Novel spectroscopies for metallome analysis of trace element:

Hyperaccumulator plants

PhD thesis by

Imam Purwadi

ResearchGate Logo

Thursday, 19 October 2023

WHOIAM?Click on markers

Enschede, Netherlands

I spent almost two years in Enschede, Netherlands, where I completed my master's degree at the ITC Faculty of Geo-Information Science and Earth Observation, University of Twente. The title of my Master's thesis was "The feasibility of targeting REEs in tailings using satellite remote sensing data: A case study of abandoned mine sites in Bangka Island, Indonesia." This research was a continuation of my undergrad project.

South Sumatra, Indonesia

More than half of my life has been spent in South Sumatra, Indonesia. I did my undergrad at the local public university, and my undergrad thesis project was about using geostatistical analysis for mapping soil pH distribution in one of the mine rehabilitation projects in Bangka Island, Indonesia.

Brisbane, Australia

Currently, I am pursuing my PhD at the Sustainable Mineral Institute, University of Queensland, Brisbane, Australia. My research is focused on developing methods to identify metal(loid)s hyperaccumulating plants that can be used for phytomining, phytoextraction, and phytoremediation.

  • WHYIDO?this research...

  • 2012

    My first visit to the island.

    I observed mining activities but stumbled upon dead corals in metal drum barrels :(

    Did you know tin has been mined there since the early 18th century?

    It wasn't until 2009 that proper mine rehabilitation regulations were issued.

    Image Description

  • 2014

    It was a turning point for me.

    I wrote my first paper about the island, and it received recognition from The Association of Indonesian Mining Professionals.

    This recognition confirmed and validated my decision to delve into the environmental aspects of Bangka Island.

    By that time, tin resources had been depleted, but rumors spoke of precious metals in the tin tailings.

    I couldn't help but wonder, where were these precious metals, and could we remine them while ensuring proper rehabilitation?

  • 2018

    Exciting results! here and here

    It turns out the tailings do contain precious metals!

    But the question arose: Could we mine them using traditional methods without deteriorating the environment?

    I explored more environmentally friendly options, like agromining.

    However, there was a challenge - no identified plants to do the job!

Yosemite National Park

Tailings:
99% quartz
1% Rare Earth Elements

2017, one of many tin tailings in the island

What's next?

Find hyperaccumulator plants!
Where to start? How to achieve?

July 2019, PhD commenced

Chapter 1:
Tools for the Discovery of Hyperaccumulator Plant Species in the Field and in the Herbarium

published in Agromining: Farming for Metals, 2nd edition, 2021

Literature review

Reasons to find hyperaccumulators

Benefits to find hyperaccumulators

Hyperaccumulators

Hyperaccumulator coined for Pycnandra acuminata (Jaffré et al. 1976)

721 identified hyperaccumulators (Reeves et al. 2017)


How?

74% of 721 identified hyperaccumulators are Nickel hyperaccumulator

An easily prepared and deployed test for Ni hyperaccumulator detection exists!

Other reasons

Tool is not the only reasons for the high number of identified Ni hyperaccumulator plants.

Soil concentrations

Wide spread nickel rich soils as the weathering product of ultramafic rocks

Economic value

Nickel is in high demand, while supply is scarce

wirestock (Feepik)

"As metal prices rise with increased demand, hyperaccumulators are gaining recognition as an alternative means of extracting metals, and so is research in this field."

While rare earth hyperaccumulators were discovered earlier than nickel hyperaccumulators, research progress in the former lags behind that of the latter

source

A new approach was proposed to expedite the identification of hyperaccumulators

Using a portable X-ray fluorescence instrument to scan herbarium specimens: Rapid analysis, Non destructive test, Abundance Sample

Image Description
Image Description

Source: X-Ray Fluorescence Ionomics of Herbarium Collections

How does a portable X-ray fluorescence instrument work?

The instrument shoots X-rays to hit the sample's atoms, and detectors catch any X-rays that come out

The outgoing X-rays from each element are distinct and can be utilized for both quantitative and qualitative analysis

The outgoing X-rays from each element are distinct and can be utilized for both quantitative and qualitative analysis

Each element spectrum was estimated using GeoPIXE. Click on legend to hide or show

Source

Portable XRF Instruments

Caveats

+ - +

Rare earth XRF peaks often observed but the instrument algorithm failed to report
Why it is important to understand the peak XRF radiation for each element?

XRF is a bulk analysis method that captures not only XRF radiation emitted from the surface of a sample but also includes some XRF radiation originating from beneath the surface that manages to reach the detectors

The depth of penetration: how far the X-ray from the instrument can penetrate the sample, and the escape depth: how far the XRF originated from an atom inside the sample can travel

The thing is...

Most of the built-in algorithm assumes the sample we prepared is thicker than the escape depth of elements we are interested in...

Red dots are synthetic samples with about same concentration but varying in thickness

The white lines are an empirical concentration reported by the instrument if the thickness is not corrected

Chapter 3:
X-ray fluorescence spectroscopy for metallome analysis of herbarium specimens

published in Plant Methods, 2022

Method development

XRF Method

For herbarium XRF scanning

  • Built-in algorithm

    Existing method

    1. 1. Already prepared by the manufacture
    2. 2. Easy to use
    3. 3. Mostly designed for rocks or soils
    4. 4. Assume samples thick enough
    5. 5. Not suitable for plants

  • Empirical calibration

    Existing method

    1. 1. Fast and easy to prepare
    2. 2. More accurate than manufacturer algorithm
    3. 3. Sample matrix expected to be the same as the standard
    4. 4. Only applicable for elements what they are prepared

  • Dynamic analysis

    Proposed method

    1. 1. Scientifically reviewed and mature algorithms
    2. 2. Used to process synchrotron XRF data
    3. 3. Fully controlled quantification
    4. 4. Once calibrated, it can be used for any elements

What is Dynamic analysis?

To solve a complex physics equation expressing a theoretical relationship between fluorescence peak intensities and the concentration of elements by providing all parameters required by the equation

In case of monochromatic beam excitation (with energy E0), the relation between the intensity of characteristic X-rays of element i (with energy Ei) and weight fraction of this element is:

Ii(Ei) = [Gε(Ei)ai(E0)I0(E0)/sinα].{1-exp[-ρd(μ(E0)cscα+μ(Ei)cscβ)]}/[μ(E0)cscα + μ(Ei)cscβ]
With: ai = Wiτi(E0)ωipi(1 - 1/ji)
Where
                Ii(Ei) – the intensity of the fluorescent radiation of the ith element,
                G – the geometry factor,
                ε(Ei) – the intrinsic detector efficiency for recording a photon of energy Ei,
                I0(E0) – the number of incident photons of energy E0 per second per steradian,
                α and β – the effective incidence and takeoff angles, respectively,
                ρ – the density of the specimen in g/cm3,
                d – the sample thickness in cm,
                μ(E0) and μ(Ei) – the total mass attenuation coefficients in cm2/g at energies E0 and Ei, respectively,
                Wi – the weight fraction of the ith element,
                τi(E0) – the total photoelectric mass absorption coefficient for the ith element at the energy E0 in cm2/g,
                ωi – the fluorescence yield of the element i,
                pi – the transition probability of the kth line of the element i,
                ji – the absorption jump at the K-edge of photoelectric absorption in ith element.
But, most, if not all of the available portable XRF instrument in the markets is polychromatic, how to solve it?

GeoPIXE

A software package developed by CSIRO using Dynamic Analysis for quantitative SXRF data

Image Description

  • Input

    Sample parameters
    Instrument parameters

  • Process

    GeoPIXE

  • Output

    Concentration 
Sample parameters:
- Density 
- Thickness 
- Chemical composition/formula
Instrument parameters
- X-ray source including: current, angle in, angle out, anode spot, filter thickness, filter density, filter formula 
- Detector including material, diameter, distance to window, area, solid angle, resolution, tilt angle, shape, array, absorber layers

Sample parameters

  • 0.9 g/cm3

    Density

    less than water
  • C6H10O5

    Formula

    Cellulose
  • ?

    Thickness

    Unknown/to be estimated
Image Description Source

Thickness?

An illustration of herbarium XRF scanning: Observe titanium plate beneath herbarium specimen

Image Description

Safety: The X-ray coming out of the instrument is not fully absorbed herbarium and even the desk. Put metal plates under specimen to absorb the excessive x-ray.

Safety: A portion of the X-ray coming out of the instrument is scattered. Put backscatter shield on the instrument to absorb backscatter radiation.

Remember: The XRF of Ti from Ti metal can travel ~2mm in dry leaves

Image Description

According to Rafał Sitko and Beata Z, 2011, emission-transmission can be used for determining matrix properties (μm), without the knowledge of the sample composition. 

(Its - Is)/It = exp[-μm]
Where:
It: the Ti intensities from the Ti plate alone
Is: the Ti intensities from the sample
Its: the Ti intensities the sample on top of the Ti plate
Ti concentration in leaves < 34 μg/g or even less (Cary and Kubota 1990; Tlustoš et al. 2011), thus not producing significant Ti fluorescence. So, equation can be simplified to:
Its = It exp[-μm]

Thickness?

Relationship between sample area density and transmitted Ti signals

Image Description

GeoPIXE requirements

  • Input

    ✓ Sample parameters
    ? Instrument parameters

  • Process

    GeoPIXE

  • Output

    Concentration

Reverse engineering

Use certified materials during calibration to reverse engineering the instrument

  • Input

    ✓ Certified reference materials
    ? Instrument parameters

  • Process

    GeoPIXE

  • Output

    ✓ Concentration

Instrument parameters

Part 1: Source

Image Description

Instrument parameters

Part 2: Filters

Image Description

Instrument parameters

Part 3: Detector

Image Description

Illustration

A simple fitting one sample: load spectrum, load calculate parameter (yield), do fitting, and open next spectrum

Source parameters x filter parameters x detector parameters = millions iterations

Leasons from reverse engineering

  1. Better than empirical calibration and built-in algorithm
  2. Exhausting and time consuming: it took 1 year to get good parameters (error < 5%)
  3. Prone to errors, if done manually
  4. work smart not hard; use pyautogui to automate typing dan clicking.

Chapter 4:
A systematic assessment of the metallome of selected plant families in the Queensland (Australia) flora using X-ray fluorescence spectroscopy

published in Australian Journal of Botany, 2023

The application of the developed method

Hyperaccumulators in Australia/Queensland

  • Few hyperaccumulator

    <10 of 721

  • Metal rich soil

    Many metal deposits

  • Existing data

    > 2000 specimens were scanned

Can we reveal any missed hyperaccumulators from the previous studies with the developed method?

Results

Newly identified hyperaccumulators by the developed methods

  • Manganese 15
  • Nickel 2
  • Cobalt 3
  • Zinc 3
  • Rare Earth 2
  • Selenium 1

The two new REE hyperaccumulators were further confirmed in another study by taking new samples from the field, subsequently measured using ICP-AES. The two are discussed a little bit here: link

Spatial distribution of Hyperaccumulators

  • Icon
    Mn Layer
  • Icon
    Co Layer
  • Icon
    Ni Layer
  • Icon
    Zn Layer

Chapter 5:
Recognition of trace element hyperaccumulation based on empirical datasets derived from XRF scanning of herbarium specimens

published in Plant and Soil, 2023

The application of the developed method

Hyperaccumulators

Plants exhibiting concentrations of at least an order of magnitude higher than that found in normal plants

  • Manganese 10000µg/g
  • Cobalt 300µg/g
  • Nickel 1000µg/g
  • Copper 300µg/g
  • Zinc 3000µg/g
  • Rare Earth 1000µg/g
  • Arsenic 1000µg/g
  • Selenium 100µg/g

Source: Baker & Brooks, 1989; Reeves, 2003b; van der Ent et al., 2013

Assessing hyperaccumulator thresholds

Using XRF and ICP data

  • XRF Data

    • 26942

      Total herbarium specimens

    • 1150

      New datasets

    • From four countries

  • ICP-AES Data

    • 1710

      Total field samples

    • From one country

N < Below detection limits

XRF has a high detection limit

Element XRF ICP-AES
Manganese 12517 1
Cobalt 26516 115
Nickel 24684 1
Zinc 23346 293
Arsenic 26865 Not available
Selenium 26861 Not available
Yttrium 26837 Not available

How to deal with below detection limit values?

Regression on Order Statistics: Lee and Helsel (2005), Lee and Helsel (2005), Dennis R. Helsel and Timothy A. Cohn (1988)

Image Description

Regression on Order Statistics vs Constant value

Regression on Order Statistics vs Constant value

Determining the threshold between Normal and Hyperaccumulator plants

Image Description

Hyperaccumulator thresholds in μg/g

Element Historical XRF ICP-AES
Manganese 10000 1210 2850
Cobalt 300 32 5
Nickel 1000 280 694
Zinc 3000 181 7
Arsenic 1000 8 Not available
Selenium 100 10 Not available
Yttrium Not available 11 Not available

The historical hyperaccumulator thresholds are higher than this study results, so we suggested to not change the historical results because higher values mean safe from false positive

Chapter 6:
Portable X-ray fluorescence (XRF) spectroscopy for intact dry leaves

under revision submitted to Ecological Research, 2023

The application of the developed method

How good is the developed method for different instruments?

Image Description

3 different instruments

114 leaves

3 different algorithms

The XRF spectra of the three instruments

Mean absolute errors to the highest errors [relative percentage error to highest errors]

Intrument Algorithm Manganese Iron Cobalt Nickel Copper Zinc
Rocksand Empirical 675.1 [5.7%] 375.4 [32.5%] 36.8 [1.4%] 1484.3 [3.2%] 3.1 [1.7%] 103.5 [7.7%]
independent pipeline 500.3 [4.2%] 270 [23.3%] 25.4 [0.9%] 636.5 [1.4%] 3.4 [1.8%] 65 [4.9%]
Manufacturer 574.5 [4.9%] 288.5 [24.9%] 73.5 [2.7%] 2930.6 [6.3%] 8.7 [4.6%] 131.5 [9.8%]
Goldd+ Empirical 395.6 [3.4%] 375.4 [32.5%] 36.4 [1.3%] 1181.9 [2.5%] 3.1 [1.7%] 81.9 [6.1%]
Independent pipeline 497.4 [4.2%] 268.9 [23.2%] 62.3 [2.3%] 707 [1.5%] 2.8 [1.5%] 132.3 [9.9%]
Manufacturer 11776.4 [100%] 1156.5 [100%] 2719.7 [100%] 46454.2 [100%] 188.1 [100%] 1338.3 [100%]
Tracer 5g Empirical 415.5 [3.5%] 376.6 [32.6%] 54.7 [2%] 1018.2 [2.2%] 3 [1.6%] 81.9 [6.1%]
Independent pipeline 276.9 [2.4%] 266.6 [23.1%] 75.6 [2.8%] 711.2 [1.5%] 5.9 [3.1%] 59.9 [4.5%]

Chapter 7:
Reflectance spectroscopy as a promising tool for 'sensing' metals in hyperaccumulator plants

published in Planta, 2023

Exploring other possibilities

Benefits of remote sensing technique compared to herbarium XRF

1. No X-ray radiation license needed

2. Applicable from individual plant species to landscape-scale

3. Capable of scanning inaccessible areas

Image Description

How a leaf reflects light: A visual analysis?

When sunlight shines on a leaf, a portion of this light is reflected.

Metal absorbance bands

Nickel Hyperaccumulator Leaves

  • Berkheya coddii 69leaves
  • Glochidion bambangan 34leaves
  • Glochidion panataran 34leaves
  • Phyllanthus rufuschaneyi 35
  • Rinorea bengalensis 32leaves
  • Rinorea javanica 24leaves
  • Actephila alanbakeri 32leaves
  • Walsura pinnata 26leaves

Nickel concentration and Spectral reflectance

Image Description

Mean spectral reflctance of Hyperaccumulator leaves per species

Reflectance vs Concentration

Image Description

Nickel concentration and Spectral reflectance

Image Description

Thank you!

Chariman: Dr Nathan Fox Examiner: Prof. Hudson W. P. de Carvalho Examiner: Dr Mark J Hackett
Supervisors:
Dr Antony van der Ent
Prof Peter D. Erskine
Dr Lachlan Casey
Laboratory Team:
Mr Vinod Nath
Ms Vanessa Glenn
Ms Natasha Ufer
Collaborators :
Prof Guillaume Echevarria * Dr Wojciech J Przybyłowicz * Dr Jolanta Mesjasz-Przybyłowicz * Dr Chris G. Ryan * Dr Gillian Kim Brown
The family of the Centre for Mined Land Rehabilitation at SMI :
Dr Philip Nti Nkrumah * Dr Adrian Paul * Dr Farida Abubakari * Dr Roger Tang * Dr Amelia Corzo Remigio * Dr Maggie-Anne Harvey * Dr Vidiro Gei * Katherine Pinto Irish
All those who have been part of this journey. Your collective support and encouragement have been invaluable to me, and I am deeply grateful for each and every one of you
These slides are built using multiple open-source data and code: WebSlides for the presentation template, LeafletJS for maps, Chart.js for charts, Font Awesome for icons, and Freepik for images.