1. Background and aim
Measurement of the chemical composition of suspended sediment in rivers usually requires the collection of large volumes of river water to obtain sufficient suspended
sediment for analysis.
In this study, we explore the use of direct XRF (X-Ray
Fluorescence) measurement of the element concentrations of suspended sediment extracted by filtration using
membrane filters, which requires a much smaller sample of suspended sediment (typically 5 – 50 mg) and, therefore, a much smaller volume of river water sample (< 2 litres).
Measurement of the chemical composition of suspended sediment using a handheld XRF analyser
MARCEL VAN DER PERK, PASCAL BORN, HANS MIDDELKOOP
Department of Physical Geography, Utrecht University, P.O. Box 80115, 3508 TC Utrecht, The Netherlands; e-mail: m.vanderperk@uu.nl
3. Findings
The XRF analysis yield usable signals for Ca, Cr, Fe, K, Mn,
Pb, Rb, Ti, Sr, and Zn; the signals for As, Ba, Cd, Cu, Co, Mo, Ni, Sb, Sn, and Zr were often or always below the detection limit.
In general, the corrected element concentrations are higher than, but within the range of the element concentrations in suspended Rhine sediment as measured by Rijkswaterstaat and Rhine sediment samples collected using a Phillips time- integrated sediment sampler during autumn 2017 (Fig. 2).
This suggests that the membrane filters are more effective to capture the finer sediment particles than other methods (Phillips sampler or continuous flow centrufuge).
In general, the precision of the analyses expressed as the coefficient of variation varies between 0.13 and 0.22. The Pb analysis is less precise with a coefficient of variation of about 0.4. The majority of the analysis uncertainty can be attributed to the uncertainty in the parameter correction factor function and uncertainty arising from sampling (Fig.
3). The inhomogeneities in sediment composition across the membrane filter only contribute for about 10% to the total uncertainty. The precision of the analyses can be improved by a more precise determination of the power-law correction factor function.
Fig. 1 Correction of element concentration as function of sediment mass on filter
Fig. 2 Mean and range (min, max) of element concentrations
in suspended Rhine sediment collected using mebrane filtering, a Phillips sampler, and measured by Rijkswaterstaat (2016)
Fig. 3 Variation in element concentrations due to uncertainty in the correction factor function, inhomogeneities in sediment composition across the membrane filter, and variation between subsamples
2. Methods
We employed an Olympus Delta-50 Premium handheld 50kV XRF analyser (60 sec tests in 3-Beam Soil Mode) for the
analysis of the suspended sediment samples. The XRF signal was corrected using a correction factor that depends on
the mass of sediment on the membrane filter. To derive the correction factor as a power-law function of sediment mass, two Rhine sediment samples of known composition were
used (Fig. 1). These sediment samples were resuspended in distilled water and different known volumes of the water with resuspended sediment were filtered through 0.45 µm mixed cellulose ester membrane filters. Each filter was analysed
three times to examine the effect of inhomogeneities in sediment composition across the filter.
In addition, to examine the precision of the element
concentrations, 13 replicate water samples were collected
from the Rhine River near Vuren, the Netherlands, in autumn 2017.
0 100 200 300 400 500
Co nc ent ra io n ( ppm )
Sr
Filter samples Phillips samples Rijkswaterstaat samples 2016 0 10000 20000 30000 40000
Co nc ent ra io n ( ppm )
K
Filter samples Phillips samples Rijkswaterstaat samples 2016
0 40 80 120 160 200
Co nc ent ra io n ( ppm )
Rb
Filter samples Phillips samples Rijkswaterstaat samples 2016
0 1000 2000 3000 4000 5000 6000 7000
Co nc ent ra io n ( ppm )
Ti
Filter samples Phillips samples Rijkswaterstaat samples 2016
0 20 40 60 80 100 120 140 160 180 200
Co nc ent ra io n ( ppm )
Pb
Filter samples Phillips samples Rijkswaterstaat samples 2016
0 200 400 600 800
Co nc ent ra io n ( ppm )
Zn
Filter samples Phillips samples Rijkswaterstaat samples 2016
3 replicate measurements on filter uncertainty in the correction factor function (simulated)
0 50 100 150 200 250
0 0.05 0.1 0.15 0.2
Co nc en tra tri on (p pm)
Dry sediment weight (g)
Sr (sediment sample S1)
0 50 100 150 200 250
0 0.05 0.1 0.15 0.2
Co nc en tra tio n ( pp m)
Dry sediment weight (g)
Sr (sediment sample S2)
0 20 40 60 80 100
0 0.05 0.1 0.15 0.2
Co nc en tra tri on (p pm)
Dry sediment weight (g)
Rb (sediment sample S1)
0 20 40 60 80 100
0 0.05 0.1 0.15 0.2
Co nc en tra tio n ( pp m)
Dry sediment weight (g)
Rb (sediment sample S2)
0 10 20 30 40
0 0.05 0.1 0.15 0.2
Co nc en tra tri on (p pm)
Dry sediment weight (g)
Pb (sediment sample S1)
0 10 20 30 40
0 0.05 0.1 0.15 0.2
Co nc en tra tio n ( pp m)
Dry sediment weight (g)
Pb (sediment sample S2)
0 4000 8000 12000 16000 20000
0 0.05 0.1 0.15 0.2
Co nc en tra tri on (p pm)
Dry sediment weight (g)
K (sediment sample S1)
0 4000 8000 12000 16000 20000
0 0.05 0.1 0.15 0.2
Co nc en tra tio n ( pp m)
Dry sediment weight (g)
K (sediment sample S2)
0 700 1400 2100 2800 3500
0 0.05 0.1 0.15 0.2
Co nc en tra tri on (p pm)
Dry sediment weight (g)
Ti (sediment sample S1)
0 700 1400 2100 2800 3500
0 0.05 0.1 0.15 0.2
Co nc en tra tio n ( pp m)
Dry sediment weight (g)
Ti (sediment sample S2)
0 20 40 60 80
0 0.05 0.1 0.15 0.2
Co nc en tra tri on (p pm)
Dry sediment weight (g)
Zn (sediment sample S1)
0 20 40 60 80
0 0.05 0.1 0.15 0.2
Co nc en tra tio n ( pp m)
Dry sediment weight (g)
Zn (sediment sample S2)
0 20000 40000 60000 80000 100000
0 0.05 0.1 0.15 0.2
Co nc ent ra tri on ( ppm )
Dry sediment weight (g)
Element (sediment sample #)
0 100 200 300 400
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Co nc ent ra tio n ( ppm )
Sample #
Sr
0 10000 20000 30000 40000 50000 60000
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Co nc ent ra tio n ( ppm )
Sample #
K
0 50 100 150 200 250 300
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Co nc ent ra tio n ( ppm )
Sample #
Rb
0 2000 4000 6000 8000
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Co nc ent ra tio n ( ppm )
Sample #
Ti
0 100 200 300 400 500 600
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Co nc ent ra tio n ( ppm )
Sample #
Pb
0 200 400 600 800 1000
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Co nc ent ra tio n ( ppm )
Sample #
