. Therefore the recycling and components reuse is realistic and

feasible consideration, while the Italtel PCB reuse is a strung assumption, but

necessary for evaluating the goodness of economic model. Just to show and

validate our methodology, it was needed to make such an assumption on the

provided case. The mentioned assumption about the Italtel’s PCBs is their market

demand, regarding either the boards or components. In this case from 4 assessed

PCBs, it is assumed that there is a market demand for 3 PCBs out of 4, and as far

as components’ demand status is concerned, the assumption is that it would be

possible to cluster all the present components on the PCBs to three different

clusters, the ICs, Capacitors and Relays. Among the components, the packages

are considered as the possible demanded components from the market, hence for

estimating the number of demanded components for each board, it is presumed

that the possible number of demanded components on a board could be a number

around one-fifth of the available number of its packages.

Data-provided de-manufacturing assessment of PCB S7201-X6001

The Power-supply PCB S7201-X6001 which is shown in figure 4.8, is considered

as one of the demanded boards among all the 4 studied boards. This board contains

18 notable components, among which 7 packages are available, thus 3

components are defined as demanded. The price of the board as a reusable product

P a g e | 141

and the mean price of the board’s reusable and demanded components are

considered as 4(€/board) and 0.55(€/component) respectively.

Figure 4. 8 Italtel’s Power-supply PCB S7201-X6001-U4-A1

Because of the data interference in de-manufacturing procedure, the possibility of

using the board or its components in the second-hand market would be possible.

Hence, after the assessment of the economic profitability of various treatment

procedures, the following graph (figure 4.9) is provided which shows the result

of treating the same board in different streams based on various conditions.

P a g e | 142

Figure 4. 9 Treatment streams allocation for (PCB S7201-X6001)

As it is visible, the 2000 boards are mainly spared in 5 de-manufacturing

treatment streams, which means that based on their conditions they were not able

to be processed in the other treatment streams. As the graph shows, among the 1st,

2nd, and 3rd streams, the decision-making system has guided the qualified PCBs

which had the needed requirements to the 1st stream. It means that among these

three streams, the first one which is the refurbishment stream for the reusable

boards has the most profit for this kind of board.

Also, regarding the 4th, 5th, 6th, and 7th streams, as the graph shows, the boards are

more sent to the 4th stream rather than the other streams. It can be understood that

the board’s price is high enough to generate profit from 1st and 4th streams which

are designed to refurbish the boards for being reused. The 4th and 5th streams are

divided from each other by a board test step, which controls the functionality of

the board after being repaired. Therefore, those boards, which have finished their

treatment in the fifth stream, were actually on the way of being reused but as their

second board test had a negative result, they had forced to be treated in the fifth

stream and be recycled. Furthermore, regarding the 2nd and 6th streams which are

the streams in which the reusable components will be prepared to be sold in the

market, the graph shows that there is no board which enters these streams.

Considering the fact that the most demanded components are ICs and the number

of this board’s ICs are too low, the reuse of component’s streams will be identified

as not profitable streams.

As regard as the 8th, 9th and 10th streams, it is seen that about one-third of the

boards are treated in the 8th stream to be recycled. The underlying reason for such

050

100150200250300350400450500550600650700

1 2 3 4 5 6 7 8 9 10

Bo

ard

Qu

anti

ty

Stream Number

Qantity

P a g e | 143

a phenomenon is that the treatment of a major part of the boards in the 8th stream,

was recognized as more profitable than sending the boards forward to be

processed in the following operational steps.

Considering the 9th stream, which is the dedicated stream for those boards which

are neither demanded nor contain the demanded components, the graph shows no

processed PCBs. The reason is that in this case our board is considered as

demanded and has the demanded components.

The 10th stream has accepted around 400 boards out of 2000 ones. It is showing

our main assumption that 20% of the boards were considered as damaged boards

and could not pass the “Board Inspection” step and had to be recycled.

Figure 4. 10 The treatment streams’ associated profit (PCB S7201-X6001)

The figure 4.10 is showing the associated profit with each one of the streams, and

as it is visible, some of the streams (1st and 4th) are generating positive profit while

the others are generating loss. As far as the profit of the 10th stream which is the

only step with the minimum costs (“Inspection” and “Recycling”), is negative, it

could be understood that even recycling of this PCB, would not be beneficial and

the only possible way to gain profit from this type of board, is to reuse it, as the

1st and 4th streams are showing the positive profit. The 5th stream is generating

such a big loss which is due to the high cost of this stream. As it was explained

before, since the boards are bearing the various costs of repairing steps and as a

matter of being failed in the test step, should be recycled; their foreseen revenue

of reusing the board will be changed with the revenue of their recycling, hence

the total outcome of this stream would be negative. The graph depicts that if the

-3,5-3

-2,5-2

-1,5-1

-0,50

0,51

1,52

2,53

3,5

1 2 3 4 5 6 7 8 9 10pro

fit

(eu

ro/b

oar

d)

Stream Number

Mean Profit

P a g e | 144

board would not be demanded from the market, the de-manufacturing procedure

for this particular board would be totally full of loss.

Data-provided de-manufacturing assessment of PCB S7338-L6030

The User-interface PCB S7338-L6030 which is shown in figure 4.11, is

considered as one of the demanded boards among all the 4 studied boards. This

board contains 90 notable components, among which 78 packages are available,

thus 18 components are defined as demanded. The price of the board as a reusable

product and the mean price of the board’s reusable and demanded components are

considered as 4(€/board) and 0.7(€/component) respectively.

Figure 4. 11 Italtel’s Power-supply PCB S7338-L6030-U2-A2

The following figure 4.12 is depicting the possible streams which our board based

on its various conditions will be processed in. As it is visible, the boards have

been processed in almost all the data needed streams (such as the 1st, 2nd, 4th, and

6th streams which are related to reuse of board or its components).

Regarding the 1st and 4th streams, it is seen that a notable number of boards have

been processed in the mentioned streams to be refurbished as a reusable board.

The reason is that the board’s market price has covered all the foreseen

refurbishment and repairing costs hence, the boards which were qualified and had

the proper condition have been guided to be processed in these streams.

P a g e | 145

By knowing the fact that the board is considered as demanded and contains the

demanded components, the graph shows that the boards which had passed the first

board test as intact boards, have been recognized as profitable to be reused as the

whole board in the 1st stream, while for the other boards which could not pass the

first test step with satisfactory result, it is realized that their profitability and

treatment stream depends on the number of intact components among its

demanded ones. Hence, we see that the boards which passed the first test step as

defective boards, are split to be processed in both the 4th and 6th streams where

the board or the components could be reused respectively.

Figure 4. 12 Treatment streams allocation for (PCB S7338-L6030)

Regarding the 7th stream which has processed almost 30% of the boards, it should

be said that after the first test step, they have been recognized as defective boards

with a big number of defective components. Hence, as their repair procedure

could be that much costly which could make its profitability negative, so, their

recycling procedure have been selected as the more profitable procedure among

the 4th, 5th, 6th and 7th streams.

The 8th and 9th streams are empty from any board, since regarding the 8th stream,

the board’s profit in reuse is more than its recycling so, it will be sent to the test

step; and regarding the 9th stream, we have assumed that the board is demanded

and as a result it does not accept any board.

The 10th stream is accepting the previously assumed percentage (20%) of

damaged boards to be recycled.

0

100

200

300

400

500

600

700

1 2 3 4 5 6 7 8 9 10

Bo

ard

Qu

anti

ty

Stream Number

Qantity

P a g e | 146

Figure 4. 13 The treatment streams’ associated profit (PCB S7338-L6030)

Data-provided de-manufacturing assessment of PCB S7510-L6023

The User-interface PCB S7510-L6023 which is shown in figure 4.14, is

considered as one of the not-demanded boards among all the 4 studied boards.

This board contains 111 notable components, among which 58 packages are

available, thus 10 components are defined as demanded. The mean price of the

board’s reusable and demanded components are considered as 0.5 (€/component).

Figure 4. 14 Italtel’s Power-supply PCB S7510-L6023-U1-A3

-0,50

0,51

1,52

2,53

3,54

1 2 3 4 5 6 7 8 9 10

Pro

fit

(eu

ro/b

oar

d)

Stream Number

Mean Profit

P a g e | 147

For making an example for the data-provided de-manufacturing assessment of a

not-demanded PCB, the mentioned PCB is considered as not-demanded but

containing some demanded components. Hence, the related graph is provided to

show the possible treatment streams in which the board can be processed.

Considering the graph, it is visible that the 1st and 4th streams which are related to

the refurbishment of the reusable PCBs, are empty of any board. The reason is

that the boards were considered as not-demanded, so, it is not logical to refurbish

a not-demanded board when there is no market need for it.

Comparing to the recycling streams (3rd and 7th), the 2nd and 6th streams which

are related to the reuse of the demanded components, have processed a fewer

number of the boards, which means that the decision regarding sending the

boards either to the 2nd and 6th or 3rd and 7th streams is dependent on the

intactness of the demanded boards. Hence the boards with a high level of

defective demanded components have been sent to the 3rd and 7th streams which

in this case have a better level of profit. Regarding the 8th, 9th and 10th streams,

the streams allocation is the same as previous mentioned PCBs.

Figure 4. 15 Treatment streams allocation for (PCB S7510-L6023)

As far as the mean profit of each board’s treatment is considered, the graph 4.16

is provided and is depicting the mean board’s profit in each stream. As it is visible,

all the processing streams, except recycling in the 10th stream, are generating a

negative profit, and around 1200 out of 2000 boards are recycled in the 7th stream.

The under lying reason for this phenomenon is that our model is working with a

set of hypothetic deterministic values which can be understood after passing the

related operational steps such as test or quality control check. In this case it is

visible that while we have the same boards with similar conditions (before sending

0

200

400

600

800

1000

1200

1400

1 2 3 4 5 6 7 8 9 10

Bo

ard

Qu

anti

ty

Sream Number

Qantity

P a g e | 148

towards 2nd and 3rd or 6th and 7th streams), the “2nd Cost Analysis” step decides to

send them towards the more profitable stream and as we have defined a hypothetic

percentage for the “Board’s Intactness”, “component’s ability to pass the

Component Quality Control step” etc. the decision making “2nd Cost Analysis”

step knows what will happen afterwards and consequently gives the optimum

decision among the possible streams. For instance, in this case board has

demanded components; after passing the test step as an intact board it will be

analyzed in the 2nd cost analysis step to decide about its most profitable stream

(2nd or 3rd). However, as we have provided the data about the component’s quality

control result for the cost analysis step, the decision would be made based on a

data that in the future we could get. As we had assumed it as deterministic and

known data, we could make this decision and send it for recycling, because the

probability for a board to be intact is 20% and the probability for a component to

pass the quality control step as an intact component is 90%, thus the probability

for the boards which are sent to the 2nd stream is (20% * 90%) and (80% * 10%)

will be the probability of the boards which are going to be sent to 3rd stream and

the probability for the boards which are sent to the 6th stream is (80% * 90%) and

(80% * 10%) will be the probability of the boards which are going to be sent to

7th stream.

Figure 4. 16 The treatment streams’ associated profit (PCB S7510-L6023)

Regarding the mean profit of treating this type of board in various treatment

streams, the graph 4.16 is provided and shows that the only profitable stream is

the 10th one. The achievable recycling profit from treating of this kind of PCB is

around one-fourth of the previous studied PCB, which is due to the lower level of

its material content. Considering the low value of its material content and bearing

-0,1

-0,05

0

0,05

0,1

0,15

0,2

0,25

0,3

1 2 3 4 5 6 7 8 9 10

Pro

fit

(eu

ro/b

oar

d)

Steram Number

Mean Profit

P a g e | 149

the operational costs of “Image Process” and “Board Test” steps in 3rd and 7th

steps, hence the profit of these streams is negative.

Data-provided de-manufacturing assessment of PCB S7510-L6047

The User-interface PCB S7510-L6047 which is shown in figure 4.17 is considered

as one of the demanded boards among all the 4 studied boards. This board contains

66 notable components, among which 39 packages are available, thus 12

components are defined as demanded. The price of the board as a reusable product

and the mean price of the board’s reusable and demanded components are

considered as 4.5(€/board) and 1 (€/component) respectively.

Figure 4. 17 Italtel’s Power-supply PCB S7510-L6047-U1-A3

The following figure 4.18 is depicting the possible streams which our board based

on its various conditions will be processed in. As it is visible, the boards have

been processed in almost all the data needed streams (such as the 1st, 2nd, 4th, and

6th streams which are related to reuse of board or its components).

Regarding the 1st and 4th streams, it is seen that a notable number of boards have

been processed in the mentioned streams to be refurbished as a reusable board.

The reason is that the board’s market price has covered all the foreseen

P a g e | 150

refurbishment and repairing costs hence, the boards which were qualified and had

the proper condition have been guided to be processed in these streams.

By knowing the fact that the board is considered as demanded and contains the

demanded components, the graph shows that the boards which had passed the first

board test as intact boards, have been recognized as profitable to be reused as the

whole board in the 1st stream, while for the other boards which could not pass the

first test step with satisfactory result, it is realized that their profitability and

treatment stream depends on the number of intact components among its

demanded ones. Hence, we see that the boards which passed the first test step as

defective boards, are split to be processed in both the 4th and 6th streams where

the board or the components could be reused respectively.

Figure 4. 18 Treatment streams allocation for (PCB S7510-L6047)

The difference between this board and the PCB S7338-L6030 is that because of

higher value of the board and its components, the treatment of the boards is more

profitable in either reusing the board or its components ea. 1st, 2nd, 4th, and 6th

streams. The reason that we do not see any treated board in the 3rd stream is that

the boards which have entered the upper part of the algorithm towards the 1st, 2nd,

and 3rd streams, had been qualified as intact previously, and as the high value of

board and components has just mentioned, the main profitable stream is the 1st

among the 3 streams. Regarding the lower part of the algorithm (4th, 5th, 6th, and

7th streams), again we observe a notable number of repairable boards which have

been sent to the 4th stream and for those boards that their repair had less profit

than its component reuse, the 6th stream has been selected.

0

100

200

300

400

500

1 2 3 4 5 6 7 8 9 10

Bo

ard

Qu

anti

ty

Steam Number

Qantity

P a g e | 151

Figure 4. 19 The treatment streams’ associated profit (PCB S7510-L6047)

As the graph 4.19 shows the mean profit of each treated board in different streams,

it is showing that the overall result of data-provided de-manufacturing of this

specific board has a high level of profit, since the only stream which has a slight

level of loss is the 5th stream and the other streams are generating profit.

Regarding the 2nd stream which has the highest mean profit, it should be noted

that the underlying reason is the price of its demanded components, hence, with

the same costs of achieving the components from the board, we are facing a higher

level of revenue, so a higher level of profit would be achieved.

The 8th and 9th streams are empty from any board, since regarding the 8th stream,

the board’s profit in reuse is more than its recycling so, it will be sent to the test

step; and regarding the 9th stream, we have assumed that the board is demanded

and as a result, it does not accept any board.

The 10th stream is accepting the previously assumed percentage (20%) of

damaged boards to be recycled.

Information provided de-manufacturing profit vs. recycling profit

comparison

Regarding the two “User-Interface” PCBs, (S7338-L6030, S7510-L6047), and

the “Power-Supply” board, as an assumption for using the model, we have

assumed them as the market demanded PCBs. As the table depicts their recycling

profit, it is visible that the two user-interface boards and the power-supply board

have a profit about 1.6 €/board, 1.2 €/board, and -0.5 €/board. Considering the

recycling of the power-supply board, not only it does not generate any profit, but

also generates loss. On the other hand, the profits earned from the designed model

for the mention boards are 1.75 €/board, 2 €/board and 0.2 €/board, respectively.

-1-0,5

00,5

11,5

22,5

33,5

44,5

5

1 2 3 4 5 6 7 8 9 10

Pro

fit

(eu

ro/b

oar

d)

Sream Number

Mean Profit

P a g e | 152

The table also shows that for the S7338-L6030 the profit of its data-provided de-

manufacturing will be increased by 10% while for the other boards (S7201-X6001

and S7510-L6047) the profit of their data-provided de-manufacturing will be

increased by 140% and 71% respectively, which will be a significant growth in

the profit earn by treating the end of life Italtel printed circuit boards.

As regard as the standard deviation, it should be noted that considering our output

profits, the standard deviation of each board’s related profit is a big number.

However, considering the higher level of profit, the impact of standard deviation

would be less.

Recycling

Profit

(€/PCB)

Model

Profit

(€/PCB)

Improvement

(€/PCB)

Improvement

Percentage

Standard

Deviation

S7338-L6030 +1.5916 +1.7505 +0.1589 +10% 0.847

S7510-L6023 +0.4372 -0.0199 -0.4571 -105% 0.1385

S7201-X6001 -0.5071 +0.2038 +0.7109 +140% 1.68

S7510-L6047 +1.1731 +2.008 +0.8349 +71% 1.376

Table 4. 30 Information provided de-manufacturing profit vs. recycling profit comparison

As it is visible in the following table 4.30, the mean unitary profit regarding the

basic recycling of each one of the PCB types is provided and compared with the

result of our model, which considers the PCB’s possible de-manufacturing

processes. Considering the first PCB (S7510-L6023) which is assumed as not

demanded from the market, the generated recycling profit is about 0.44 €/PCB

while its de-manufacturing related profit (Model based treatment) is negative and

about -0.02 €/PCB. As it can be seen, although we have exploited the data in the

simulation of de-manufacturing processes, the difference of the recycling and de-

manufacturing processes, is a negative value.

Therefore, the possible question, which can be arouse would be “What is the

benefit of having and exploiting the data and information in treating such a

board?” To answer this question, it is important to keep in mind that this

phenomenon has happened since the cost of the “Board Inspection”, “Image

Process”, and “Test” processes which are needed to evaluate the PCB’s condition,

will cause this loss of profit. To explain more, it would be better to make a

hypothetic example for processing this type of board. Considering the mentioned

PCB which contains some demanded components but does not have a market

demand for the whole PCB, it should be processed in primary work stations to be

P a g e | 153

qualified about the functionality and intactness of its demanded components,

which it means generating the cost. In this case, if the number of demanded

components were high and among them there would be just a few intact

components, hence we must bear a considerable amount of cost for just testing

them, while there would be no benefit of reusing them as intact components.

4.6 Sensitivity analysis

The profitability of applying the data-provided de-manufacturing model is

dependent on various factors and in this part of study we are focusing on the

impact and the level of various factors’ influence on the overall result.

To introduce the influential factors, it is better to cluster and group them together

to make the profit sensitivity graphs easier to be comprehended. For this reason,

we will have three clusters of different variables which are influencing the profit

level.

As regards as the first category of variables, we have clustered the data which are

related to the condition of the board together, and in the following graph (figure

4.20) the profit’s level of sensitivity to the change of each one of the factors is

depicted.

Profit’s sensitivity analysis regarding the condition-related

variables

The board’s condition-related factors are the a. Board’s physical intactness (which

is identified as “BIR” (Board’s Inspection Result)), b. Board’s functionality

intactness (which is recognized by “BINT” (Board’s Test Result)), c. Number of

present demanded components on the board (NPDC), d. Board’s functionality

status after being repaired in the process (which is recognized in test step after

being repaired by “BRR” (Board’s Repair Result)).

Considering the logic which is behind the incrementally changing a binary data

in the graph, and showing it in the percentage type is that, for instance for the

physically intactness of the boards (BIR) (which could be 1 for being not-

damaged and 0 for being damaged) we decided to show the percentage of not-

damaged boards from the whole quantity of processed boards. Hence as an

example, when in the horizontal axis, the “Board’s Inspection Result” (BIR) is

introduced as 20% (percentage of being damaged) it means that from all the

processed boards, 20% were damaged (which is the basic scenario) and while it

increases, the number of damaged boards will be increased.

P a g e | 154

Figure 4. 20 Profit’s sensitivity analysis regarding the PCB’s condition-related variables

Regarding the graph, it is visible that the profitability increases dramatically with

the change of “BINT” factor which is related to the board’s intactness percentage.

In the base scenario we have assumed the BINT equal to 20%, and the graph

shows that we can have 38% increase in the profit by improving the BINT to 40%,

which means that the more the boards are intact, the more profit would be

achieved. The next influential factor is the percentage of NPDC, which its 0%

means that none of the board’s components are demanded and consequently its

100% means all the present components on the board are considered as demanded.

The graph shows that with a 20% increase in the percentage of present demanded

components, the profit will be increased by 12%.

Considering the decreasing trends in the graph, it is noteworthy that their behavior

is almost the same, hence their level of influence on the profit are the same.

The trend of BIR which is the representative of the percentage of damaged boards,

shows that by having 40% of damaged boards, we will face a 13% decrease in our

profit than that of base scenario with 20% of damaged boards.

Similarly, the trend of BRR, which is the symbol of the percentage of defective

boards after repair, shows that the more defective boards that we have, the worse

our profit would be. Because the 0% of BRR means that all the repaired boards

0,3

0,5

0,7

0,9

1,1

1,3

1,5

1,7

1,9

2,1

0 % 2 0 % 4 0 % 6 0 % 8 0 % 1 0 0 %

PR

OFI

T(E

UR

O/B

OA

RD

)

Board Inspection Result Board Intactness

Demanded Components Board Repairing Result

P a g e | 155

after being repaired will pass the test step with satisfactory result, and 100% of

BRR means that all the boards after being repaired, will fail in the test step.

Profit’s sensitivity analysis regarding the market-related

variables

The board’s market-related factors are the a. Price of the board (POB), b. Price of

Component (POC), c. Component Material Value (the mean value of the board’s

components, except the high value ones “CMV”), d. Bare Board Value (BBV),

and e. HCMV (the mean value of the board’s high value components “HCMCV”).

Before addressing the sensitivity analysis graph, it is better to explain about the

values of base scenario’s variables. As we had mentioned them before in the

“model application on the case” part, we had used various values for POB, POC,

CMV, BBV, and HCMCV for each board. Thus, for the influential values in

calculating the profit of base scenario, we considered the mean value of each one.

For instance, the value of the POC, are different and ranging from 0.5 to 1

(€/component) depending on the boards type; to calculate the revenue of selling

the component in the base scenario, we have considered the mean value of

components. Similarly, the value of each variable in the base scenario is its mean

value.

Regarding the graph (figure 4.21), it is obvious that the more valuable are the used

components, materials, and totally the board, the more profitable de-

manufacturing it could have. As regard as the trend of basic recycling, it should

be noted that the only influential variables in its trend, are all the material-related

variables which are the source of recycling revenue (CMV, BBV, and HCMCV).

The PCB’s traditional treatment profit without data exploitation is depicted in the

graph and shows a smooth upward slope, since its only source of revenue is the

board’s material content.

Regarding the revenue-related factors, the graph shows that the POB is

influencing the profit the most. While the POB increases by 100%, the board’s

de-manufacturing profit increases by 152% as well, which is related to the

revenue of selling the board.

The POC is seen as the next important variable which significantly influences the

de-manufacturing profit. As it is visible, the trend of both the POB and POC until

50% increase are almost coincident, but as the change level increases, the graphs

difference becomes bigger. Considering the extent of profit increase that the POC

can be made, it is noteworthy to say that with a 100% increase in POC, the profit

will be increased by 120%.

P a g e | 156

Figure 4. 21 Profit’s sensitivity analysis regarding the market-related variables

As regard as the next variable, the CMV is influencing the level of profit. The

overall trend of POB, POC, and CMV are almost the same with a slight difference.

As the trend shows, by 100% increasing the CMV, it would be possible to increase

the profit by 99%. Regarding the other variables such as BBV and HCMCV, the

level of difference in the profit is negligible. All in all, it is comprehended that the

most important variables which influence the de-manufacturing profitability of a

PCB, are the price of the board, price of the component and the value of its

components, which in this case, it is obvious that the profit is more influenced by

the market price of reusable boards or components.

0

1

2

3

4

5

6

7

8

9

10

- 1 0 0 % - 5 0 % - 2 5 % 0 % 2 5 % 5 0 % 1 0 0 % 2 0 0 % 4 0 0 %

PR

OFI

T (E

UR

O/B

OA

RD

Price of Board

Price of Component

Component Material Value

Bare Board Value

High Concentrated Material Component Value

Traditional

P a g e | 157

Profit’s sensitivity analysis regarding the provided-data level

The most interesting graph would be the following graph which is showing the

achievable mean profit from treating a product (which in our case is a PCB), with

different types of data regarding the product’s related data. The product’s data

could be about its overall information (all the needed data for its recovery) or

about its subassemblies or parts and components (all the needed data for their

recovery) or at the lowest level, the data about its material content (the mass of

different material types which is used in it).

As it is visible, the graph shows the basic PCB recycling profit as a constant value

without any dependency on the presence of any data or information about the

PCB. The reason of its constant value is that, the recycling profit of a PCB is just

related on the material content of the PCB.

Figure 4. 22 Profit’s sensitivity analysis regarding the provided-data level

To simplify the graph’s drawing, we have assumed that the demand from the

market is highly dependent on the data which is provided for the de-manufacturer.

It means that we have assumed no value for the de-manufacturer’s output which

has acquired with no data (the shredded object with no data about its formative

material, has no value). For instance, when it is said that the data about

-0,8

-0,6

-0,4

-0,2

0

0,2

0,4

0,6

0,8

1

C B M CB MC MB MCB

Pro

fit

(eu

ro/b

oar

d)

Information Level

De-manufacturing Basic Recycling

C: Components` information is provided

B: Boards` information is provided

M: Materials` information is provided

CB: Components` and Boards` information are provided

MC: Materials` and Components` information are provided

MB: Materials` and Boards` information are provided

MCB: Materials`, Components` and Boards` information are provided

P a g e | 158

components is provided for the de-manufacturer, it means that the market will just

accept the de-manufacturers recovered components, just because they are

identified as known components. Consequently, when the data about the board,

components, and the materials are provided for the de-manufacturer, also there is

a market need for the board, component and the materials. In the figure 4.22 the

material, component and board are abbreviated by the words M, C and B

respectively. CB is the combination of the component and the board. The same

codification is valid for the rest possible combinations such as MC, MB and MCB.

Regarding the various types of provided-data for the PCB de-manufacturing, it is

visible that if the de-manufacturer will be provided by just the information about

the board’s components (the revenue comes just from the demanded components),

the profitability would be highly negative, since there is neither the revenue from

selling the boards nor the revenue from recycled materials. The graph shows that

regarding the 4 analyzed PCBs, the profitability will be positive.

While the data regarding just the material is provided for the de-manufacturer, the

only profit comes from recycling, which is our base scenario.

As regard as the other types of data which can be provided for the de-

manufacturer, it is shown that if the data about just the components and the board

or the material and the components will be provided for the de-manufacturer, the

profitability level will be lower that the base scenario.

As the graph is depicting the profit levels, it is shown that, if the data about just

the material and the board will be provided for the de-manufacturer, it would be

possible to make a higher level of margin respecting the recycling margin, and

finally the highest level of margin would be achieved when all the data regarding

the PCB is provided.

The mean profit of recycling a PCB, in our case is about 0.67 €/PCB, while based

on our model application, which had been provided by all the needed data, it could

be possible to get 0.98 €/PCB which shows an increase of 46% in margin.

P a g e | 159

5 Conclusion

The thesis main focus was how to exploit the manufacturer’s data by the de-

manufacturers to optimize the EOL products de-manufacturing process. As the

type of EOL product had a significant impact on the data exploitation procedure,

we investigated the EOL telecommunication PCB as the EOL products which

were the subject of recycling process. The aim of this thesis was to define the

optimum level of needed data and information for the de-manufacturer which

should be provided by the manufacturer and also how to take advantage of the

provided data to make a higher level of margin than that of basic recycling

scenario.

The base of data exploitation was the structure of needed data and a database

which its gathering was the first task of this work.

In current industrial practices and scientific literature, information gathering on

input products is often missing, resulting a loss of opportunity in treating the EOL

products and returning them to the loop of circular economy in an optimized way.

The needed database for each kind of product should be provided based on the

needs of its de-manufacturer; hence, the appropriate knowledge about the

products’ de-manufacturing process can be considered as crucial. After defining

the proper database and considering its size and the working environment of the

product, the decision about the most suitable method of information transfer,

would be possible to make.

The presented novel de-manufacturing model regarding the EOL PCB treatment

was defined based on the possible recoverable elements from PCBs. Regarding

the model, 3 different outcomes could be achieved. The achievable elements from

PCB de-manufacturing process were defined as the whole product, its

components, and its formative materials. The defined de-manufacturing model

which is equipped with various decision making and operational steps, works as

an optimization model which suggests the optimum procedure of treating the PCB

based on its condition and the market demand. In order to make the decision about

the most profitable processing streams for treating the PCBs, an economic model

was needed to be defined. Hence, based on the structure of the de-manufacturing

model, the economic model was defined.

After evaluating 4 PCBs of the Italtel’s EOL communication PCBs by the model,

it was seen that providing data would unlock the various streams of revenue for

the de-manufacturer. To be more precise about unlocking the profitable de-

manufacturing streams, it should be said that in the traditional de-manufacturing

P a g e | 160

procedure, the de-manufacturer is just getting profit from one of the profitable

streams which is recycling, while the other two margin generator streams (product

reuse and component reuse) are locked because of lack of information.

Based on the sensitivity analysis results, it is conceived that the de-manufacturing

model’s profit is highly sensitive to the market demand and the price of reusable

product or its parts and components. Hence, considering the obsolete products

that there is no market demand for them or inexpensive products that the purchase

of their used one does not make sense, applying the de-manufacturing model is

not suggested. But on the other hand, the more expensive the product or its

components are, the more the generated margin would be.

Consequently, the level of data which could be profit generator for the de-

manufacturer is related to the price of the product, and its reused-one market

demand. Also, as far as our economic model was based on PCB de-

manufacturing, all of its sensitivity analysis results do not apply on all types of

products. For instance, the acquisition process of a PCB’s component is complex

and costly while its price is low.

The thesis results showed that in case of providing the appropriate database for

the de-manufacturer in order to unlock all the three profitable streams of de-

manufacturing, there would be 46% improvement in margin, which is quiet

interesting. The potential achievable margin could be a motivation for the both

the manufacturer and de-manufacturer. There could be a business model, based

on which the generated margin is split among the data provider (manufacturer)

and the data user (de-manufacturer).

The successful business model which could be among the manufacturer and the

de-manufacturer, would be a win-win relation, from which both sides get profit.

On the one hand, the manufacturer which is obliged to take care of its products

lifecycle management and EOL product management, can provide the data for the

de-manufacturer for 2 reasons, 1. By providing the data for the de-manufacturer

it can motivate the de-manufacturer to treat its EOL products to earn more, 2. By

charging the de-manufacturer for providing the data for him, and on the other

hand the de-manufacturer who is interested in earning the more profit, is satisfied

by splitting the margin, since in any case its margin would be increased.

P a g e | 161

6 Bibliography

[1] Murray, Alan, Keith Skene, and Kathryn Haynes. "The circular

economy: an interdisciplinary exploration of the concept and

application in a global context." Journal of Business Ethics 140.3

(2017): 369-380.

[2] Citterio, "Reuse oriented circular economy strategy of waste PCBs

supported by knowledge based system" (2016)

[3] Winans, Kiara, Alissa Kendall, and Hui Deng. "The history and current

applications of the circular economy concept." Renewable and

Sustainable Energy Reviews 68 (2017): 825-833

[4] Stahel, Walter R. "The circular economy." Nature News531.7595

(2016): 435.

[5] Sarkis, Joseph, and Hanmin Zhu. "Information technology and systems

in China's circular economy: Implications for sustainability." Journal

of Systems and Infoarmation Technology 10.3 (2008): 202-217.

[6] Huo, X., Peng, L., Xu, X., Zheng, L., Qiu, B., Qi, Z., Zhang, B., Han,

D., Piao, Z., 2007.

[7] Leung, A., Cai, Z.W., Wong, M.H., 2006. Environmental

contamination from electronic waste recycling at Guiyu, southeast

China. J. Mater. Cycles Waste Manage. 8, 21-33.

[8] Guo, Q., Yue, X., Wang, M., Liu, Y., 2010. Pyrolysis of scrap printed

circuit board plastic particles in a fluidized bed. Powder Technol. 198,

422-428.

[9] Guo, Y., Huang, C., Zhang, H., Dong, Q., 2009b. Heavy metal

contamination from electronic waste recycling at Guiyu, Southeastern

China. J. Environ. Qual. 38, 1617-1626.

[10] Huang, Kui, Jie Guo, and Zhenming Xu. "Recycling of waste

printed circuit boards: A review of current technologies and treatment

status in China." Journal of hazardous materials164.2-3 (2009): 399-

408.

[11] Chung, Shan-Shan, and Chan Zhang. "An evaluation of legislative

measures on electrical and electronic waste in the People’s Republic of

China." Waste Management 31.12 (2011): 2638-2646.

[12] Tolio, Tullio, et al. "Design, management and control of

demanufacturing and remanufacturing systems." CIRP Annals66.2

(2017): 585-609.

P a g e | 162

[13] Geng, Yong, and Brent Doberstein. "Developing the circular

economy in China: Challenges and opportunities for achieving'leapfrog

development'." The International Journal of Sustainable Development

& World Ecology 15.3 (2008): 231-239.

[14] Li, Jianzhi, et al. "Printed circuit board recycling: a state-of-the-art

survey." IEEE transactions on electronics packaging

manufacturing 27.1 (2004): 33-42.

[15] CUCCHETTO, ELISA. "PCB design for recycling. Strategie per

il recupero di schede a circuito stampato mediante sistemi

automatizzati." (2012).

[16] https://www.amitroncorp.com/

[17] https://www.pcbcart.com/article/content/PCB-introduction.html/

[18] http://www.madehow.com/Volume-2/Printed-Circuit-

Board.html/

[19] Tullio, Tolio, et al. "Design, management and control of

demanufacturing and." CIRP Annals. Manufacturing

Technology (2017).

[20] ZHAO, Zi-wen, et al. "Development and Application of

Unsoldering Equipment for Printed Circuit Board Scraps [J]." Modular

Machine Tool & Automatic Manufacturing Technique 10 (2009): 025.)

[21] Layiding, Wa, et al. "Disassembling approaches and quality

assurance of electronic components mounted on PCBs." Electronics

and the Environment, 2005. Proceedings of the 2005 IEEE

International Symposium on. IEEE, 2005.

[22] Wang, Chen, et al. "An innovative approach to predict technology

evolution for the desoldering of printed circuit boards: A perspective

from China and America." Waste Management & Research 34.6

(2016): 491-501.

[23] Wang, Jianbo, Jie Guo, and Zhenming Xu. "An environmentally

friendly technology of disassembling electronic components from

waste printed circuit boards." Waste Management 53 (2016): 218-224.

[24] Okuyama, Shigeru. "Infrared heater." U.S. Patent No. 5,028,760.

2 Jul. 1991.

[25] Park, Seungsoo, et al. "Apparatus for electronic component

disassembly from printed circuit board assembly in e-waste."

International Journal of Mineral Processing 144 (2015): 11-15.

[26] Zhang, Xiaojiao, et al. "Effective dismantling of waste printed

circuit board assembly with methanesulfonic acid containing hydrogen

P a g e | 163

peroxide." Environmental Progress & Sustainable Energy 36.3 (2017):

873-878.)

[27] Beckett, P. M., et al. "Practical application of laser soldering in

electronics assembly." Industrial Electronics, 1997. ISIE'97.,

Proceedings of the IEEE International Symposium on. Vol. 1. IEEE,

1997.

[28] Huang, Haihong, et al. "Study on disassembling approaches of

electronic components mounted on PCBs." Advances in Life Cycle

Engineering for Sustainable Manufacturing Businesses. Springer,

London, 2007. 263-266.

[29] Colledani, Marcello, Giacomo Copani, and Tullio Tolio. "De-

manufacturing systems." Procedia CIRP 17 (2014): 14-19.

[30] Ijomah, W. L., and J. D. Chiodo. "Application of active

disassembly to extend profitable remanufacturing in small electrical

and electronic products." International Journal of Sustainable

Engineering 3.4 (2010): 246-257.

[31] Ghosh, B., et al. "Waste printed circuit boards recycling: an

extensive assessment of current status." Journal of Cleaner Production

94 (2015): 5-19.)

[32] Kaya, Muammer. "Recovery of metals from electronic waste by

physical and chemical recycling processes." International Journal of

Chemical, Molecular, Nuclear, Materials and Metallurgical

Engineering 10.2 (2016): 232-243.

[33] Williams, Julie Ann Stuart. "A review of research towards

computer integrated demanufacturing for materials

recovery." International Journal of Computer Integrated

Manufacturing20.8 (2007): 773-780.

[34] Parlikad, A.K., McFarlane, D., Harrison, M. and Thorne, A.

(2009) ‘The role of AIDC technologies in product recovery – an

information quality perspective’, Int. J. Product Lifecycle

Management, Vol. 4, Nos. 1/2/3, pp.129–145.

[35] https://elmatica.com/2017/02/20/price-development-printed-

circuit-boards-2007-2017/

[36] http://www.zeph.com/

[37] https://www.nxp.com/docs/en/application-note/AN3298.pdf/

[38] https://project-management.com/project-management-huts/

[39] White, Charles David, et al. "Product recovery with some byte: an

overview of management challenges and environmental consequences

P a g e | 164

in reverse manufacturing for the computer industry." Journal of

Cleaner Production 11.4 (2003): 445-458.