International Journal of Occupational and Environmental

Health

ISSN: 1077-3525 (Print) 2049-3967 (Online) Journal homepage: http://www.tandfonline.com/loi/yjoh20

Case study of occupational mercury exposure

during decontamination of turnaround in refinery
plant

Maytiya Muadchim, Wantanee Phanprasit, Mark Gregory Robson, Dusit

Sujirarat & Rawee Detchaipitak

To cite this article: Maytiya Muadchim, Wantanee Phanprasit, Mark Gregory Robson, Dusit
Sujirarat & Rawee Detchaipitak (2018): Case study of occupational mercury exposure during
decontamination of turnaround in refinery plant, International Journal of Occupational and
Environmental Health, DOI: 10.1080/10773525.2018.1425657

To link to this article: https://doi.org/10.1080/10773525.2018.1425657

Published online: 15 Jan 2018.

Submit your article to this journal

Article views: 7

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=yjoh20
International Journal of Occupational and Environmental Health, 2018
https://doi.org/10.1080/10773525.2018.1425657

ORIGINAL RESEARCH PAPER

Case study of occupational mercury exposure during decontamination of

turnaround in refinery plant
Maytiya Muadchima  , Wantanee Phanprasitb  , Mark Gregory Robsonc  , Dusit Sujiraratd and
Rawee Detchaipitaka 
a
Faculty of Public Health, Department of Occupational Health and Safety, Mahidol University, Bangkok, Thailand; bFaculty of Public
Health, Department of Occupational Health and Safety, Mahidol University, Bangkok, Thailand; cRutgers- School of Environmental
and Biological Sciences, New Brunswick, NJ, USA; dFaculty of Public Health, Department of Biostatistics, Mahidol University, Bangkok,
Thailand

ABSTRACT ARTICLE HISTORY

Decontamination during turnaround (TA) can lead to high exposure to toxic chemicals among Received 11 May 2017
workers. The decontamination process in refinery plants usually comprises two types, i.e. steam Accepted 5 January 2018
and chemical decontamination. No matter the method used, concentrations of toxic chemicals KEYWORDS
must be measured using the direct reading instrument at the end of each decontamination Mercury; decontamination
cycle, which maybe repeated several times until the readings are in acceptable level. To worker; turnaround; refinery
evaluate mercury exposure of decontamination workers during decontamination procedure plant
in comprehensive turnaround of a refinery plant. Thirty personal and 16 area air samples were
collected using passive dosimeters and absorbent tubes, respectively, during 8  days of the
decontamination in comprehensive turnaround. All samples were analyzed using inductively
coupled plasma mass spectrometry (ICP-MS). Good practices and standard procedures for air
sampling and analysis were performed. The TWA concentrations were calculated and compared
between steam and chemical decontamination workers. All area samples were well below TLV
(ND – 0.0016 mg/m3) while 5 of 14 samples collected at the steam team and 1 of 16 samples taken
from the chemical team exceeded the TLV. The geometric mean (GSD) of TWA concentration of
the steam team was 0.0057(10.4906) mg/m3, which is about twice as high as that of the chemical
team, 0.0031(6.9422) mg/m3. The highest mercury concentration, 0.1037 mg/m3, was collected
from a steam decontamination worker. According to the activities and observation, the steam
team may have high exposure while reading the chemical concentrations at the end of the
decontamination cycle.

Introduction times higher than normal work routines and during the
comprehensive turnaround (TA) workers could have sig-
Mercury is a chemical of global concern due to its tox-
nificant exposure. TA shutdown of refineries to allow
icity and prevalence in atmosphere and environment.
for decontamination, repairs, replacements, inspec-
This rare element occurs naturally in the earth’s crust
tions, and overhauls to increase equipment reliability
combined with rock, fossil fuels, etc. The concentra-
to maintain production integrity and reduce the risk of
tion of mercury in crude oil, condensate and natural
catastrophic failures [4–7].
gas varies globally e.g. in the Gulf of Thailand and the
After shutdown, the remaining feeds in the produc-
Middle East presented 593 μg/kg [1] and 0.8 μg/kg [2]
tion system will be drained in underground storage
respectively. These products are used as feedstock in pet-
tanks to start the decontamination. Two decontami-
rochemical industries in Rayong Province, Thailand. In
nation methods are employed, i.e. chemical and steam.
2010, approximately 13,000 kg of mercury was released
For a unit with high concentration of mercury, chem-
from anthropogenic sources in Thailand and about 5%
ical decontamination would be employed (Figure 1).
was from oil and gas production, e.g. petroleum refin-
The hydrocarbon decontamination chemical is pumped
eries [3]. The release of chemicals in the environment
in the cleaning loop to remove residual oil, hydrogen
could occur due to leaks, spills, and fugitive sources
sulfide (H2S), benzene, combustible gasses, and pyroph-
during normal operations and maintenance. Certainly
oric iron sulfides including mercury accumulated on the
chemical exposure during maintenance could be several

CONTACT  Wantanee Phanprasit  [email protected], [email protected]

© 2018 Informa UK Limited, trading as Taylor & Francis Group
2   M. MUADCHIM ET AL.

utilities were divided in five sections. Each section was

controlled by each control panel, and only panels 1, 2,
3, and 5 were covered by this study. They comprised 34
chemical decontamination loop systems (15 columns
and 32 vessels) located in the same area and each panel
comprised several units as shown in Figure 2. Panel 4
was excluded because as utility unit, it does not involve
mercury.
When setting up a TA plan, high, and low mercury
contaminated equipment or units are identified. The low
mercury contaminated units would be cleaned up using
the steam method and the chemical decontamination
would be employed for those units with high mercury
contamination. Decontamination workers were divided
in two groups; steam decontamination and chemical
decontamination. Decontamination workers of the
steam decontamination team measured the concentra-
tions of the toxic chemicals at a manhole using direct
reading instruments at the end of each decontamination
cycle. The decontamination process ended only when
toxic chemical, e.g. mercury and benzene concentra-
tions are lower than their TLV (0.025 mg/m3, 0.5 ppm),
otherwise the decontamination cycle must be repeated.

Participants
The decontamination workers participating in the
Figure 1. Chemical decontamination process of the studied site. study were from two sources; the steam decontamina-
tion workers comprised of the refinery plant staff and
the chemical team comprised of subcontracted work-
surface of equipment, parts. Then the system is flushed ers. However, both teams were familiar with the work
with water and the wastewater is sent to waste storage procedures well although some of them had worked in
tanks or cooled with nitrogen gas. For a unit with high this position less than 1 year. Moreover, the contracted
hydrocarbon compounds but low mercury, the steam workers must pass the safety training course before
decontamination method would be used. The process recruitment to the TA project. While working in the
is similar, but hot water and steam are used to flush and TA area, they wore working coverall suits and half face
purge the system instead of using chemicals, and the respirators with 3 M mercury cartridges, 6009 for the
wastewater is drained to a storage tank or wastewater steam team and 6096 [mercury and organic vapor] for
treatment system depending on the concentration of the the chemical team.
toxic chemicals. No matter of what method is used, the Due to short duration of TA execution and high
decontamination process will be repeated until the toxic intense work, the work ran around the clock. For the
chemicals are at acceptable levels. two shifts, workers were split in two teams, day and night
Reports have been published of mercury exposure shifts. Both worked for 12 h daily everyday throughout
during TA for welding, confined space entry, and manual 58 days of the TA project.
removal of scale from vessels [8], but not for decon-
tamination. This study focused on mercury exposure
during decontamination of decontamination workers Sampling strategy
in a refinery plant. Approximately 110 decontamination workers were
employed each shift, 70 in the chemical team and 40 in
Material and methods the steam team. Thirty passive sampling badges (SKC
Inc., model # 520-02A, Lot No. 9962) were attached at
Study site and decontamination process the coverall collar of 16 steam decontamination workers
This study was conducted in a refinery plant located in and 14 chemical decontamination workers when they
Map Ta Phut District, Rayong Province, Thailand, where signed in to work and collected before they left the TA
a comprehensive turnaround was conducted between site. Sixteen Hydrar packed sorbent tubes (Carulite, Lot
May and June 2016. The decontamination work was per- 10187 Cat. No. 226-17-1A) were used to collect area air
formed in the first 8 days. The production process and samples according to the NIOSH method #6009. The air
 INTERNATIONAL JOURNAL OF OCCUPATIONAL AND ENVIRONMENTAL HEALTH   3

Figure 2. Coarse process diagram of a petroleum refinery.

Note: Panel 1: Crude Distillation Unit (CDU), High Vacuum Unit (HVU), Visbreaker Unit (VBU), Hydrodeslphurization Unit (HDS) and HDS fractionation
Section (HDF); Panel 2: Crude Re-contacting Unit (CDR), Naphtha Hydrotreater (NHT), Platformer Unit (PLF), LPG Treating Unit (LPG); Panel 3 consists of
Hydrocracking Unit (HCU), HCU Fractionation section (HCF), Adip Regeneration Unit (ADIP), Sulfur Recovery Unit (SRU), Shell Claus Off-gas Treating (SCOT)
and Sour Water Stripping (SWS); Panel 4: Hydrogen Manufacturing Unit (HMU), Hydrogen Purification Unit (PSA), Raw/Potable/Service Water System, Steam
Generation & Distribution Unit, Boiler Feed System and Nitrogen System and Panel 5: Condensate Residue Splitter (CRS), Kerosene Merox Unit (KMU) and
Deep Hydrodesulphurization Unit (DHDS).

sampling units were attached to a tripod and the sam- between 92 and 109%. The analysis conditions of plasma
pler was approximately 1.2 meter above the ground in were: RF power = 1500 W, RF Matching = 1.60 V, make
the decontamination areas where the workers operated up gas  =  0.32 L/min, nebulizer pump  =  0.1 cps, and
but without obstructing the work. The personal pumps flow of the carrier gas was 0.6 L/min. In addition, the
were calibrated before and after air sampling and the detector conditions were: discriminator = 6.0 mV, analog
average flow rate was used for time weight average HV = 1850 V, and pulse HV = 1440 V. Field blanks were
(TWA) concentration calculation. Both personal and prepared and analyzed in parallel with the samples.
area air samples were collected throughout 8  days of
decontamination. Statistical analysis
All workers who carried the badge were informed
about the study and asked for their voluntary enroll- The data were analyzed and described using descriptive
ment before signing a consent form approved by the statistics, e.g. mean and standard deviation. Correlations
Ethics Committee on Human Rights Related to Human between variables were assessed using Pearson’s corre-
Experimentation, Mahidol University. This research lations coefficient. Differences of mercury exposure
proposal was reviewed and approved under No. MUPH between groups was assessed using the paired simple
2016-029. T-test, Mann–Whitney test, and the Kruskal–Wallis test,
respectively. Furthermore, to analyze all data, the “not
detected (ND)” samples were replaced with half of the
Laboratory analysis
LOD [11]. Statistical analyses were performed using the
The analyzer was an ICP-MS (Agilent 7500 series, software package SPSS, version 22.0v.
JP82802997); thus, the sample preparation procedure
in ID 140 [9] and NIOSH method #6009 [10] were mod- Result
ified so that the samples were suitable for the analyzer. A
series of known concentrations of 2.5, 5, 10, 15, 20, and Due to the nature of the refinery plant, the production
25 ppb [mercury standard solution, Hg (NO3)2], were process was established outdoors; therefore, air move-
used to prepare standard curves. The r2 of the linear ment in the site could affect chemicals concentration.
regression of the standard curve was 1. The LOD was less The results of the air samples collected and analyzed
than 0.3 μg/m3, and the desorption efficiency ranged in for mercury are presented in Table 1. Among 16 area
4   M. MUADCHIM ET AL.

Table 1. Mercury concentrations in air samples collected during system decontamination.

Area sample Steam decontamination workers Chemical decontamination workers
Hg conc. (mg/ Responsibility Hg conc. (mg/ Responsibility Hg conc. (mg/
Panel Sampling area Datea m3) area Datea m3) area Datea m3)
1 CDU, De-salter 3rd ND CDU, HVU, VBU, 2nd 0.0646 CDU, De-salter 3rd 0.0116b e
#1 HDS, HDF
CDU, De-salter 3rd ND 3rd 0.0046 CDU, De-salter 3rd NDb
#2
CDU,V1093, 6th ND 4th ND CDU,V1093, 6th 0.0211b
C1091 C1091
CDU,V1691, 6th ND 6th 0.0027 CDU,V1093, 6th 0.0128b
V1693 C1091
CDU, FFU4, 8th ND
T5542
2 CDR, V1053 1st 0.0005 CDR,NHT,PLF 1st 0.0020 CDR, S1402 2nd 0.0017b
CDR # 1 2nd ND 5th ND CDR, S1402 2rd 0.0027c
CDR # 2 2rd ND 5th 0.0034
NHT, S1402 2rd 0.0016 7th ND
3 ADIP, C2001 5th ND HCU, ADIP, SRU 1st 0.1037 AD- 5th 0.0093b
IP,C2001,C2051
SCOT, C2301 5th ND 3rd 0.0457 SCOT, 5th 0.0116d
C2301,V2301
SCOT, 5th 0.0060c
C2301,V2301
5 CRS # 1 1rd 0.0006 CRS, KMU, DHDS 4th 0.0609 CRS, C2501 1st 0.0102b
CRS # 2 1rd 0.0008 5th 0.0237 CRS, C2501 1st NDb
DHDS, C2804 4th ND 6th 0.0381 CRS, V2501 7th 0.0065b
#1
DHDS, C2804 4th ND 7th 0.0094 CRS, V2501 7th NDb
#2
DHDS, V2808 4th ND CRS, V2501 8th NDb
DHDS, C2804 4th 0.0329b
DHDS, C2804 4th 0.0078b
Note: ND: not detected, (LOD ≤ 0.0003 mg/m3).
a
Sampling date.
b
Chemical cleaning.
c
Set up pipe system of chemical cleaning.
d
Safety supervision.
e
Maintenance.

samples, only 4 were detectable and the highest con- analyze the difference between the median of TWA con-
centration was 0.0016 mg/m3, which was lower than the centrations of the steam and the chemical teams, showed
TLV. Thus, workers in the TA site not involved with the no significant differences (p-value = 0.357), while those
system decontamination were less likely to be exposed of the area and the personal samples analyzed by paired
to mercury above TLV. simple T-test differed significantly (p-value = 0.001). In
Six members of the decontamination team were addition, the Kruskal–Wallis test was used to analyze
exposed to mercury above the TLV, and one of these was the difference between mercury exposures of workers
in the chemical team. The highest exposure (0.1037 mg/ in each panels presented in Table 3. The median values
m3) was among the steam decontamination worker who of mercury exposure of workers working in each panel
operated near the HCU, ADIP, and SRU Units, whereas did not significantly differ (p-value = 0.060).
the highest among the chemical team was 0.0329 mg/
m3 obtained from a worker who cleaned at the DHDS Discussion
Unit, C2804 of panel 5. The statistical characteristic of
the mercury concentrations are presented in Table 2. Chemical concentrations in the decontamination area
The data of all mercury concentrations, comprising log were anticipated to be high due to the activities of
normal distribution. Mann–Whitney test was used to decontamination; thus, area samples were collected

Table 2. Summary of statistical characteristics and test of the data.

Mercury concentration (mg/m3)
Sample n Median AM SD GM GSD Range p-value
Personal sample 0.357a
Steam team 14 0.0070 0.0256 0.0324 0.0057 10.4906 ND-0.1037
Chemical team 16 0.0071 0.0084 0.0088 0.0031 6.9422 ND-0.0329
Area sample 0.001b
Area 16 0.0002 0.0003 0.0004 0.0002 2.1723 ND-0.0016
Note: ND: not detected (significance at p-value < 0.05).
a
By Mann-Whitney Test.
b
Different test of area and personal sample (paired sample-t test).
 INTERNATIONAL JOURNAL OF OCCUPATIONAL AND ENVIRONMENTAL HEALTH   5

Table 3. Mercury exposure of workers in each panel.

Mercury exposure (mg/m3)
Panel N Median AM SD GM GSD Range p-value
1 8 0.0081 0.0147 0.0214 0.0039 2.6379 ND-0.0646 0.060
2 6 0.0018 0.0014 0.0015 0.0006 1.9667 ND-0.0034
3 5 0.0116 0.0353 0.0415 0.0198 1.6811 ND-0.1037
5 11 0.0094 0.0173 0.0195 0.0048 2.7532 ND-0.0609
Notes: AM; arithmetic mean, SD; standard deviation, GM; geometric standard deviation; ND: not detected; The mean difference was significant at p level
0.05.

to estimate exposure. Moreover, higher mercury con- Acknowledgments

centrations than those of the personal samples were This work was supported by the Office of the Higher
expected because they were at the site at all times. Education Commission, Thailand including the grant of the
However, the result showed an incompatibility between petrochemical company. The authors wish to thank all of
area and personal sampling, i.e. mercury concentra- petrochemical staff and contractors and laboratory staff from
tions obtained from the area samples were signifi- the Bureau of Occupational and Environmental Disease,
Reference Laboratory and Toxicology Center, Thailand and
cantly different and lower. This could be because the
support from NIEHS P30 Center Grant ES005022 and the
work was conducted in a well ventilated area [12] and New Jersey Agricultural Experiment Station.
another possibility is the contamination of the work-
ers clothing with mercury-contaminated wastewater or
chemicals which lead to continued exposure associated Disclosure statement
with evaporation of mercury from the contaminated No potential conflict of interest was reported by the authors.
clothing [13].
Based on the statistical analysis, both teams had sim- Funding
ilar exposure; however, most exceeding TLV samples,
five of six, were collected from the steam decontami- This work was supported by the PTT Global Chemical Public
Company Limited.
nation team. Interestingly, the chemical decontamina-
tion team worked with higher mercury contamination
units but did not exhibit elevated exposure. Thus, it ORCID
seems unlikely that the high mercury exposure among Maytiya Muadchim http://orcid.org/0000-0001-7443-0116
the steam decontamination team would result from Wantanee Phanprasit http://orcid.org/0000-0002-3928-3111
decontamination activities. The activity apart from the Mark Gregory Robson http://orcid.org/0000-0001-5702-4781
decontamination of the steam decontamination team Rawee Detchaipitak http://orcid.org/0000-0003-2962-3626
included measuring toxic chemicals in the manhole at
the end of the decontamination cycle, both for steam References
and chemical decontamination. The highest mercury
 [1]  Lopez-Garcia R, Radford R. Managing mercury in
exposure involved the steam decontamination workers hydrocarbon processing plants during turnarounds.
responsible for toxic chemical measurement and steam Hydrocarbon World [Internet]. 2012 [cited 2017 April
decontamination at the HVU, ADIP Unit, and SRU of 20]; 7(1):19–22. Available from: https://www.slideshare.
panel 3. net/rradford/hydrocarbon-world
Another possibility of exposure source of the steam   [2] Littpage T. Mercury in crude oils, MCA spring seminar
series [Internet]. 2013 [cited 2017 April 26]. Available
decontamination workers could have been the drain-
from: http://www.marinechemist.org/Mercury.pdf
age of wastewater after purging the system; wastewater   [3] Wongsoonthornchai M, Kwonpongsagoon S,
with low concentrations of toxic chemicals would be Scheidegger R. Modeling mercury flows in Thailand
drained for wastewater treatment. Because the waste- on the basis of mathematical material flow analysis.
water was hot and agitated due to the movement, CLEAN – Soil, Air, Water. 2016;44(1):16–24.
mercury vapor could have been released from the  [4]  Duffuaa SO, Daya MB. Turnaround maintenance
in petrochemical industry: practices and suggested
wastewater [14–15]. improvements. J Qual in Maint Eng. 2004;10(3):184–
190.
Conclusion  [5]  Chung EK, Jang JK, Koh DH. A comparison of
benzene exposures in maintenance and regular works
Mercury exposure was several times higher than the at Korean petrochemical plants. J Chem Health Safety.
TLV and the highest levels were found among steam 2017;24(3):21–26.
 [6]  Akerstrom M, Almerud P, Andersson EM, et al.
decontamination workers. Potential exposure sources
Personal exposure to benzene and 1,3-butadiene
could be the decontaminated units when measuring during petroleum refinery turnarounds and work
toxic chemical concentrations and in the wastewater in the oil harbour. Int Arch Occup Environ Health.
drained in the water treatment system. 2016;89(8):1289–1297.
6   M. MUADCHIM ET AL.

  [7] Chung EK, Shin JA, Lee BK, et al. Characteristics of [11] Glass DC, Gray CN. Estimating mean exposures from
occupational exposure to benzene during turnaround censored data: exposure to benzene in the Australian
in the petrochemical industries. Saf Health Work. petroleum industry. Ann Occup Hyg. 2001;45(4):275–282.
2010;1(1):51–60. [12] Cherrie JW. The beginning of the science underpinning
 [8]  Estevez FJ. Mercury in gas operation-occupational occupational hygiene. Ann Occu Hyg. 2003;47:179–185.
health risk assessment [Graduate Theses & Non- [13] Stopford W, Bundy SD, Goldwater LJ, et al.
Theses]. Montana Tech of the University of Montana; Microenvironmental exposure to mercury vapor. Am.
2015. Industr. Hygiene Assoc. J. 1978;39:378–384.
 [9]  Occupational Safety and Health Administration. [14] Kim MK, Zoh KD. Fate and transport of mercury in
Mercury vapor in workplace atmospheres ID140. Salt environmental media and human exposure. J Prev Med
Lake City, Utah; 1991. Public Health. 2012;45(6):335–343.
[10] National Institute of Occupational Safety and Health. [15] Hargreaves AJ, Vale P, Whelan J, et al. Mercury and
NIOSH manual of analytical methods (NMAM): antimony in wastewater: fate and treatment. Water Air
mercury, method 6009, 4th ed. 2; 1994. Soil Pollut. 2016;227:89.