Evaporation of Mercury from CCFLs during Recycling of LCD

Television Sets
Kristofer Elo1, Erik Sundin1
1
Division of Manufacturing Engineering, Department of Management and Engineering,
Linköping University, Linköping, Sweden

Abstract
The element mercury is one of the most hazardous substances known. Still, it is common in the air, water,
soil and products we use in our daily life. LCD TVs is one of these products. To prevent the mercury in
the LCD TVs from polluting the environment, the LCD TVs are recycled. This is done through automatic
shredding or manual disassembly where the mercury can spread in the work environment, the process
equipment or to the recycled material. This is due to broken CCFLs in the LCD TVs which contain the
mercury. The aim of this paper is to investigate, through a literature review and an empirical study, the
amount of mercury released into the work environment due to broken CCFLs from LCD TVs. In the
literature review there were found the amount of mercury other researchers has found in CCFLs from
LCD TVs, and also where the mercury was found. In the empirical study, the amount of mercury in a
LCD due to broken CCFLs were measured and validates the results from other researcher and states that
the mercury is difficult to predict.

Keywords:
Mercury, recycling, disassembly, liquid crystal display, television set, cold cathode fluorescent lamps.

1 INTRODUCTION some precautions not are taken. The automatic recycling

Mercury (the element) is one of the most hazardous
substances in the world [1]. It is naturally present in air,
water and soil [2]. It is estimated that the natural mercury
emissions to the atmosphere are between 800 and 6000
tonnes annually [3-6]. Additionally, 2320 tonnes of
mercury are emissions caused by man [6]. Mercury is
absorbed by all living organisms and becomes more
concentrated in organisms higher in the food chain,
especially in aquatic animals. Humans can be exposed to
mercury by eating, drinking water and breathing air. The
general concentration of mercury vapours in the air is
expected to be between 2-4 ng/m3 in remote areas and
about 10 ng/m3 in urban areas [7-8].
Mercury is also present in many products. Examples of
such products are thermometers, switches, pressure
guards, flow meters, valves, rectifiers, thermostats, cold
cathode fluorescent lamps (CCFLs), paint, dental
amalgam and skin care products [4, 9-11]. CCFLs are used
in several lighting applications, e.g. liquid crystal display
(LCD) television sets (TVs), scanners, LCD monitors, and
digital cameras [11].
The common recycling of LCD TV containing CCFLs in
Europe is done through either automatic processes
(shredding) or manual disassembly [12-15]. During such
processes and operations the environment and the work
environment are at risk of being exposed to mercury if
processes contain technologies for removal of mercury
vapours [12-14, 16]. If the mercury vapours amalgamate
to the recycled materials or the recycling equipment itself,
then the mercury degrades the material and pollutes the
surrounding environment [14]. In the case of manual
disassembly, personal have the risk of mercury exposure
due to the possibility of treating LCD TVS with broken
CCFLs inside. The CCFLs could break, for example,
during transportation, preparation before disassembly and
during disassembly.
1.1 Aim
The aim of this paper is to investigate the amount of
mercury vapours released from broken CCFLs from LCD
TVs.
1.2 Methodology
During the work presented in this paper a literature review
and an empirical study were performed. This was done to
investigate the amount of mercury released into the
environment and the work environment due to broken
CCFLs. Following the review, a study was performed to
complement, verify or contradict the literature.
2 MERCURY VAPOURS – A REVIEW
If a human breathes in mercury vapours, about 80% of the
vapours will be absorbed by the body via the lungs [4, 7,

Proceedings of EcoDesign 2013 International Symposium

17-18]. Exposures of concentrations between 15 g/m3
and 30 g/m3 during long periods of time have shown
negative effects on the human health [8]. Exposure of
mercury vapours with a concentration between 1 mg/m3
and 3 mg/m3 during a few hours may cause pulmonary
irritation and destruction of the tissues in the lungs, and
occasionally disorders in the central nervous system.
Exposure during longer periods of time may cause shaking
and mental disturbances. These effects are observed
during long-term exposure to mercury vapours above the
concentration 0.1mg/m3 [5].
According to the World Health Organisation, the annual
level for mercury vapour exposure is 1 g/m3 before a risk
of affecting the health occurs [8]. Table 1 contains the
different work environmental levels in five different
countries of mercury allowed, over a period of eight hours,
five days a week.
Table 1: Work environment levels for mervury (Hg)
vapours from country to country
Country Organic Hg Non-organic Hg Ref.
Sweden 0.01 mg/m3 0.03mg/m3 [19]
Taiwan 0.01mg/m3 - [20]
Germany 0.02mg/m3 0.02mg/m3 [21]
UK 0.02mg/m3 0.02mg/m3 [22]
Japan 0.025mg/m3 - [23]
2.1 Mercury in CCFLs
The New Jersey Department of Environmental Protection
states in a report published in 2004 that between 17% and
40% of the mercury is vaporised if CCFLs are broken.
This is over a period of several hours, and a third of the
mercury is released over the first eight hours. [24] In the
literature, the amount of mercury in CCFLs varies
between 0.1 mg/lamp and 10 mg/lamp [11, 14, 17, 25-26].
According to two manufactures of CCFLs used in LCD
TVs the amount of mercury in the lamps is less than 3
mg/lamp [27-28]. The number of CCFLs in LCD TVs
differs in the literature between 2 and 28 lamps, depending
on the size and design of the TVs [13-14, 25, 29]. The
designs of LCD TVs are either a design with several
CCFLs directly illuminating the LCD screen from the
inside [25, 30], or a design were the light from two CCFLs
is guided to the screen via a light guide [31].
According to Dang et al., the highest concentration of
mercury in a fluorescent lamp is located in light emitting
powder [32]. Raposo et al. state that there are differences
in the concentration of mercury vapours from the emitting
powder depending on the age of the CCFLs [33]; the
newer CCFLs have a higher concentration, while the older
CCFLS have a lower one. The authors also state that this
is due to the mercury in the powder from the older CCFLs
being absorbed by surrounding components. McDonnel
and Williams state that the highest concentration of
mercury is located in the electrodes in CCFLs, and not the
light emitting powder. The authors also state that due to
the disagreement in the subject the entire CCFL should be
considered during end-of-life treatment. [25]
Another investigation of the amount of mercury vapour
released during end-of-life treatment of CCFLs in LCD
monitors and LCD TVs came to the conclusion that
mercury vapours have a highly volatile behaviour. During
a test, 76% of the mercury was accounted for during the
end-of-life shredding process. The statement of mercury
losses is based on the assumption that each CCFL contains
12 mg of mercury. During the same investigation, the
exposure of mercury vapours for personal performing
disassembly of LCD monitors was measured. The highest
concentrations were measured in the container for broken
CCFLs. The other measurement points were the
workplace, and in the direct surroundings of intentionally
broken CCFLs, but did not show concentrations above the
recommended level in the work environment. It should be
added that the workplace was fitted with a suitable
ventilation system. [17]
Thaler et al. found mercury in the glass of fluorescent
lamps which had been used up to about 25,000 hours. The
amount of mercury bound in the glass depends of the
weight of the fluorescent powder and the length of time
the lamps have been used. Lower weight and a longer
period of time results in more mercury bound to the glass
in the lamps. [34]
3 EMPIRICAL STUDY
The study contained three parameters which were changed
during the study while measuring the mercury vapour
concentration. The parameters in the study were:
 Number of broken CCFLs
 Size of the container
 Length of time the broken CCFLs were allowed to
evaporate mercury into the containers
Only one parameter was changed between the
measurements to make it possible to understand the
parameters effect on the mercury concentration. However,
this approach makes it difficult to see any synergy effects
between the parameters.
3.1 Equipment used
To measure the mercury vapours a gold film mercury
vapour analyser called JEROME® 431-X™ was used.
The instrument has a range of detection between 0.003
mg/m3 and 0.999 mg/m3 and has a resolution of 0.001
mg/m3. The precision of the device is a 5% standard
derivation at 0.100 mg/m3 of mercury. The air flow
through the device is 0.75 0.05 litre/min during a
measurement which takes 12 seconds. [35]
The different containers used to create air-tight volumes
were 14 litre, 31 1itre and 72 litre containers with lids. The
volumes of the containers were selected to make an even
distribution of volumes during measuring. The reason for
the usage of containers, and not performing the
measurements directly in a room, was to be able to
simulate broken CCFLs inside a LCD TV. Another reason
was to investigate if there is scalability depending on the
change of volume. A direct illuminated 42 inch LCD TV
with CCFLs has a volume of approximately 14.5 litres.
This is calculated with a depth of 30mm and the screen
ratio 16:9.
All measurements were performed in an air cabinet with a
constant flow of air at a minimum of 0.56 m/s. The
JEROME instrument and the 31 litre container are
illustrated in Figure 1 below.
Figure 1: Image of the setup
3.2 Material used
The CCFLs used were gathered from several LCD TVs of
different brands, years of manufacturing, designs and
sizes. The dimensions of the CCFLs are illustrated in
Table 2 below.
Table 2: Dimensions of CCFLs used during the study
# CCFLs Diameter Length Series
13 4mm 715mm 1-6
16 3mm 640mm 7
9 3.4mm 745mm 8 and V1
14 4mm 835mm 9-10
2 3.3mm 835mm 10
All lamps were collected from a discarded LCD TVs
provided by a recycling centre in Linköping, Sweden. A
recycling centre is where the community can hand in
waste products such as chemicals as well as electronic and
combustible products for future treatment, recycling or
incineration. The number of hours the LCD TVs were
used during thier life-times is unknown.
3.3 Plan of the empirical study
The study was divided into 10 series, with each series
representing the different setups of parameters. During
Series 1-7, the influence of changing the number of
broken CCFLs in a 31 litre container was investigated.
Series 8-10 were performed to investigate the influence of
changing the volume eight CCFLs were broken in. The
additional series was performed to validate if the
behaviour of mercury in this context was possible to
predict.
The study was performed to span a range of measurements
to support the understanding of the behaviour of the
outcome in the test series. When the behaviour was
assumed to be understood a hypothesis was created. To
falsify or validate the hypothesis a statement was made.
This statement was then investigated. The result of the
investigation then falsified the hypothesis if the
investigation contradicted the hypothesis, and validated
the hypothesis if the outcome was in line with the
hypothesis.
3.4 Measurement procedure
Before the study began the measurement device was
calibrated and warmed up according to the device
instruction manual and fitted with an extension tube.
Holes were also drilled in the containers to fit the
extension tube at the measurement device, and the air
cabinet was cleaned. The tests were made according to the
following procedure.
1. The container and the measurement instrument were
placed inside the air cabinet, the extension tube of the
device was fitted through the hole in the container,
and the end of the tube was placed as close to the
centre of the box as possible.
2. The background level of mercury in the air cabinet
and container was measured and documented.
3. The CCFL or CCFLs was or were placed in the centre
bottom of the container and broken carefully to try to
make a uniform cross section of the fracture. The box
was closed with a lid as fast as possible after the
CCFLs were broken.
4. Initially 10 measurements were made and
documented. The 10 measurements were repeated at
15min, 30min and in the complementing series also at
45min and 60min.
5. After all measurements were performed the CCFLs
were removed from the container and storage. The
container and the air cabinet were then cleaned from
possible glass and other materials.
3.5 Uncertainties and known factors affecting the test
outcome
During the study there were several uncertainties which
were difficult to quantify and anticipate. The identified
uncertainty was the amount of mercury in the CCFLs
before the study. The following known factors affect the
outcome of the study.
 The temperature in the room affects the evaporation
rate for mercury [36]. Room temperature was 20-21
degree Celsius.
 The air pressure also affects the evaporation rate for
mercury. The air pressure in the room was 101.3 kPa.
 The amount of air “flushing” the boxes during a
single measurement was 15cl. With 10 measurements
the amount becomes 150cl or 1.5 litre of air flushing
the containers.
4 RESULTS FROM THE EMPIRICAL STUDY
4.1 Results from Series 1-10
The results from the measurements conducted in Series 1-
7 and 9 are illustrated in Figure 2 while the values are
listed in Table 3. These series were performed to
investigate the influences the number of broken CCFLs
has on the mercury vapour concentration within a 31 litre
container. The results from Series 8-10 are presented in
Figure 3, while the values are listed in Table 4. These
series were performed to investigate the influences of
different volumes of containers when 8 CCFLs were
broken inside. Each pile in Figure 2 and Figure 3 together
with the numbers in Table 3 and Table 4 are average
values.
4.2 Expected results and results from the validating
series
The value in the validating series was expected to be
0.022±0.005mg/m3. The calculation of the expected value
is based on the measurements performed in Series 8 and 1-
5 and 9. The value in Series 8 after 60 min is 0.107mg/m3.
Compensating for the air flush issue, the value should be
0.230 mg/m3 as illustrated in equation below.

This is the expected value for 8 lamps broken in a 14 litre

container after 60 min. To recalculate the value of 1 CCFL
instead of 8, the relationship between Series 1-5 and 9 can
be used. The closest values with a relationship are after 30
min, with 0.010mg/m3 and 0.097mg/m3 respectively. The
increase in mercury concentration when changing from 1
CCFL to 8 is 918% based on the compensated values. The
same increase is assumed to be identical at 60 min. This
assumption, together with the compensated value from
Series 8 at 60 min gives the expected value of
Figure 2: Results from Series 1-7 and 9. *Series 9
0.0226 0.023mg/m3. The expected value without the air
Table 3: Results from Series 1-7 and 9 flush compensation is 0.022mg/m3. The variation of
# ±0.005 is based in the standard variation of the instrument.
CCFL Instan
Series s t 15min 30min 45min 60min The result from the validation series is 0.058mg/m3, 163%
1-5 1 0,013 0,010 0,010 higher than expected.
6 8 0,043 0,098 0,103
9 8 0,018 0,096 0,097 0,099 0,086
5 DISCUSSION
7 16 0,064 0,210 0,250
In the literature it is stated that the mercury occurrence as
a vapour and amalgamated to materials. [24-25, 32-33]
Mercury has not only been found in vapours from
fluorescent lamps; it has also been found in the electrodes
[17], in the glass [34] and in the fluorescent powder [32-
33] of the lamps. This indicates that this is the case with
CCFLs in LCD TVs as well. This makes it important to
take care of the entire CCFL, and not only the powder in
the lamps, something also stated by the previous authors
[17]. Even if end-of-life treatment processes are equipped
with ventilation systems which remove mercury vapours,
there will still be mercury present in the material for
recycling and in the processes equipment itself [17].
During the empirical study the following relationships
were found. If the number of broken CCFLs increases,
then the mercury concentration also increases. If the
container volume increases, then the mercury
concentration decreases. The highest value measured was
Figure 3: Results from Series 8-10. 0.250mg/m3 for 16 broken CCFLs in a 31 litre container
Table 4: Results from Series 8-10. after 30 min. This will make the concentration even higher
Cont if 16 CCFLs is broken inside a 42 inch TV. As the
Serie ainer Insta 15mi 30mi 45mi 60mi expected value and the validation series illustrates, the
s s nt n n n n concentration is difficult to predict as stated by Böni and
Widmer [17]. The expected value was not the same as the
8 14 0,038 0,138 0,133 0,121 0,107 validating value which clearly states that the behaviour of
9 31 0,018 0,096 0,097 0,099 0,086 mercury in CCFLs from LCD TV s is not fully
10 72 0,008 0,052 0,061 0,077 0,083 understand. The validation series also illustrates that the
concentration of mercury caused by a single broken
CCFLs in a 42 inch TV can be 0.058mg/m3 after
accumulation over one hour.
What to prefer for statistic validation, many measurement
in few CCFLs or few measurements on many CCFLs? The
drawback with our selected approach was the issue of air
flushing the sample volume each time a measurement was
performed. Why did we select this approach? This is due
to the limitation of the number of CCFLs and the variation
due the unpredictable nature of mercury. The reason for
not breaking more than 16 CCFLs in a container was the
high concentrations measured and that JEROME device
was saturated and had to be regenerated after a few
measurements with more than 16 CCFLs.
6 CONCLUSION
The literature reviews showed that not only does the
fluorescent powder in CCFLs in LCD TVs contain
mercury; in fact the entire CCFL, especially older and
used CCFLs, have mercury amalgamated to the glass and
the electrodes as well [25, 32-34]. The amount of mercury
in CCFLs varies in the literature from 0.1mg to 10mg per
lamp [11, 14, 17, 25-26]. The empirical study revealed the
difficulty in trying to predict the concentration if one or
several CCFLs are broken inside a container. The highest
concentration measured, 0.250mg/m3, was for 16 CCFLs
broken inside a 31 litre container measured after 30 min.
The concentration measured for a single CCFL, broken in
a 14 litre container representing a direct illuminated 42-
inch LCD TV, was 0.058mg/m3 when measured after 60
min. If a higher number of CCFLs are broken in a LCD
TV, then the mercury concentration can be expected to be
higher.
7 FUTURE RESEARCH
Ideas for future research are to complement the empirical
study and mathematical describing the relationship of
broken CCFLs, container and the concentrations of
mercury.
8 ACKNOWLEDGMENTS
The authors would like to thank the Swedish
Governmental Agency for Innovation Systems
(VINNOVA), Swedish foundation of strategic research
and ProViking for financing the research. Secondly, the
authors whole also like to thank everyone who supported
them during the empirical study and in the HPLA and
AutoDisA research projects, which were related to this
research.
REFERENCES
[1] Chang, T.C., You, S. J., Yu, B. S., Chen, C. M.,
Chiu Y. C. (2009) Treating high-mercury-
containing lamps using full-scale thermal
desorption technology, Journal of Hazardous
materials, 162, pp 967-972.
[2] Lindqvist, O., Johansson, K., Aastrup, M.,
Andersson, A., Bringmark L., Hovsenius, G.,
Håkanson, L., Iverfeldt, Å., Meil, M., Timm, B..
(1991) Mercury in the Swedish environment –
Recent research on causes, consequences and
corrective methods. Water, air and soil pollution,
55.
[3] Liu, G., Cai Y., O´driscoll, N., Feng X., Jiang, G.
(2012) Advances in environmental chemistry and
toxicology of Mercury, Chapter One Overview of
Mercury in the Environment, First Edition, John
Wiley and Sons.
[4] World Health Organization (1991), Inorganic
mercury, International Programme on Chemical
Safety, Geneva.
[5] Lindberg S., Stokes P., Goldberg E., Wren C.
(1987) Group report: Mercury, Lead, mercury,
cadmium and arsenic in the environment, John
Whily and sons, New York, Chichester Brisbane
Toronto, Scope 31, pp 17-34.
[6] Pirrone, N., Cinnirella, S., Feng, X., Finkelman, R.
B., Friedli, H. R., Leaner, J., Mason, R.,
Mukherjee, A. B., Stracher, G. B., Streets, D. G.,
Telmer, K. (2010) Global mercury emissions to the
atmosphere from anthropogenic and natural
sources, Atmoshperic Chemical and Physics,
Copernicus publications.
[7] World Health Organisation (2005) Mercury in
drinking-water, WHO/SDE/WSH/05.08/10.
[8] World Health Organisation (2000) Air Quality
Guidelines for Europe, Second edition, WHO
Regional Publications, European Series, No. 91,
ISBN 92 890 1358 3, ISSN 0378-2255.
[9] Swedish Environmental Protections Agency (2003)
Hitta kvicksilver i tekniksa varaor och produkter,
ISBN 91 620 5279 9, ISSN 0282-7298, CM
Digitaltryck AB, Bromma, Sweden.
[10] Unites States Environmental Protection Agency
(1997) Mercury Study – Report to congress.
Volume 1; Executive Summery.
[11] Chang, T. C., You, S.J., Yu, B. S., Kong, H. W.
(2007) The fate and management of high mercury-
containing lamps from high technology industry,
Journal of Hazardous Materials 141, Elsevier, 27
July 2006, pp 784-792.
[12] Peeters, J. P., Vangegas, P., Dewulf, W., Duflou,
J. R. (2012) End-of-Life Treatment Strategies for
Flat Screen televisions: A Case Study, The 9th
 Global Conference on Sustainable Manufacturing,
Saint Petersburg, Russia.
[13] Peters, J. P., Vanegas, P., Duflou, J.P., Mizuno, T.
Fukushige, S., Umeda, Y. (2013) Effects of
boundary conditions on the end-of-life treatment
of LCD TVs, CIRP Annual - Manufacturing
Engineering, 62, pp 35-38.
[14] Cryan, J.,Freegard, K:, Morrish, L., Myles N.
(2009) Demonstration of Flat panel display
recycling technologies, WRAP Final project
report, Project code MDD014.
[15] Kopacek, P., Kopacek, B. 2006 Intelligent,
Flexible Disassembly. International journal of
advanced manufacturing technology, 30, pp 554-
560.
[16] Stena technoworld. 2011.
www.stenatechnoworld.com, Accessed 2011-
11-08.
[17] Böni, H., Widmer, R. 2011 Disposal of flat panel
display monitors in Switzerland, Swiss federal
laboratories for material science and technology,
Swico recycling.
[18] Teisinger, J., Fiserova-Bergerova, V. (1965)
Pulmonary retention and excretion of mercury
vapours in man, Industrial Medicine and Surgery,
34, page 580.
[19] Swedish Work Environmental Authority (2005)
Hygienic limits and measures against air
contamination (org. title Hygieniska gränsvärden
och åtgärder mot luftföroreningar).
[20] Council of labour affairs 2006 Establishment of
risk assessment-oriented exposure assessment
strategy, Available at www.cla.goc.tw, Accessed
2013-09-06.
[21] Ausschuss für Gefahrstoffe 2006 Die Technischen
Regeln für Gefahrstoffe Arbeitsplatzgrenzwerte
TRGS 900, Available at www.buau.de. Accessed
2013-09-06.
[22] Health and safety executive 2013 EH40/2005
Workplace exposure limits, Available at
ww.hse.gov.uk, Accessed 2013-09-06.
[23] Journal of occupational health 2012
Recommendation of occupational exposure limits,
Available at joh.sanei.or.jp, Accessed 2013-09-06.
[24] Aucott, M., McLinden, M., Winka, M., 2004
Release of Mercury from broken fluorescent bulbs,
State of New Jersey Division of Science Research
and Technology Department of Environmental
Protection.
[25] McDonnel, T. J., Williams, K. S. (2010) The
location and character of mercury in waste LCD
backlights, WARP Project report, Project code:
MDD028.
[26] Figi, R., Nagel, O., Hagendorfer, H. 2012 A
straightforward wet-chemistry method for the
determination of solid and gaseous mercury
fractions in Backlight Cold Cathode Fluorescence
Lamps, Talanta, 100, pp 134-138.
[27] Monix 2013 www.monix.com.tw, Responsible
publisher Monix technology co. ltd. Accessed
2013-09-04.
[28] GIO 2013 web site www.giotaiwan.com.tw,
Responsible publisher GIO optoelectronics corp.,
Accessed 2013-09-04.
[29] Ko, J.-H., Ryu, J.-S., Yu, M.-Y., Park, S.-M., Kim,
S.-J. 2010 Initial Photometric and Spectroscopic
Characteristics of 55-inch CCFL and LED
Backlights for LCD-TV Applications, Journal of
the Korean institute of illumination and electrical
installation engineers, 24, 3, pp 8-13.
[30] Ryan, A., O’Donoghue, L., Lewis H. 2011
Characterising components of Liquid crystal
displays to facilitate disassembly, Journal of
cleaner production, 19, pp 1066-1071.
[31] Matharu, A. S., Wu, Y. 2009 Liquid crystal
displays: from devices to recycling, Hester R. E.,
Harrison R. M. (eds.). Electronic waste
management, The royal society of chemistry, pp
180-211.
[32] Dang, T. A., Frisk, T. A., Grossman, M. W. (2002)
Applications of surface analytical techniques for
study of the interaction between mercury and
fluorescent lamps material, Analytical and
bioanalytical Chemistry, number 373, issue 7, pp
560-570
[33] Raposo, C., Carvalinho, Windmöller, C., Durão,
W. A. Jr. (2003) Mercury speciation in fluorescent
lamps by thermal release analysis, Waste
management, 23, pp 879-886.
[34] Thaler, E. G., Wilson, R. H., Dought, D. A., Beer,
W. W. (1995) Measurement of mercury bound in
the glass envelope during operation of fluorescent
lamps, Journal of the Electrochemical Society,
142, 6, pp 1968-1970.
[35] Arizona Instrument LLC 2011 JEROMA® 431-X™
Mercury vapour analyser operation manual,
Available at www.azic.com, Accessed 2013-09-06.
[36] Lindberg, S. E., Turner, R. P. (1977) Mercury
emissions from chlorine-production solid waste
deposits, Nature, 268, pp 133-136.