Environmental Pollution 210 (2016) 246e252

Contents lists available at ScienceDirect

Environmental Pollution
journal homepage: www.elsevier.com/locate/envpol

Chemical composition of scales generated from oil industry and

correlation to radionuclide contents and gamma-ray measurements of
210
Pb*
Lina Al Attar*, Bassam Safia, Basem Abdul Ghani
Department of Protection and Safety, Atomic Energy Commission of Syria, Damascus, P.O. Box 6091, Syria

a r t i c l e i n f o a b s t r a c t

Article history: Scale generated from the maintenance of equipment contaminated by naturally occurring radioactive
Received 3 September 2015 materials may contain also chemical components that cause hazardous pollution to human health and
Received in revised form the environment. This study spotlights the characterisation of chemical pollutants in scales in relation to
12 November 2015
home-made comparison samples as no reference material for such waste exists. Analysis by energy
Accepted 7 December 2015
dispersive x-ray fluorescence, with accuracy and precision better than 90%, revealed that barium was the
Available online 29 December 2015
most abundant element in scale samples, ranging from 1.4 to 38.2%. The concentrations of the toxic
elements such as lead and chromium were as high as 2.5 and 1.2% respectively. Statistically, high cor-
Keywords:
Scale
relation was observed between the concentration of Ba and Sr, sample density, radionuclide contents
Chemical pollutant (210Pb and 226Ra) and self-attenuation factor used for the radio-measurements. However, iron showed a
Comparison sample reverse correlation. Interpretation of data with regards to the mineralogical components indicated that
226
EDXRF Ra and 210Pb co-precipitated with the insoluble salt Ba0.75Sr0.25SO4. Since both Ba and Sr have high Z,
Self-attenuation samples of high density (r) were accompanied with high values of self-attenuation correction factors (Cf)
for the emitted radiation; correlation matrix of Pearson reached 0.935 between r and Cf. An attempt to
eliminate the effect of the elemental composition and improve gamma measurements of 210Pb activity
concentration in scale samples was made, which showed no correction for self-attenuation was needed
when sample densities were in the range 1.0e1.4 g cm
3. For denser samples, a mathematical model was
developed. Accurate determinations of radionuclide and chemical contents of scale would facilitate
future Environmental Impact Assessment for the petroleum industry.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction
Naturally occurring radioactive materials (NORM) are widely
recognised as a problem in countries with developing oil and gas
industries. A significant contamination by radionuclides from 232Th
and 238U series, which are present in varying concentration in hy-
drocarbon reservoirs, arises from their accumulation and co-
precipitation with barium and strontium salts, forming a dense
caked layer (ranging from mm to a few mm in depth) on the inner
surfaces of processing equipment, separator vessels, wellheads,
tubular and pipelines (Al-Saleh and Al-Harshan, 2008; Gazineu and
Hazin, 2008; Heaton and Lambley, 1995). Descaling operations give
*
This paper has been recommended for acceptance by Charles Wong.
* Corresponding author.
E-mail address: prscientific@aec.org.sy (L. Al Attar).
rise to scale wastes. The activity concentrations of scales vary be-
tween several becquerels to thousands Bq/g of sample (Al Attar
et al., 2015a; EPA, 2001; Godoy and Cruz, 2003; Zielinski and
Otton, 1999). In addition to NORM, scales contain corrosion prod-
ucts, oil and heavy metals that would be more abundant than the
radionuclides (i.e. mainly Ba and Sr as sulphates) (Gazineu et al.,
2005; O'Brien et al., 2011; Varskog and Gellermann, 2011). NORM
of low activity concentrations can be discharged to the environ-
ment, however, those of high concentrations are often classified as
low level radioactive waste (IAEA, 2004; Liland et al., 2012; Read
et al., 2004; Varskog and Gellermann, 2011; Zielinski and Otton,
1999).
Appropriate handling and disposal of scale waste require
adequate chemical and radiochemical characterisation, which is a
challenging task owing to their radioactivity content which
greatly exceeds the environmental level, complex chemical
composition and high density (Al Attar et al., 2015a; Gazineu and

http://dx.doi.org/10.1016/j.envpol.2015.12.007
0269-7491/© 2015 Elsevier Ltd. All rights reserved.
 L. Al Attar et al. / Environmental Pollution 210 (2016) 246e252
Hazin, 2008; EPA, 2001; Obiajunwa et al., 2002). Determination
of the activity concentration in scale samples by gamma spec-
trometry is susceptible to sample-absorption, so-called self-
attenuation due to the presence of heavy metals (with high Z),
especially for low g-energy lines (i.e. <100 keV); this effect can
be corrected experimentally or theoretically (Al Attar et al., 2014;
El Afifi and Awwad, 2005; Gazineu and Hazin, 2008; Habib et al.,
2010, 2014; Hrichi et al., 2013). However, under- or over-
estimation of the activity concentration of radionuclides cannot
be ruled out.
Literature surveys show that limited studies have focused on
estimating the chemical composition of oil-produced scale by en-
ergy dispersive x-ray fluorescence (EDXRF) (Al-Saleh and Al-
Harshan, 2008; Al-Masri and Aba, 2005; El Afifi and Awwad,
2005; Hamlat et al., 2001; Obiajunwa et al., 2002). Although no or
inappropriate reference materials have been used for elemental
analysis, correlation to self-attenuation correction factors has not
been considered.
As part of an on-going research programme to treat and manage
oil solid wastes from the Syrian oil fields, a primary study (Al Attar
et al., 2015a) reveals that the activity concentrations of 226Ra and
210
Pb, in scale samples, were as high as 2922 and 1794 Bq g
1,
respectively. Self-attenuation correction factors were estimated in
another work (Al Attar et al., 2014) by transmission measurements
using a 152Eu point source and found to reach 5.08 for the most
dense scale samples at gamma energy line of 45.5 keV.
In this study, an extensive effort was made to prepare a number
of home-made comparison samples for elemental characterisation
of scale samples by EDXRF. Correlation of elemental composition to
density of scale samples, self-attenuation correction factors and
radionuclide concentrations was investigated in detail. Finally, an
approach was made to optimize the value of self-attenuation
correction factors for the determination of the radioactivity con-
centration of 210Pb in scale samples by gamma measurements.
2. Materials and methods
2.1. Scale sampling and preparation
A total of 32 scale samples were collected from a number of
drums that are kept at the interim storage site of NORM decon-
tamination facility (NDF) of Al Furat Petroleum Company, i.e.
500 km to the northeast of Damascus. The drums were chosen to
cover the entire recorded range of the measured gamma dose rates
(at contact, using Rados 110), i.e. 1e700 mSv h
1. Sampling was
carried out, in duplicate or triplicate, using a pressurised-stainless
steel Auger (inner diameter of 8 cm) along the centre of the
drum. Each sample, weighted ca. 80e120 g, was double-bagged in
plastic and tightly sealed to ensure secure containment.
The scale samples were then transported to the laboratory, air-
dried, ground and homogenised. Samples were divided into two
halves. One half was used to determine the elemental composition
using EDXRF as described later. The other half was dissolved ac-
cording to the protocol mentioned in an earlier work (Al Attar et al.,
2015b) and then counted via HPGe gamma spectrometry after 4
weeks to allow establishment of secular equilibrium between 222Rn
and progeny.
2.2. Sample preparation for elemental analysis
At present, there is no reference material available with speci-
fied concentration of chemical elements for NORM-scales
(Landsberger et al., 2013). Hence, it was important to prepare so
called “home-made comparison samples” using grade-reagent
anhydrous inorganic salts, i.e. CaCO3, Na2CrO4, Fe2O3, ZnCl2,
247
Sr(NO3)2, BaSO4, PbSO4, Na2SO4, K2CO3, and SiO2. Six home-made
comparison samples were prepared by weighing an aliquot mass of
the appropriate salt (Table 1), that were dried overnight at 40  C in
the oven (Carbolite, England), ground by a mechanical mill (Fritsch,
Germany), sieved through 185 mm (Retsch, Germany) and homo-
genised using Turbula mixer (Basel, Schweiz) for 48 h. About 2 g of
the fine powder were then pressed into a pellet (of 2 cm diameter)
using hydraulic press (MP 250 Massen GmbH, Germany) at 15
tonne. Each pressed pellet was finally covered by a Mylar foil and
placed into a polyethylene cup of 4.91 cm2. The scale samples were
prepared in the same manner.
2.3. EDXRF measurements
The EDXRF used for elemental analysis was TX-2000 spec-
trometer (Ital structures, Italy) equipped with a Si(Li) semi-
conductor detector, with an energy resolution of 160 eV at 5.9 keV.
Excitation of Ka and La-lines (in the case of Ba and Pb), within the
energy range 3.3e16 keV, was generated by a 2 kW Mo-tube
operated at a current of 20 mA and 35 keV voltage. Spectra were
collected for 500 s live-time and analysed using QXAS-AXIL
(Quantitative X-ray Analysis System and Analysis of X-ray spectra
by Iterative Least square fitting) package (version 3.6), designed by
the International Atomic Energy Agency (IAEA, 2007).
Precision and accuracy of the analysis technique were estimated
by preparing five replicates of each home-made comparison sample.
The theoretical and the measured concentrations mean (in % mass)
of elements studied in the six samples are shown in Table 2; triple
measurements were performed for each of the five replicate sam-
ples. The remaining matrix components were oxygen, sulphur and
smaller amounts of various other elements (Table 3).
2.4. Correlation of the chemical composition of scale samples with
self-attention correction factors
When a gamma photon passes through a sample it undergoes
interactions, within the sample itself, that reduce the g-rays
reaching the detector, this effect is refereed as self-attenuation. The
degree of self-absorption/attenuation depends on material density,
elemental composition and gamma-ray energy (Cutshall et al.,
1983; Kitto, 1991). In recent studies (Al Attar et al., 2014, 2015a),
the activity concentrations of 210Pb and 226Ra in scale samples and
self-attenuation correction factors for gamma emitters (i.e. Cf, see
equation (1)) were determined via a transmission method using a
152
Eu point source. Herein, the effect of the elemental composition
of scale samples on the activity concentration of the radionuclides
and Cf values was investigated in detail. Statistical approaches viz.
principal component analysis (PCA) and analysis of variance
(ANOVA) were applied using XLSTAT Programme (2014).
1
ðI=I0 Þ
Cf ðEÞ ¼ (1)
lnðI0 =IÞ
where Cf(E) is the selfeattenuation correction factor for the effi-
ciency at a particular g-energy, I and I0 are the attenuated and
unattenuated photon intensities, respectively.
2.5. Optimising the estimation of self-attenuation correction factor
of scale samples at 46.5 keV
To achieve better estimation of self-attenuation correction fac-
tors for radiochemical characterisation of scale by gamma spec-
trometry, scale samples were totally dissolved according to the
protocol stated in a previous work (Al Attar et al., 2015b). This
eliminated the effect of elemental composition and gave the
248 L. Al Attar et al. / Environmental Pollution 210 (2016) 246e252

Table 1
Weight (g) of the corresponding inorganic salts of the studied elements for the home-made comparison samples.

BaSO4 CaCO3 Fe2O3 K2CO3 Na2CrO4 Na2SO4 PbSO4 SiO2 Sr(NO3)2 ZnCl2

CS1 40.000 2.497 5.146 e 24.922 e 8.782 e 14.492 4.160.

CS2 42.488 2.498 5.146 12.665 3.115. 12.500 2.927. e 14.492 4.169
CS3 42.488. 2.498 18.587 7.225 3.115 4.500 2.927 e 14.492 4.169
CS4 16.946 4.995 43.321 e 12.461 e 8.782 e 7.241 6.253
CS5 5.099 4.995 50.041 12.500 6.230 8.486 1.464 5.809 1.208 4.169
CS6 5.098 14.9850 28.595 17.500 3.115 10.635 1.464 4.894 1.207 12.506

Weighing was carried out using 4 digits electronic balance (Sartorius) with uncertainty of ±0.006 mg.

Table 2
Elemental concentrations (in % mass) of the home-made comparison samples by EDXRF.

Na Ca Sr Ba Cr Fe Zn Pb K

CS1 7.1 1.0 6.0 23.5 8.0 3.6 2.0 6.0 0

e 1.03 ± 0.05 6.12 ± 0.07 24.07 ± 0.40 8.08 ± 0.24 3.77 ± 0.18 2.03 ± 0.06 6.10 ± 0.11 e
CS2 4.9 1.0 6.0 25.0 1.0 3.6 2.0 2.0 7.2
e 1.07 ± 0.10 6.27 ± 0.10 25.06 ± 0.40 0.98 ± 0.03 3.79 ± 0.10 2.04 ± 0.03 2.04 ± 0.20 6.93 ± 0.20
CS3 2.3 1.0 6.0 25.0 1.0 13.0 2.0 2.0 4.0
e 1.07 ± 0.08 6.11 ± 0.10 25.48 ± 0.40 0.91 ± 0.02 13.40 ± 0.20 2.13 ± 0.03 2.05 ± 0.03 3.92 ± 0.20
CS4 3.5 2.0 3.0 10.0 4.0 30.3 3.0 6.0 0
e 2.05 ± 0.10 2.75 ± 0.10 10.24 ± 0.50 4.16 ± 0.20 29.42 ± 0.60 3.13 ± 0.03 6.23 ± 0.60 e
CS5 4.5 2.0 0.5 3.0 2.0 35.0 2.0 1.0 7.1
e 2.24 ± 0.10 0.52 ± 0.02 3.21 ± 0.10 2.19 ± 0.10 33.26 ± 0.20 2.19 ± 0.10 1.05 ± 0.10 7.18 ± 0.20
CS6 4.3 6.0 0.5 3.0 1.0 20.0 6.0 1.0 9.9
e 6.07 ± 0.09 0.52 ± 0.04 3.15 ± 0.10 1.06 ± 0.03 20.08 ± 0.20 6.05 ± 0.02 1.09 ± 0.10 10.14 ± 0.30

Top values on each raw are the theoretical values, whereas the bottom ones represent the measured values.

Table 3 For the scale samples, the data recorded was the average of
The remaining matrix components (in % mass) for EDXRF analysis. triple measurements (with duplicate for each sample). Using AXIL
C O S Si H programme, the net peak area intensities for the characteristic lines
CS1 0.3 32 6.4 e 3.0
of the elements were calculated by spectra fitting; ES (Elemental
CS2 1.4 33 9.0 e 3.0 Sensitivities) enabled to calibrate the XRF sensitivities for a group of
CS3 1.0 31 7.0 e 3.0 elements from the standards (home-made comparison samples)
CS4 0.6 30 3.0 e 3.0 taking into account the x-ray attenuation. The sensitivities of the
CS5 1.7 33 2.8 2.7 3.0
standards were then used to determine the concentrations of the
CS6 3.3 33 3.3 2.3 3.0
elements in the unknown samples using QXAS. Fig. 1 illustrates the
The values were generated by QXAS-AXIL software of the XRF depending on EDXRF spectra of a scale sample and its comparison home-made
Compton and Rayleigh scattering ratio.
sample with the fitted x-ray lines of the elements studied.
Box plots of the concentration of the elements studied in scale
opportunity to ignore the high density and counting-level (i.e. at samples were constructed and are shown in Fig. 2. The elemental
the detector face or 5 and 10 cm away from the detector) of gamma composition in scales was, in general, enriched when compared to
measurements. The activity concentration of 210Pb (46.5 keV) in the soil or sediments contents. Clearly, barium was the most abundant
dissolved scale samples was then determined using N-type HPGe
detector (Bruker company, model GCD-60 230, Latvia) with relative
efficiency of 60% and FWHM of 0.89 and 2.0 keV at 122 and
1332.5 keV, respectively. Counting was carried out for a sufficient
period of time to attain uncertainty of 5e10%. The count rates of
blank and reference solutions were simultaneously determined.
The spectra obtained were analysed by InterWinner7 PC software
(ITECH-Instruments, 2011).

3. Results and discussion

3.1. Elemental composition of scale samples by EDXRF

The elemental concentration of Ca, Sr, Ba, Cr, Fe, Zn, Pb and K in
the home-made comparison samples were evaluated by EDXRF as
summarised in Table 2. The results obtained of the five replicates
(n ¼ 5) of each home-made sample were in good agreement with
the theoretical concentrations giving errors less than 10%. The
precision of XRF measurements was estimated in terms of the
relative standard deviation (RSD) and did not exceed 10% for all Fig. 1. EDXRF spectra of a home-made comparison sample (bottom) and a scale sample
elements of interest. (top) with the main excited Ka and La-lines (in the case of Ba and Pb).
 L. Al Attar et al. / Environmental Pollution 210 (2016) 246e252 249

No reference materials yet exist for the characterisation of scale

produced from oil and gas industries; in addition, the use of typical
Median Mean Min./Max. 25-75%
NORM-soil or geological samples is not an adequate comparison.
45
Thus, an attempt was made in this study to prepare home-made
40 comparison samples that somewhat simulate the matrix composi-
tion of scales, so that chemical content of scales would be more
35 representative. Depending on the mineralogical composition of
scale samples stated earlier (Al Attar et al., 2015b), the remaining
30
Concentration %

matrix components were oxygen, sulphur and smaller amounts of

25 various other elements (see Table 3). These components were, more
or less, compatible to those stated by Al-Saleh and Al-Harshan
20 (2008), where sulphur, silica and oxygen concentrations were
1.8e16.9, 0.8e1.5 and 40%, respectively. Carbon content was esti-
15 mated in this study to be ca. 0.3e3.3%, while Mansour et al. (2012)
reported carbon content via ignition to be 6.2%.
10

5
3.2. Correlation of the chemical composition of scale samples with
0 self-attention correction factors
Ca2+ S r2+ Ba2+ Cr3+ Fe 3+ Zn2+ Pb2+
In order to verify the effect of chemical composition of scale
Fig. 2. Boxplots of the concentration (%) of the studied elements in the scale samples. samples on self-attenuation correction factors used for the
determination of radionuclide concentrations by gamma spec-
troscopy, it was necessary to apply statistical approaches such as
element, ranging from 1.4 to 38.2%; iron content varied between 1.1
Principal Component Analysis (PCA) and correlation matrix of
and 37.9%. The concentrations of strontium and calcium in scale
Pearson.
samples gave mean values of 3.5 and 4.8%, respectively. The toxic
PCA explains the covariance structure of a set of variables. Data
elements, i.e. lead and chromium, were present with mean values
are represented as a Correlation circle on axes F1eF2 (Fig. 3). The
of 2.5 and 1.2%, respectively.
The chemical contents in scale samples described in the litera-
ture are shown in Table 4. Obviously, the data obtained in this study
fitted in the worldwide ranges, with the fact that barium was the
most abundant element in most countries (maximum ranging
22e40%, in Algeria, Syria and the USA). In some cases, the con-
centration of calcium was also high reaching 43% in Red Sea Re-
finery (Mansour et al., 2012) and 30% in Syria (Al-Masri and Aba,
2005). In the latter case, the scales were collected by scrubbing
the NORM-contaminated equipment, whereas in our study the
scale samples were collected from the drums accumulating scales
that are generated by decontamination. Using high pressure water
jets (HPWJ) for decontamination would cause a partial dissolution
of calcium minerals due to the fact that solubility of calcium salts is
higher than barium and strontium. This has been seen when co-
produced water of Syrian oil-fields were investigated in details
(Al Attar and Safia, 2013) and found that the concentration of Ca
ranged 4e20 g L
1, while Sr and Ba concentrations were lower by
one or two order of magnitude. The high content of barium
observed in scale samples confirmed previous x-ray diffraction
findings (Al Attar et al., 2015b) where strontian-baryte
(Ba0.75Sr0.25SO4) was the main component in addition to either
(Ba0.69Pb0.31SO4) or Ca5(PO4)3OH. The occurrence of lead, zinc and Fig. 3. Correlation (on axes F1eF2) of the activity concentrations, % elemental con-
iron sulphides was also reported. centration, density (r) and attenuation correction factor (Cf) of the scale samples.

Table 4
Worldwide elemental concentrations (in % mass) of scale petroleum waste.

Country Ca Sr Ba Cr Fe Zn Pb Reference

Algeria 1.1 6.4 39.5 e e e 3.9 Hamlat et al., 2001

Egypt (Red Sea) 42.8 6.7 10.9 e 21.1 0.8 1.3 Mansour et al., 2012
Egypt (Abu Rudies) 7.8e17.9 0.7e1.7 2.1e6.4 e 27.1e30.0 e e El Afifi and Awwad, 2005
KSA (Riydah City) 1.0e4.0 e e 0.01e0.02 15.6e90.2 0.08e0.09 e Al-Saleh and Al-Harshan, 2008
Nigeria (Niger Delta) 19.7 0.1 736a 11.2 171a 39.9a Obiajunwa et al., 2002
Syria 1.0e30.0 0.2e9.2 0.4e25.0 e e e 0.0e5.0 Al-Masri and Aba, 2005
This work (NDF) 0.5e16.0 0.1e10.7 1.4e38.2 0.04e5.2 1.1e37.9 0.9e7.7 0.2e6.5 e
USA (West Texas) 1.5 5.3 22.0 e 3.1 e e Landsberger et al., 2013
a
Values are expressed in mg g
1.
250 L. Al Attar et al. / Environmental Pollution 210 (2016) 246e252

variables studied were elemental concentration (in %), the density
of samples, i.e. ranging 1.04e3.08 g cm
3, Cf values at 45.5 keV, i.e.
1.17e5.08 (Al Attar et al., 2014) and radioactivity concentration of
226
Ra that was in the range 2e2922 Bq g
1 and 210Pb that ranged
from minimum detectable activity (i.e. 0.065) to 1794 Bq g
1 (Al
Attar et al., 2015a). Fig. 3 shows that the applied projection
explained 73.96% of the variability, with eigenvalues of l1 ¼ 6.349
(on F1: 58.54%) and l2 ¼ 1.694 (on F2:15.42%). High correlation
occurred between F1 axis and six variables, they are 226Ra, 210Pb, Ba,
Sr, Cf, and r, with factor loadings of 0.973, 0.894, 0.971, 0.943, 0.96
and 0.849, respectively. Iron represented a reversible correlation on
F1 axis (
0.832) and so did Ca but to a less extent (
0.635). The two
variables Pb and Cr were in correlation with F2, accompanied with a
factor loadings of 0.869 and 0.835, respectively. However, Zn
showed no correlation to either axis.
The preceding data were confirmed by Pearson correlation
matrix shown in Table 5. Clearly, the radioactivity concentration of
226
Ra illustrated a significant correlation with both of Ba and Sr, in
addition to Cf. To a lesser extent, 210Pb showed similar behaviour.
This ascertained that 226Ra and 210Pb co-precipitated with the main
mineralogical component, i.e. Ba0.75Sr0.25SO4, which in turn would
be the mineral affecting the attenuation phenomena in scale
samples. Godoy and Cruz (2003) also stated a stepwise linear cor-
relation between 210Pb and Ba content in scale generated by the
petroleum industry. Since these two elements (Sr and Ba) have high
Z, their presence in scale would result in high density of samples.
Consequently, correlation between Cf and density was considerably
high (0.935). Reverse correlation was noted between Ba and Fe
contents (
0.787), which was reflected also in a negative propor-
tion between the latter element and Cf (
0.774). The high con-
centration of Fe found in scale samples, a mean of 16.8% (Fig. 2),
illustrated that its source may not be from the geological formation
of the reservoir as for other elements but from the maintenance job
performed on the contaminated equipment using HPWJ, which
may cause scraping a thin metallic layer from the inner surface of
the equipment. Notably, when the concentration of Fe was high, the
activity concentration of 226Ra, 210Pb and the concentration of Ba
and Sr were low. Correlation of Cf with the rest of the elements
studied (Cr, Zn and Pb) was absent.
Plots of ANVOVA are illustrated in Fig. 4, where a linear rela-
tionship was seen between Cf and both of Ba and Sr (associated
with regression coefficients of 0.894 and 0.883, respectively).
However, Fig. 5 shows a logarithmic function between Cf and Fe,
whereas fractional function demonstrated the relation with Ca.
3.3. Optimizing the estimation of self-attenuation correction factor
of scale samples at 46.5 keV
As aforementioned in the experimental section, an attempt was
made to verify the values of self-attenuation correction factors
obtained elsewhere, i.e. Cf, (Al Attar et al., 2014) by eliminating the
influence of the chemical matrix of scale samples (i.e. the effect of
sample density). This was performed by totally dissolving the 32
scale samples that have apparent density in the range
1.04e3.08 g cm
3. When counting by HPGe detector, emphasis was
placed upon the activity concentration of 210Pb since self-
attenuation would be high at this low g-energy line (i.e. 46.5 keV).
Counting the dissolved scale samples enabled us to compare the
activity concentrations of 210Pb in these samples with those prior to
dissolution (in the solid state, without transmission). The ratio of
the activity concentrations of 210Pb in the dissolved scales to the
concentration of 210Pb in the solid samples (without correction)
gave a new correction factor that is designated as CD, which was
found to range from 1 to 4.1.
This confirmed the following two cases;
 for samples with density 1.0e1.4 g cm
3 and count per second
(cps) up to 1300, an overestimation in determining the activity
concentration of 210Pb in scale samples by applying the trans-
mission method (using Cf) did occur to an extent of 40%. Hence,
no correction for self-attenuation should be made, which would
then give an acceptable activity concentration within 10e15%
error;
 for samples with density 1.45e3.08 g cm
3 and cps between
1350 and 14,000, the obtained correction factor by transmission
measurement (Cf) would be adequate with 20% error. To
improve the error to less than 15%, the mean correction factor
(CM, see equation (2)) was plotted versus cps of solid scale
samples (Fig. 6). This gave a simplified model of a logarithmic
function with a regression coefficient of 0.962. From Fig. 6, the
obtained mathematical CM would be the attenuation correction
factor to adjust the activity concentration of 210Pb in the solid
scale samples instead of that obtained by the transmission
method.
Cf þ CD
CM ¼ (2)
2
To summarise, the dissolution process improved the determi-
nation accuracy of the activity concentration of 210Pb in scale
samples, i.e. with density of 1.0e1.4 g cm
3, by a factor of 25e30% in
comparison to the transmission method as no correction should be
done. For denser samples (1.45e3.08 g cm
3), using the mathe-
matical approach gave attenuation correction factor (CM) in the
range 1.7e4.1, somewhat similar to that attained by dissolution, CD
(i.e. 1.4e4.1). However, self-attenuation correction factors obtained
via. transmission (Cf) were previously reported to be 1.6e5.1 (Al
Attar et al., 2014). The approach used herein improved the
accompanied error in the activity concentration of 210Pb to less than

Table 5
Correlation matrix of Pearson “n” of the studied parameters.

Variables r (g mL
1) 226

Ra (Bq g
1) 210
Pb (Bq g
1) %Ca2þ %Sr2þ %Ba2þ %Cr3þ %Fe3þ %Zn2þ %Pb2þ Cf

r (g mL
1) 1
226
Ra (Bq g
1) 0.766 1
210
Pb (Bq g
1) 0.711 0.871 1
%Ca2þ
0.434
0.618
0.494 1
%Sr2þ 0.709 0.969 0.842
0.616 1
%Ba2þ 0.750 0.981 0.847 ¡0.658 0.985 1
%Cr3þ 0.117 0.135 0.074
0.286 0.083 0.164 1
%Fe3þ ¡0.662 ¡0.784 ¡0.725 0.299 ¡0.760 ¡0.787
0.127 1
%Zn2þ
0.118
0.163 0.076 0.038
0.034
0.111
0.235 0.232 1
%Pb2þ 0.298 0.108 0.219 0.170 0.087 0.138 0.671
0.328
0.063 1
Cf 0.935 0.908 0.855
0.574 0.844 0.884 0.148 ¡0.774
0.198 0.232 1

Values in bold are different from 0 with a significance level alpha ¼ 0.001.
 L. Al Attar et al. / Environmental Pollution 210 (2016) 246e252 251

6 6
(a) (b)
5 5

4 4

3 3
Cf

Cf
2 2

1 1
R2=0.894 R2=0.883
0 0
0 10 20 30 40 0 2 4 6 8 10 12
%Ba+2 %Sr+2
Conf. interval (Mean 95%) Conf. interval (Obs. 95%) Conf. interval (Mean 95%) Conf. interval (Obs. 95%)

Fig. 4. ANOVA of Cf of scale samples and the % elemental concentration of: (a) Ba2þ and (b) Sr2þ.

6 6
(a) y = -1.258ln(x) + 5.3897
(b) y = 2.7441x-0.364
5 R² = 0.8093 5 R² = 0.7104

4 4

3 Cf 3
Cf

2 2

1 1

0 0
0 5 10 15 20 25 30 35 40 0 5 10 15 20
%Fe 3+ %Ca2+

Fig. 5. Relationship of Cf of scale samples with the % elemental concentration of: (a) Fe3þ and (b) Ca2þ.

4. Conclusion
5

y = 1.1339ln(x) - 6.706 Six “home-made comparison” samples were prepared to deter-

R² = 0.9622 mine the elemental composition of scales generated from the pe-
4 troleum industry since no reference material for scales is available.
The precision and accuracy of the analysis technique were found to
be ca. 90%. Characterisation by EDXRF illustrated that Ba was the
3 most enriched element with 1.4e38.2%; the mean concentrations
CM

of Sr, Fe and Ca reached 3.5, 16.8 and 4.8% respectively. The toxic
elements Pb and Cr were found with mean values of 2.5 and 1.2%,
respectively. A comparison to the reported worldwide data was
2
made. The high content of these elements made scale samples of
hazardous impact to human health and the environment in addi-
tion to its radiological risk.
1 Applying principal component analysis and correlation matrix
0 2000 4000 6000 8000 10000 12000 14000
of Pearson reflected high correlation between Ba, Sr, sample den-
CPS sity, Cf values and radionuclides contents (210Pb and 226Ra). With
reference to XRD analyses, previously reported, the data herein
Fig. 6. Dependence of CM on cps for activity determination at gamma peak energy of
46.5 keV in scale samples.
ascertained that 226Ra and 210Pb co-precipitated with the main
mineralogical component, i.e. Ba0.75Sr0.25SO4.
To optimize the determination of the activity concentration of
15%. Consequently, the mathematical approach would facilitate the
210
Pb in high density scale samples (1.45e3.08) by g-measure-
evaluation of self-attenuation correction factors for 210Pb deter- ments, a mathematical approach was made which improved the
mination in the solid scale samples. This approach may also be relative error to be less than 15%.
preferable to other technical methods for several reasons; no need
for a point source with the interested g-energy lines, quicker for
Acknowledgement
routine analysis (counting the sample once, while counting twice
for the transmission method), less radiation exposure to high ac-
Great appreciation is due to Prof. I. Othman (Director General of
tivity levels and save the life-time of the detector.
AECS) for his invaluable encouragement and funding the project.
The authors would like also to thank Mr. H. Khalili for his assistance
252 L. Al Attar et al. / Environmental Pollution 210 (2016) 246e252
in sample preparation and Mr. H. Issa for XRF analyses. Prof. A. Dyer
is sincerely thanked for reviewing the manuscript.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.envpol.2015.12.007.
References
Al Attar, L., Al Abdullah, J., Doubal, W., Safi, B., Abdul Ghani, B., Khalily, H., 2014.
Evaluation of self-attenuation coefficients of oil produced scales for gamma-ray
spectroscopic analysis. Instrum. Sci. Technol. 42, 562e575.
Al Attar, L., Doubal, W., Al Abdullah, J., Khalily, H., Abdul Ghani, B., Safia, B., 2015a.
Characterisation of NORM solid waste produced from the petroleum industry.
Environ. Technol. 36, 1104e1113.
Al Attar, L., Safia, B., 2013. Sorption of 226Ra from oil effluents onto synthetic cation
exchangers. J. Environ. Manag. 124, 156e164.
Al Attar, L., Safia, B., Abdul Ghani, B., Al Abdullah, J., 2015b. Recovery of NORM from
scales generated by oil extraction. J. Environ. Radioact. http://dx.doi.org/
10.1016/j.jenvrad.2015.12.014, (in press).
Al-Masri, M.S., Aba, A., 2005. Distribution of scales containing NORM in different
oilfields equipment. Appl. Radiat. Isotopes 63, 457e463.
Al-Saleh, F.S., Al-Harshan, G.A., 2008. Measurements of radiation level in petroleum
products and wastes in Riyadh city refinery. J. Environ. Radioact. 99, 1026e1031.
Cutshall, N.H., Larsen, I.L., Olsen, C.R., 1983. Direct analysis of 210Pb in sediment
samples: self-absorption corrections. Nucl. Instrum. Methods Phys. Res. 206,
309e312.
El Afifi, E.M., Awwad, N.S., 2005. Characterization of the TE-NORM waste associated
with oil and natural gas production in Abu Rudeis, Egypt. J. Environ. Radioact.
82, 7e19.
Gazineu, M.H.P., De Araújo, A.A., Brand~ ao, Y.B., Hazin, C.A., Godoy, DeO. J.M., 2005.
Radioactivity concentration in liquid and solid phases of scale and sludge
generated in the petroleum industry. J. Environ. Radioact. 81, 47e54.
Gazineu, M.H.P., Hazin, C.A., 2008. Radium and potassium-40 in solid wastes from
the oil industry. Appl. Radiat. Isotopes 66, 90e94.
Godoy, J.M., Cruz, R.P., 2003. 226Ra and 228Ra in scale and sludge samples and their
correlation with the chemical composition. J. Environ. Radioact. 70, 199e206.
Habib, A.S., Bradley, D.A., Regan, P.H., Shutt, A.L., 2010. The use of MCNP and gamma
spectrometry in supporting the evaluation of NORM in Libyan oil pipeline scale.
Nucl. Instrum. Methods Phys. Res. Sect. A 619, 245e251.
Habib, A.S., Shutt, A.L., Regan, P.H., Matthews, M.C., Alsulaiti, H.A., Bradley, D.A.,
2014. Characterization of naturally occurring radioactive materials in Libyan oil
pipe scale using a germanium detector and Monte Carlo simulation. Radiat.
Phys. Chem. 95, 352e355.
Hamlat, M., Djeffal, S., Kadi, H., 2001. Assessment of radiation exposures from
naturally occurring radioactive materials in the oil and gas industry. Appl.
Radiat. Isotopes 55, 141e146.
Heaton, B., Lambley, J., 1995. TENORM in the oil, gas and mineral industry. Appl.
Radiat. Isotopes 46, 577e581.
Hrichi, H., Baccouche, S., Belgaied, J.E., 2013. Evaluation of radiological impacts of
TENORM in the Tunisian petroleum industry. J. Environ. Radioact. 115, 107e113.
IAEA, 2004. Radiation Protection and the Management of Radioactive Waste in the
Oil and Gas Industry. Safety Reports Series, No. 49. IAEA, Vienna.
IAEA, 2007. Quantitative X-Ray Analysis System (QXAS): User's Manual and Guide
to X-Ray Fluorescence Technique (Version 3.6) accessed on line: 14th July 2015.
www-pub.iaea.org/MTCD/publications/PDF/IAEA-CMS-21_CD_web/PDF/CMS_
21.pdf.
Kitto, M.E., 1991. Determination of photon self-absorption corrections for soil
samples. Appl. Radiat. Isotopes 42, 835e839.
Landsberger, S., Brabec, C., Caion, B., Hashem, J., Lu, C., Millsap, D., George, G., 2013.
Determination of 226Ra, 228Ra and 210Pb in NORM products from oil and gas
exploration: problems in activity underestimation due to the presence of
metals and self-absorption of photons. J. Environ. Radioact. 125, 23e26.
Liland, A., Strand, P., Amundsen, I., Natvig, H., Nilsen, M., Lystad, R., Frogg, K.E., 2012.
Advances in NORM management in Norway and the application of ICRP's 2007
recommendations. Ann. Int. Comm. Radiat. Prot. 41 (3e4), 332e342. http://
dx.doi.org/10.1016/j.icrp.2012.06.021. Elsevier Ltd.
Mansour, N.A., Ahmed, T.S., Fayez-Hassan, M., Hassan, N.M., Gomaa, M.A., Ali, A.,
2012. Measurements of radiation level around the location of NORM in solid
wastes at petroleum companies in Egypt. J. Am. Sci. 8, 252e261. Accessed on
line: 25 August 2015. http://www.jofamericanscience.org/journals/am-sci/
am0806/031_9225am0806_252_260.pdf.
O'Brien, R.S., Melbourne, A.J., Johnston, A., 2011. Development of NORM manage-
ment in Australia. In: Proceedings Series of an International Symposium NORM
VI, 22
26 March 2010 Marrakesh, Morocco, IAEA, Vienna, pp. 163e178.
Obiajunwa, E.I., Pelemo, D.A., Owolabi, S.A., Fasasi, M.K., Johnson-Fatokun, F.O.,
2002. Characterisation of heavy metal pollutants of soils and sediments around
a crude-oil production terminal using EDXRF. Nucl. Instrum. Methods Phys. Res.
Sect. B 194, 61e64.
Read, D., Rabey, B., Black, S., Glasser, F.P., Grigg, C., Street, A., 2004. Implementation
of a strategy for managing radioactive scale in the hinca clay industry. Miner.
Eng. 17, 293e304.
US EPA, 2001. Oil and Gas Production Wastes. Accessed on line: 27th February 2015.
www.epa.gov/radiation/tenorm/oilandgas.html.
Varskog, P., Gellermann, R., 2011. Identification, handling and disposal of norm in
the Norwegian petroleum industry. In: Proceedings Series of an International
Symposium NORM VI, 22
26 March 2010, Marrakesh, IAEA, Vienna,
pp. 387e394.
XLSTAT, 2014. (Addinsoft™) Version 2014.4.07. Accessed on line: 1st April 2014.
http://www.xlstat.com/en/download.html.
Zielinski, R.A., Otton, J.K., 1999. Naturally Occurring Radioactive Materials (NORM)
in Produced Waster and Oil-field Equipment-An Issue for the Energy Industry.
U.S. Geological Survey. Accessed on line: 21st March 2015. www.pubs.usgs.gov/
fs/fs-0142-99/fs-0142-99.pdf.