Citation
Does the multi-functionality of smartphones justify their environmental impacts?

Material Information

Title:
Does the multi-functionality of smartphones justify their environmental impacts?
Creator:
Martens, Joshua W. ( author )
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file (36 pages). : ;

Subjects

Subjects / Keywords:
Smartphones ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Advances in mobile communication technology have brought about a new distinction between a device with mobile communication capabilities, called a feature phone, and a device which has the computing power to perform a multitude of tasks, called a smartphone. Undoubtedly, one would expect the impacts of a smartphone to be greater than a feature phone, but if a smartphone user utilizes the capabilities of the device, then he/she may replace many individual items to perform the separate functions. The goal of this paper is to calculate the impacts of using a current smartphone, while applying credits for the emissions avoided by the multi-functionality of the device. The impacts and credits were calculated using an economic input output life cycle assessment (EIO-LCA) method. The results of this LCA showed that under "optimal replacement behavior", the avoided emissions were greater than the attributable impacts from smartphone use in 7 of 10 TRACI categories. In the other 3 categories, the avoided emissions offset at least 89% of the impacts. However, under current, survey-reported rates of replacement, the typical consumer does not replace enough of the dedicated devices to offset the impacts in any TRACI category.
Thesis:
Thesis (M.S.)--University of Colorado Denver. Civil engineering
Bibliography:
Includes bibliographic references.
System Details:
System requirements: Adobe Reader.
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Joshua W. Martens.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
900788696 ( OCLC )
ocn900788696

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

DOES THE MULTI FUNCTIONALITY OF SMA RTPHONES JUSTIFY THE IR ENVIRONMENTAL IMPACTS ? by JOSHUA W. MARTENS B.S., University of Michigan, 2007 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Civil Engineering Program 2014

PAGE 2

ii This thesis for the Master of Science degree by Joshua W. Martens has been approved for the Civil Engineering Department by Arunprakash Karunanithi Chair Indrani Pal Azadeh Bolhari May 18 th 2014

PAGE 3

iii Martens Joshua W. ( M.S., Civil Engineering ) Does the M ulti functionality of S martphones J ustify T heir E nvironmental I mpacts? Thesis directed by Associate Professor Arunprakash Karunanithi ABSTRACT Advances in mobile communication technology have brought about a new distinction between a device with mobile communication capabilities, called a feature phone, and a device which has the computing power to perform a multitude of tasks, called a smartphone. Undoubt edly, one would expect the impacts of a smartphone to be greater than a feature phone, but if a smartphone user utilizes the capabilities of the device, then he/she may replace many individual items to perform the separate functions. The goal of this paper is to calculate the impacts of using a current smartphone, while applying credits for the emissions avoided by the multi functionality of the device. The impacts and credits were calculated using an economic input output life cycle assessment (EIO LCA) m emissions were greater than the attributable impacts from smartphone use in 7 of 10 TRACI categories. In the other 3 categories, the avoided emissions offset at least 89% of the impacts. However, under current, survey reported rates of replacement, the typical consumer does not replace enough of the dedicated devices to offset the impacts in any TRACI category. The form and content of this abstract are approved I recommend its publication. Approved: Arunprakash Karunanithi

PAGE 4

iv DEDICATION I dedicate this work to my wife and my daughters. This work is only possible because of the support you have provided and sacrifices you have made for me. I thank you and I lo ve you all dearly.

PAGE 5

v TABLE OF CONTENTS CHAPTER I INTRODUCTION ................................ ................................ ................................ 1 Progress of Mobile Communications Technology ................................ ....... 1 Literature Review ............... .................................................. ....................... 2 II RESEARCH METHODOLOGY ................................ ................................ ......... 5 System Boundary ................................ ................................ ......................... 5 Stakeholders ................................ ................................ ................................ 6 System Function and Functional Unit ................................ .......................... 6 Life Cycle Inventory (LCI) ................................ ................................ ........... 8 Original LCA Approach: Hybrid LCA ................................ ............. 8 Revised LCA Approach: Pure EIO LCA ................................ .......... 9 Reference Flows ................................ ................................ ............................ 12 Defined Impacts and Credits ................................ ........................... 12 Price Estimation ................................ ................................ ................ 13 E mission Credit Scaling ................................ ................................ .... 14 III ANALYSIS ................................ ................................ ................................ ......... 16 Life Cycle Interpretation and Anlaysis (LCIA) Using TRACI .................... 16 Dominance of Attributable Impacts ................................ .............................. 17 Dominance of Avoided Emissions (Credits) ................................ ................ 19 ...................... 20 ............................ 20 IV DISCUSSION ................................ ................................ ................................ ..... 23 Extrapolation of Results ................................ ................................ ................ 23

PAGE 6

vi Study Limitations ................................ ................................ .......................... 23 Future Areas of Research ................................ ................................ .............. 25 REFERENCES ................................ ................................ ................................ ....... 26

PAGE 7

vii LIST OF TABLES Table 1 Literature Review of Smartphone LCAs ................................ ..................... 3 2 Sources of LCI data from original iteration of the study ............................ 9 3 Advantages and disadvantages of Process Based and EIO LCA ............. 11 4 Impacts (+) and credits ( ) considered for net impact of 2 year smartphone life span ................................ ................................ ..................... 13 5 Consumer survey results of dedicated device replacement by equivalent smartphone function ................................ ................................ ... 22

PAGE 8

viii LIST OF FIGURES Figure 1 a) Pew Research Center survey results of cellular phone ownership b) North Ameri can cellular subscriptions 2011 2013 actual; 2014 2019 predicted (Ericsson 2013) ................................ ................................ ............... 1 2 Basic process flow diagram with proper system boundary for each product/process considered ................................ ................................ ............ 5 3 Relative TRACI scores for smartphone production and 2 year use .......... 17 4 D ................................ ................................ ................................ 18 5 TRACI Scores for emission credits for 2 year smartphone use ................ 19 6 Range of TRACI scores for impacts and emission credits for 2 conditions ......................... 21 7 Range of TRACI scores for impacts and emission credits for 2 year use ................................ ............ 21 8 Net GHG (kg CO2 e) From All N. American Smartphones ..................... 23

PAGE 9

1 CHAPTER I INTRODUCTION Progress of Mobile Communica tions Technology Technological advances have transformed terminology we use to describe mobile cellular technology Feature phone is the term used today to describe a basic cellular phone that is capable of sending/receiving cal ls and text messages, and possibly access ing the internet, all without a touchscreen. The new distinction is meant to achieve the dichotomy in the cellular phone market between a smartphone which has the computing power and wireless networking capabilitie s to perfo rm a multitude of tasks via a touchscreen interface While the true necessity of these tasks is a debate for another paper, the fact remains that citizens of modern societies are relying on some device or devices to perform these tasks (Ericsson 2013) Additionally, the growth of overall smartphone ownership has risen drastically in recent years. Figure 1 shows two separate sources depicting the upward trends in smartphone ownership. a) b) Figure 1: a) Pew Research Center survey results of ce llular phone ownership b) North American cellular subscriptions 2011 2013 actual; 2014 2019 predicted (Ericsson 2013)

PAGE 10

2 While these results may hardly be surprising for any casual observer of consumer habits, it does confirm the growing market penetration o f smartphones and the transition of many consumers from owning a feature phone to a smartphone. With the increased proliferation of any technology, research naturally takes notice of impacts associated with its production and use. While studies on the ef fects of technology growth may produce conflicting and nebulous results on the economic and societal impacts (Mattila 2013), a more powerful and more complex technology is unequivocally expected to generate greater resource demands, higher energy demands, and mo re severe environmental impact. Multiplying these impacts by a population of smartphone users forecasted to exceed 5 billion by 2019 (Ericsson 2013) gives reason for study into the environmental consequences of such complex devices. Literature Revie w Several researchers have attempted to quantify the impacts of smartphone production and use. Table 1 summarizes the most thorough studies available and enumerates some of the Sustainable ICTs and Management Systems for Green Computing on data from a variety of studies of the individual components. They evaluated the quality of the data from each source and attributed a particular level of un certainty relative to the quality of the data. Each study included was performed in a process sum makes the tabulated result a lower bound rather than a complete estimate (Hu e t al. 2012). Climate change and energy usage were the only reported impact categories.

PAGE 11

3 Table 1: Literature Review of Smartphone LCAs Device Source of data LCA calculation performed by LCA Method GHG/Device Production ( kg CO2 e ) Energy/Device Productio n ( MJ ) "Generic smartphone after 2010" Physical Teardown by authors of PROSUITE study Mattila Judl Seppl (2013) Process Sum using ecoinvent 43.5 Aggregated studies from 2004 2008 Hu Hu Kaabouch (2012) Process Sum from varied sou rces 7.0 9.7 130 176 Nokia Lumia 1020 Corporate Report Nokia (2012) Process Sum according to ISO 14040 32.0 466.0 iPhone 5S Corporate Report Apple (2012) Process Sum according to ISO 14040 56.7 iPhone 5S Physical Teardown by iSuppli Martens (2014) EIO LCA 71.1 76.7 1044.1 1132.2 Samsung Galaxy S4 Physical Teardown by iSuppli Martens (2014) EIO LCA 91.4 1313.8 "Typical 2012 smartphone" Nomura Research institute Martens (2014) EIO LCA 49.4 61.9 719.9 898.6 The authors Mattila, Judl, and Seppl have published a series of papers (Mattila 2013, PROSUITE (2009 2013) is a pluridisciplinary EC supported "integrating project" whose goal is to develop a cohe rent, scientifically sound, and broadly accepted methodology for the sustainability assessment of current and future technologies over their life cycle, applicable to different stages of maturity Mattila, Judl, and Seppl worked on t he case study of multifunctional mobile devices. The authors performed a The mass of the components was then used in ecoinvent 2.2 to determine the

PAGE 12

4 environmental impa cts of production. While their results on the environmental impacts were of particular concern to this paper the authors describe the goal of PROSUITE as such: In the PROSUITE approach, a five pillar approach to sustainability was used (which) measured (sustainability) through the aspects of human health, social well being, natural environment, natural resource availability and economic prosperity their values fo recasted until 2020 and 2030 based on different growth possibilities. These results were ultimately converted into normalized endpoint scores based on European data from ecoinvent v2.2 for their main report for the PROSUITE project. Based on the he impacts on human health and natural environment would seem to be minor the functionality of the s 2012). This basket included a basic mobile phone, MP3 player, digital camera, GPS navigator, and laptop/notebook. In this paper, Judl reported results on a relative basis, finding different resu Original equipment manufacturers (OEMs), such as Nokia and Apple, also release their own internally produced environmental reports (Nokia 2012, Apple 2012). Both companies report greenhouse gas emis sions on a CO 2 e basis and Nokia also reports embodied energy. Details of methodology and individual pollutants are not described

PAGE 13

5 CHAPTER II RESEARCH METHODOLOGY System Boun dary features (e.g. camera improvements, near field communication, fingerprint scanner, ambient light sensors, gesture recognition) and software features (e.g. voice controls multimedia editing, account syncing and cloud computing applications). Undoubtedly, one would expect the impacts of a smartphone to be greater than those of a feature phone, but if a smartphone user utilizes the capabilities of the device, then he/she wi ll not have to purchase many individual items to perform the separate functions. Energy Emissions Figure 2 : Basic process flow diagram with proper system boundary for each product/process considered (Credit: Abel Chavez 2012) The goal of this pa per is to calculate the impacts of producing and using a current smartphone, while applying credits for the emissions avoided by the multi functionality to (LCA) The boundaries of this stu dy will include all upstream emissions from raw material extraction, production, transportation, as well as emissions associated with the energy and materials required during the use phase. Data on disposal and/or recycling of smartphones is very

PAGE 14

6 sparse a nd currently lacks a high degree of certainty. Impacts associated with the disposal and end of life processes (e.g. recycling, reuse, landfilling) for smartphones will b e considered in a qualitative section of the discussion but will not be included withi n the system boundary for the technical LCA Stakeholders By purchasing and using a smartphone, a consumer bears the responsibility for the impact of their high tech, hand held lifeline. The consumer is the main stakeholder in this study, as they should be concerned with their impact on the environment. Smartphone OEMs are additional stakeholders as the environmental perception of their devices may directly influence their profitability. System F unction and F unctional U nit This paper operates under th e premise that a smartphone may perform many unique functions (e.g. digital music player, calendar, alarm clock, digital photography) that would otherwise be performed by a variety of independent products. This LCA was performed under the assumption that the smartphone user would purchase and use many digital camera) to accomplish these functions if they did not have the option of using the smartphone. If the smartphone user does not purchase the individual goods, then the multifunctionality of the smartphone has effectively eliminated the emissions associated with the production and use of each good. For example, if 5 kg CO 2 e result from production of a personal calen dar, then 5 kg CO 2 e should be subtracted from the life cycle inventory (LCI) a smartphone for every year that the smartphone is used to replace the personal calendar.

PAGE 15

7 Other papers have compared the impacts of performing a particular task digitally vs. a t raditional method with very illuminating results (Andrae 2013; Kim 2012), but for this study, it is not necessary to calculate the impact of performing each task on the smartphone. When performed on a smartphone, each task may require three things: 1) a c ertain amount of electricity (which was delivered to the battery during charging) 2) online data storage and/or 3) wireless network activity. The emissions will be calculated from the electricity used during all of these unit processes, as well as an attr ibutable amount of emissions from data center and wireless network manufacturing. These emissions, along with the emissions from the production and distribution of the actual smartphone, will encompass the impact over the standard 2 year life time of curr ent smartphones. Thus, the activity of a smartphone over a 2 year period will serve as the functional unit of this paper. The 2 year time frame was chosen for several reasons. The iterative nature of smartphone innovation, combined with the current busin ess model of cellular providers in the United States results in a product cycle which has proven very lucrative to original equipment manufacturers (OEMs) and cellular providers. American wireless customers can purchase a smartphone at a subsidized price from most cellular providers upon signing a contract for two years of wireless voice and data service ( http://www.verizonwireless.com/b2c/device/smartphone ). These subsidies generally c over close to 2/3 of the standard cost of the newest models, and often cover the entire cost of models which have been released for more than one year. This subsidization model encourages many consumers to purchase a new device every two years. Some rese arch suggests that Americans replace their smartphone even more frequently (Wu,

PAGE 16

8 2012). Assuming the consumer will immediately replace a smartphone every 2 years also achieves a very conservative approach for the influence of these avoided emissions, as th Life Cycle Inventory (LCI) Original LCA Approach: Hybrid LCA Data collection poses inherent difficulties in conducting any LCA. Valid results require completeness, reliability, and consistency of the data. The complexity and variability of smartphone production, combined with approach of incorporating the impacts from many other products, generates even greater challenges in producing quality data for this study. Originally, the a uthor attempted an LCA of an iPhone 5 based on a compilation of data for the production and use of the iPhone and the products which it could replace. These data came from a variety of reliable sources, including individual process sum LCAs as well as eco nomic input output ( EIO ) data which was used to calculate LCI using the EIO LCA tool provided by Carnegie Mellon University (A comprehensive explanation of this method is contained in the following section). The baseline data for the emissions associate d with iPhone 5 production were based on data provided by Apple in its environmental report for the iPhone 5. The report provides the overall emissions in terms of CO 2 e without reporting the individual levels of each pollutant. EIO LCA data from oadcast and Wireless Communications provided the p roportion o f each pollutant relative to the CO 2 e produced. The individual emissions were then back calculated according to these proportions.

PAGE 17

9 Other processes were analyzed based on pro cess sum data when available. Otherwise, EIO LCA from appropriate sectors was used. Table 2 shows the breakdown of data sources. Table 2 : Sources of LCI data from original iteration of the study ( ) => emission credit; ( + ) => attributable emission s C ombining both process sum LCA and EIO LCA data produced a hybrid LCA. Hybrid LCAs have been shown to be more effective in capturing a larger system boundary than using either process sum LCA or EIO LCA methods independently (Williams 2004), however, great attention to detail is required to ensure that emissions are not double counted or that certain unit processes are not ignored. The lack of more robust data in the initial version of this study required assumptions that did not allow for the proper level of reliability necessary to confidently report the results. Revised LCA Approach: Pure EIO LCA In an attempt to synthesize the data in a more reliable manner, the entire LCA was revised to calculate all impacts based on EIO LCA methodology. Econom ic input output tables model the economic interactions between different sectors of an economy. These tables track the financial exchanges between the sectors and can then be used to approximate the proportional contribution of each sector to the economic activity of the

PAGE 18

10 sector of interest. The Bureau of Economic Analysis releases the EIO tables for the US economy every five years. 2002 data is the most recently released, as 2007 data was due out by the end of 2012, but is now expected to be released near the end of 2014. An LCA practitioner utilizes the work done by the researchers from the Green Design Institute (GDI) at Carnegie Mellon University to derive environmental impacts from a given quantity of economic activity. The GDI has connected the EIO data with emissions data from each sector and made this data openly available from their website: www.eiolca.net The total emissions of each sector are divided by the total economic activity of that sector, producin g an average emission factor per dollar of economic activity in that sector. These calculations are done for a variety of different emissions and impacts including greenhouse gas emissions, energy demands, TRACI midpoint scores, land use requirements, and common air pollutants. When the LCA practitioner chooses to examine the impacts of a product made by sector A, the EIO LCA tool utilizes EIO tables to determine the proportional economic activity generated in other sectors (e.g. X,Y,Z). The EIO LCA too l then generates reference flows for the environmental impacts created per $1M of activity by sector A by combining the impacts generated by sectors X, Y, and Z. Then by dividing those reference flows by the price of a single product, one can determine a n LCI for that product using this aggregated data. EIO LCA can produce powerful results that can enable a researcher to determine systems level conclusions with minimal resource investment. To conduct a comparable process sum LCA would take considerably more time and data collecting efforts without the possibility of capturing the system boundary as completely as an EIO LCA. Clearly, the

PAGE 19

11 aggregation of data across large sectors does limit the conclusions which can be drawn about the results of any one EI O LCA. Table 2 lists some inherent benefits and limitations of EIO LCA compared with a process sum model. Table 3: Advantages and disadvantages of Process Based and EIO LCA (From Hendrickson 2006 via http://www.eiolca.net/Method/Limitations.html ) Process Based LCA EIO LCA Advantages results are detailed, process specific results are economy wide, comprehensive assessments allows for specific product comparisons allows for systems level comparisons identifies areas for process improvements, weak point analysis uses publicly available, reproducible results provides for future product development assessments provides for future product development assessments provides information on ev ery commodity in the economy Disadvantages setting system boundary is subjective product assessments contain aggregate data tend to be time intensive and costly process assessments difficult difficult to apply to new process design must link monetary values with physical units use proprietary data imports treated as products created within economic boundaries cannot be replicated if confidential data are used availability of data for complete environmental effects uncertainty in data difficult to ap ply to an open economy (with substantial non comparable imports) uncertainty in data For the purposes of this study, EIO based calculations were a far more effective choice for several reasons: 1) This study was not intended to make claims about the imp acts of one specific smartphone model, but rather attempt to describe the impacts of the average/typical smartphone currently in use.

PAGE 20

12 2) Process specific data would not be very reliable because the particular production methods/processes of each smartphone su pplier are proprietary. 3) Accurately modeling the production of a complex device requires capturing a broad system boundary from complex supply chains, which would not be possible from a process based analysis. 4) Accurate analysis of the various replaced produ cts requires data from consistent system boundaries. EIO tables enable a high level of confidence in this consistency. 5) This study a lso did not intend to compare the impacts of specific replaced products. This mad e aggregated data a more appropriate choice. Reference F lows Defined Impacts and Credits All associated emissions in this study are converted into their impact over the 2 year functional unit. The emissions for which the smartphone is culpable (impacts) will be limited to the following processes: 1) P roduction of the smartphone 2) A ll electricity generated for charging the smartphone battery, running the data centers and wireless networks for 2 years 3) A portion of the production from data centers and wireless n etworks (GHG only) The emissions for which the smartphone will be credited with eliminating (credits) will be limited to the following processes:

PAGE 21

13 1) P roduction of the replaced products 2) Electricity/battery production associated with the use of the replaced pro ducts Table 4: Impacts (+) and credits ( ) considered for net impact of 2 year smartphone life span Production and Use Impacts Emission Credits from Replaced Products Smartphone Production Personal Calendar Smartphone battery charging (electricity) 8 MP camera Wireless Network Use (electricity) Landline Phone Cloud Storage Operation (electricity) Alarm clock Wrist watch Flashlight GPS unit Digital music player (iPod) Use of replaced products (electricity or battery) Price Estimation Robust EIO LCA results necessitate reliable price data. These data can be either producer or consumer prices. Consumer prices refer to the actual retail price paid by a consumer at the time of purchase excluding an additional sales tax. The producer price only encompasses the value of the materials and the costs of production for the product, but does not include the markup from production to retail value. The profit margin from that markup may undoubtedly vary over the range of products and firms within an ag gregated sector. Thus, producer price models are preferred over consumer price models when possible.

PAGE 22

14 The Nomura Research Institute generated a bill of materials (BoM) for the typical smartphone produced in 2012. A bill of materials describes the compon ents which make up a complex device. This list can be in terms of mass but the Nomura BoM listed a range of monetary values for each component. Applying these values to the producer price models for the appropriate sectors allowed for a very precise desc ription of the LCI range for the typical 2012 smartphone. For the prices of the replaced products, consumer price models were the only readily available data. In order to produce a conservative and reliable estimate of the avoided impacts, several measu res were taken to ensure that the prices of the replaced products were not overestimated. First, the products chosen as replacement products were assumed to have basic functionality (e.g. alarm clocks only with clock and AM/FM capabilities, MP3 player wit hout touchscreen/gaming capabilities). Second, the range of prices was taken from online retailers (excluding shipping and tax) by searching for the 40 most popular items. Third, the high and low price were eliminated before finding a mean price for each product. Last, the mean value was used for the upper bound estimate and the lower bound estimate was calculated as one standard deviation less than the mean. Emission Credit Scaling Clearly, every product does not have the same 2 year life cycle as the smartphone. Thus, span. For example, the life span of the alarm clock is estimated at 8 years and therefore, the 2 year activity of the smartphone only offsets 2/8 or 25% of the emissions associated with the production of the alarm clock. In the case of a 12 month personal calendar, the

PAGE 23

15 smartphone replaces 2 calendars during its 2 year life span, so the emission credits will be 200% of the emissions associated with th e production of a single calendar.

PAGE 24

16 CHAPTER III ANALYSIS Life Cycle Impact Assessment (LCIA) Using TRACI TRACI 2.0, the Tool for the Reduction and Assessment of Chemical and other Environmental Impacts 2.0 quantification of stresso rs that ha ve potential effects, including ozone depletion, global warming, acidification, eutrophication, tropo spheric ozone (smog) formation, human health criteria relate d effects, human health cancer, human health noncance r, ecotoxicity, and fossil fuel depletion effects. (Bare 2011) Thus, all emissions which contribute to particular environmental effect can be combined on an equivalent basis. Consider global warming as an example. CO 2 CH 4 NO x etc. all contribute to global warming. Using the relative global wa rming potentials for all of these pollutants their collective effect can be reported in terms of CO 2 equivalent or CO 2 e. The TRACI tool from EPA allows for similar equivalence calculations for the other potential effects mentioned above. Each of the pr oducts/processes was evaluated in order to determine the midpoint scores in ten TRACI categories. The scores are presented using the high impact score in each category as the 100 % reference level. This presentation was chosen for two reasons. First, usi ng relative scores allowed for the preservation of clarity and conciseness in the visual presentation of the data. The absolute scores vary by several orders of magnitude from one TRACI category to another. To display the scores with appropriate axes wou ld require ten separate graphics. Presenting all ten TRACI scores at relative levels allows the reader to clearly compare the trends in the data across categories. Second, following the example of the PROSUITE authors (Judl 2012), the relative score preven ts any overstatement about the absolute nature of the results. The study was meant to illustrate

PAGE 25

17 whether the attributable impacts could be offset by the avoided impacts, and thus, a relative comparison of the impacts and credits is a more appropriate resu lt to present. Dominance of Attributable Impacts Figure 3 displays the low and high values of the attributable impacts from the 2 year functional unit. The values represent the range of calculations based on high and low price estimations used in the EIO production clearly dominant every category. Production accounted for at least 62% of every score at the high estimate level (global warming and air acidification), and, in the case of water eutrophication, production generates 99.6% of the CFC 11e. Cloud computing showed to be the second most significant process, accounting for 17 21% of the high level impacts in five categories (global warming, air acidification, human health criteria (air), air eutrophica tion, and smog). Figure 3: Relative TRACI scores for smartphone production and 2 year use The dominance of the production aspect makes a closer evaluation of the individual components of the smartphone a worthwhile exercise. Figure 4 shows the contrib ution

PAGE 26

18 from each of the 24 components, with the most significant values highlighted. Several trends arise from close inspection. First, the components involved in the display and 35% of the impacts in any one category. The size, performance, and complexity of smartphone screens has grown since the first iPhone, and this trend could continue to drive up the overall impacts of a smartp hone. Second, the other elements that dominate each category are ones that will continue to be scaled up as ity as smartphones increase in processor speed, RAM, and graphics capabilities. LCA calculations using Nomura Research Institute range for bill of materials.

PAGE 27

19 Dominance of Avoided Emissions (Credits) demonstrates the dominance of the different products considered. These scores are also presented relative to the high impa ct level. Focusing first on the importance of individual contributors (the sum of the scores will be discussed in the following section), it is clear that production of the digital camera and the electricity for operation of the landline phone and alarm c lock contribute the most heavily to almost every category. In the the digital camera almost offset the entire high level impacts for the two year life cycle of the smartphone. Figure 5: TRACI Scores for emission credits for 2 year smartphone use

PAGE 28

20 Figure 6 shows the range of results for the impacts and emission credits from a single consumer who accomplishes the tasks of all the discussed products with a smartphone over the course of the 2 year functional unit. It is also assumed that the consumer would have purchased the products if the sm artphone could not perform the function (i.e. The smartphone consumer requires a digital camera and does not take pictures only because From these result s, one can see that in seven of the categories the lowest emission credit estimate would still offset the highest assumed impact of smartphone production and use. In the other three categories, ecotoxicity, human health cancer, and human health cancer, th e minimal emission credits were 95%, 89%, and 92% of the respective maximum impacts. Considering the conservative nature of the assumptions in this study, these results definitively prove that any individual smartphone user can completely offset their imp acts by fully utilizing the capabilities considered here. It would be nave and somewhat irresponsible to assume that every smartphone user exhibits optimal replacement behavior. According to consumer s urvey studies, the rate at which smartphone users are replacing dedicated devices with a single, multi functional device are growing but are hardly 100%. The percentages of respondents who replaced their devices with a smartphone function are shown in tab le 5. Applying these rates of replacement to the optimal behavior results produces the range of emission credits for the

PAGE 29

21 Figure 6: Range of TRACI scores for impacts and emission credits for 2 year use Fi gure 7: Range of TRACI scores for impacts and emission credits for 2 year use

PAGE 30

22 can be seen in figure 7. I t can be seen that the typical smartphone consumer does not replace enough of the dedicated devices to offset the impacts. As smartphones grow more and more incorporated into the everyday life of consumers, it can be expected that the rates of replacement will increase. Table 5: Consumer survey results of dedicated device replacement by equivalent smartphone function Device Survey Results ( /+ 5% ) Survey Source Digital Camera 43.0% Oracle MP3/iPod 34.0% Oracle GPS Unit 24.0% Oracle Landline Phone 35.8% CDC Calendar Flashlight Wristwatch Alarm Clock 34.2% Average of other devices

PAGE 31

23 CHAPTER IV DISCUSSION Extrapolation of Results While it has been clearly stated that the results of this study are informative on a relative basis, let us consider some possibilities for the global warming impact of all smartphone across North Amer ica. Starting with data as it stood in 2013, figure 8 depicts the net CO 2 e emissions from the number of smartphones projected to be in use by 2019 (Ericsson 2013). The chart displays the yearly trend in greenhouse gas emissions assuming different levels of annual product replacement increases. The results show that in order to achieve a net negative impact, the number of consumers replacing their dedicated devices would have to increase by 7 % annually until 2019. Figure 8: Net GHG (kg CO2 e) Fro m All N. American Smartphones ( Based on Smartphone Growth Projections (Ericsson Mobility Report 2013) and Growing Levels of Product Replacement ) Study Limitations Some of the limitations of this study have been discussed throughout the paper. The method of EIO LCA invites an unavoidable level of uncertainty. In particular, using data from 2002 EIO sectors required inflation scaling over a 10 12 year period. While the

PAGE 32

24 inflation rate was reliably based on consumer price indices, it is difficult to confiden tly state whether the prices of good followed these indices closely. The interactions between sectors, upon which the EIO LCA results are based, may have also changed significantly since 2002. Smartphones were just an emerging, niche product in 2002. Th e complexity of the devices has evolved greatly in just the past few years which may have conflicting consequences. On one hand, the energy and resource demands may be greater than the data from 2002 would suggest. On the other hand, maturation of supply chains and advancement in manufacturing techniques may have increased the efficiency of production. This paper did not attempt to capture those effects. (EoL) were not co nsidered. Studies from OEMs have shown very limited impacts to global warming from device recycling (Apple 2012, Nokia 2012), but these studies may be missing much more significant aspects of EoL activities. While the statistics on electronic recycling r ates are difficult to judge, EPA estimates only 25% of ICT devices recycled, with the majority of the rest ending up in landfills (EPA 2007). These low rates are due in part to consumer ignorance about where/how to recycle electronics and concerns about w hat actually happens with the devices (Hu 2012). Those concerns are legitimate as research suggests 50 80% of recycled electronics are shipped to developing countries for disassembly (Tong 2004). Poor, untrained people in these countries take apart these devices in order to sell useful parts or even chemically separate the precious metals without any of the proper safety materials. The ultimate fate of toxic chemicals released from EoL treatment, whether disassembled for material recovery, landfilling, o r incineration, can result in severe human health and ecotoxicity impacts (Ogilvie 2004).

PAGE 33

25 Future Areas of Research With 2007 EIO data due for imminent release, the results of this paper could be easily updated to reflect activity of more modern ICT produ ction. A comparison of more recently released individual devices could also produce illuminating results that OEMs could use for beneficial marketing. Companies like Google and Apple are aiming to project a sustainable, environmentally friendly image wit h commitments to powering their data centers with renewable energy (Greenpeace 2014). The optimal behavior results of this study could be used to further that image if applied to specific Android or Apple smartphones and including the reduced impacts of u sing the device powered by a OEMs or cellular service providers might also be interested in the true rates of replacement. One of these companies could benefit from the results of a survey which is specific to consumer habits regarding how smartphone users are utilizing the multifunctional nature of their devices. The survey results included in this study do not confirm whether consumers are actually had intentions to purchase the devices which had vey could determine what other products and services have consumers substituted with an equivalent smartphone function. The potential for these devices is practically limitless and there is definitely room for more research to better define the key issues and answers.

PAGE 34

26 REFERENCES Andrae, A. S. G. (2013). Comparative Micro Life Cycle Assessment of Physical and Virtual Desktops in a Cloud Computing Network with Consequential, Efficiency, and Rebound Considerations. Journal of Green Engineering 3 (2), 193 218. Apple Environmental Reports; http://www.apple.com/environment/reports/ [Accessed 12 April, 2013] Bare, J. (2011). TRACI 2.0: the tool for the reduction and assessment of chemical and other envi ronmental impacts 2.0. Clean Technologies and Environmental Policy 13 (5), 687 696. Barron, K., Hazen, J. ( 2011 ) Opportunity Calling: The Future of Mobile Communications Take Two. Oracle Communications. Bureau of Labor and Statistics; http://www.bls.gov/data/inflation_calculator.htm [Accessed 1 May, 2013] Carnegie Mellon University Green Design Institute. (2013) Economic Input Output Life Cycle Assessment (EIO LCA) US 2002 (428) model [Inte rnet], Available from: http://www.eiolca.net/ [Accessed 1 May, 2013] Center for Disease Control and Prevention. ( 2013 ) Behavioral Risk Factor Surveillance System. Comparability of Data BRFSS 2012. Cook, G., Dowdall, T., Pomerantz, D., Wang, Y. ( 2014 ) Clicking Clean: How Companies are Creating the Green Internet. http://www.greenpeace.org/usa/Global/usa/planet3/PDFs/clickingclean.pd f Gilstrap, D. ( 2013 ) Ericsson Mobility Report: On the Pulse of the Networked Society. http://www.ericsson.com/res/docs/2013/ericsson mobility report novembe r 2013.pdf Hankey, R. (2014). Electric Power Monthly with Data for January 2014 Hendrickson, C. T., Lave, L. B., Matthews, H. S. (2006). Environmental Life Cycle Assessment of Goods and Services: An Input Output Approach. Hu, W. C., Hu, W. C., & Kaabo uch, N. (2012). Sustainable ICTs and Management Systems for Green Computing. CH 11: 252 283. Kim, J., & Rohmer, S. (2012) Environmental Life Cycle Assessment on Paper and Electronic Billing & Payment System.

PAGE 35

27 Jakhanwal, V., Leung, V., Lam, W. ( 2013 ) http://technology.ihs.com/430692/samsung galaxy s4 carries 236 bill of materials ihs isuppli virtual teardown reveals Judl, J., Mattila, T., Seppala, J., Koskela, S ., & Kautto, P. (2012 ). Challenges in LCA comparisons of multifunctional electronic devices. In Electronics Goes Green 2012+(EGG), 2012 (pp. 1 5). IEEE. Maga, D., Hiebel, M., & Knermann, C. (2013). Comparison of two ICT s olutions: desktop PC versus thin client computing. The International Journal of Life Cycle Assessment 18 (4), 861 871. Malmodin, J., Moberg, ., Lundn, D., Finnveden, G., & Lvehagen, N. (2010). Greenhouse gas emissions and operational electricity use in the ICT and entertainment & media sectors. Journal of Industrial Ecology 14 (5), 770 790. Mansfield, I. ( 2013 ) Apple's iPhone 5s Carries $199 BOM and Manufacturing Cost. http://www.cel lular news.com/story/Handsets/62182.php Mattila, T.,Judl, J., Seppl, J. ( 2013 ) Final Deliverable W6, D6.3: Case Study: Information technology (Multifunctional mobile devices) Final sustainability assessment. http://prosuite.org/c/document_library/get_file?uuid=9102d135 b389 4bed 9a07 a456e0db2214&groupId=12772 Mattila, T., Judl, J., & Seppl, J. (2014). Carbon Footprint of Mobile Devices: Open Questions in Carbon Footprinting of Emerging Mobile ICT Technologies. In Assessment of Carbon Footprint in Different Industrial Sectors, Volume 1 (pp. 151 166). Springer Singapore. Moberg, ., Johansson, M., Finnveden, G., & Jonsson, A. (2010) Printed and tablet e paper newspaper from an environmental perspective A screening life cycle assessment. Environmental impact assessment review 30 (3), 177 191. Nokia. Eco Profile Nokia Lumia 1020. ( 2013 ) http://download.fds ncom.nokia.com/supportFiles/eco_declaration/files/eco_declaration_phones/Lumi a_1020_Eco_profile.pdf Ogilvie, S.M. (2004). WEEE and Hazardous Waste A report produced for DEFRA. Queen's Printer and Controller of HMSO. March 2004. opower.com http://blog.opow er.com/2012/09/how much does it cost to charge an iphone 5 a thought provokingly modest 0 41year/#Methodology Rassweiler, A. ( 2010 ) IHS Technology. Press Release. https://technology.ihs.com/388837/isuppli estimates new ipod nano bill of materials at 4373 Shah,R., Ho, S., Chaurasia, S., Jefferey, S. ( 2012 ) Smartphone Guide. Nomura Research Institute.

PAGE 36

28 http://images.businessweek.com/bloomberg/pdfs/nomura_smartphone_poster_201 2.pdf Smith, A. ( 2013 ) Smartphone Ownership 2013 Update. Pew Research Center. http://www.pewinternet.org/2013/06/05/smartphone ownership 2013/ Tong, X. and J. Wang. ( 2004 ) Transnational flows of e waste and spatial patterns of recycling in China. Eurasian Geography and Economics 45 (8): 589 602. U.S. Environmental Protection Agency. (2007). MANAGEMENT OF ELECTRONIC WASTE IN THE UNITED STATES: APPROACH TWO Draft Final Report Available on: http://www.epa.gov/osw/conserve/mat erials/ecycling/docs/app 2.pdf Williams, E. (2004). Energy intensity of computer manufacturing: Hybrid assessment combining process and economic input output methods. Environmental science & technology 38 (22), 6166 6174.