Webinar

Investigation of Concrete by
Means of micro-XRF
Dr. Andrew H. Menzies and Dr. Max Buegler
Bruker Nano Analytics, Berlin, Germany
Webinar, June 3rd of 2020

Innovation with Integrity

Introduction
Presenters / Moderators

Andrew Menzies, PhD

Sr. Application Scientist Geology and Mining,
Bruker Nano Analytics, Berlin, Germany

Max J.L. Bügler, PhD

Applications Specialist,
Bruker Nano Analytics, Berlin, Germany

3
Construction Materials - Concrete
Outline

• Brief introduction

• M4 TORNADO

• Micro-XRF

• Construction Materials - Concrete

• Alkali-Silicate- (ASR) and Alkali-Carbonate-Reaction (ACR)

• AMICS (mineralogy)

• SEM-EDX

• Mapping of relevant elements

• Analytical approaches

• Comparison of anodes materials

• Practical demonstration

• Summary

4
 M4 TORNADO
Micro-XRF Spectrometer

30 W micro-focus Rh tube with polycapillary lens

for excitation spot sizes < 20 µm (for Mo-Kα)
Optional 40 W micro-focus W tube with collimator
for excitation of ‘heavy’ elements, embedded in
lighter matrices
Up to two Silicon drift detectors (SDD)
with 30 or 60 mm² active area each energy
resolution < 145 eV
(for Mn-Kα @ 130 kcps throughput)
Sealed sample chamber with adjustable
pressure between 1 mbar and atmospheric
pressure for detecting elements down to Na
Sample stage with measureable area of 200 mm x 160 mm, maximum
sample height 120 mm, maximum sample weight 5 kg, and sample
stage speed up to 100 mm/s, minimum step size 4 µm
5
 M4 TORNADO Instrument Design
Focusing by Poly-Capillary Lens

Polycapillary lenses

< 20 µm for 17,5 keV

~7 mm

SEM image of
Polycapillary lenses

Poly-capillary lens collects

large solid angle of tube radiation
Concentrates beam at sample into small spot
Small spot size guarantees
high brilliance for data collection

@brukerXMA 6
M4 TORNADO Instrument Design
Design and Safety

1 vertical port for

optical microscopes
Micro-focus X-Ray tube
with poly-cap lens and
Be window

Vacuum flange

4 side ports
for excitation Crash protection
sources and
detectors

7 @brukerXMA 7
HyperMap Data Accumulation
Position-Tagged Spectrometry (PTS)

Data mining both on- and offline

What is a HyperMap?
Complete X-ray
spectrum for every
measured pixel saved.

Generates 4D data cube

2D – Spatial Distribution
3D – Spectral
Distribution (E)
4D – Intensity of every
spectral channel
Measurement time/pixel
Data down to 1 ms
compression
during
measurement

@brukerXMA 8
 Introduction
XRF on Different Scales and Fields

X-TRACE
M4 Tornado M6 Jetstream Tracer
(on SEM)
19 mm 25 mm 37.5 cm

38.7 mm

46.5 cm
19 mm

Information from different Trace element

No sample preparation
depths of the sample sensitive
9
 Concrete
Applications

• Concrete has been used as building

material since the Mesopotamian
ages and became very popular with
Greek and Roman builders
• It withstands immense pressure
and has a high durability
• It can be poured into almost any
shape
• If reinforced with steel cable or
mesh as pre- or post-tensioned it
can bridge wide distances

10
 Concrete
Weathering

• Concrete can last thousands of years under

the right conditions
• Nowadays we frequently find images of
cracked and weathered concrete
• As concrete is a mixture of many complex
chemical compounds its durability depends
on environmental conditions
• Saltwater for example leads to chemical
reactions that change the volume and lead to
stress and strain in the structure

11
 Concrete
Alkali–Silica Reaction (ASR)

• Aggregates containing certain forms of silica will react with alkali

hydroxide in concrete to form a gel that swells as it adsorbs water from
the surrounding cement paste or the environment.
• These gels can induce enough expansive pressure to damage concrete.
• The final reaction can be schematically represented by:
Ca(OH)2 + H4SiO4 → Ca2+ + H2SiO42− + 2H2O

Ca2+ + H2SiO42− + 2H2O → CaH2SiO4 • 2H2O

• Which is a reaction that will terminate during maturation of the concrete

as the acidity goes down

12
 Concrete
Alkali–Silica Reaction (ASR)

• In the presence of alkaline hydroxides (NaOH / KOH) the pH level is

significantly increased, and the reaction can continue. In the presence of
such alkaline hydroxides a sodium silicate can form:
2Na(OH) + H4SiO4 → Na2H2SiO4 • 2H2O

Na2H2SiO4 • 2H2O + Ca(OH)2 → CaH2SiO4 • 2H2O + 2NaOH

• The hygroscopic sodium silicate will expand as it absorbs water.

• This expansive pressure in the solid matrix may induce spalling of
material, formation of macroscopic cracks, and a consequently a loss of
structural integrity.
• However, this ASR can be controlled using certain supplementary
cementitious materials.

13
 Concrete
Alkali–Silica Reaction (ASR)

In order of decreasing potential to cause

alkali-silica reaction.
Important Mineral Phases:
• Amorphous silica: volcanic glass, opal,
• Unstable silica polymorphs (tridymite and
cristobalite),
• Chalcedony,
• Other forms of cryptocrystalline silica,
• Deformed quartz Grains of aggregate with different mineral
• Strained quartz composition and morphology:
a) initial condition,
• Quartz with crystallization imperfections, b) with ASR - induced characteristic micro-cracks

• Quartz in crystalline forms

DARIA JÓWIAK-NIEDWIEDZKA ,KAROLINA GIBAS MICHAL A. GLINICKI, 2017, PETROGRAPHIC IDENTIFICATION OF REACTIVE MINERALS
IN DOMESTIC AGGREGATES AND THEIR CLASSIFICATION ACCORDING TO RILEM AND ASTM RECOMMENDATIONS, DOI:
10.7409/rabdim.017.015

14
 Concrete
Alkali–Carbonate Reaction (ACR)

• Dedolomitization, the breaking down of dolomite, is normally associated

with expansion.
• The reaction in this case can be schematically represented by:
CaMg(CO3)2 + 2NaOH → Mg(OH)2 + CaCO3 + Na2CO3
• In another step the sodium-carbonate in combination with calcium
hydroxide (portlandite) can react again into :
Na2CO3 + Ca(OH)2 → CaCO3 + 2NaOH
• Under the presence of calcium hydroxide, the dolomite reacts as follows:
CaMg(CO3)2 + Ca(OH)2 → Mg(OH)2 + 2CaCO3

• This reaction and subsequent crystallization of brucite (Mg(OH)2) may

cause considerable expansion.

15
 Advanced Mineral Identification and
Characterization System (AMICS)

• AMICS software package applies finger printing

to spectra for pixels in order to identify the
minerals in a mapping
Manual analysis of
• The system comes with a preloaded first sample

comprehensive set of typical spectra for more

than 2000 minerals for quick starting a custom
database
Second sample
• For each material system an individual semi-automatically , but
database is developed as samples are tuning necessary

characterized
• Based on the sample the database is extended
by standard spectra created from the identified Automated
identification
regions of a specimen. learns with each
sample of similar
• Statistical tools to calculate mass fractions and type

areal coverage across the specimen

16
AMICS
How does AMICS work?

Micro-XRF Analysis: Hypermap

Mineralogy Workflow Schematic

Classification

• Acquires spectrum for each segment • Identifies minerals based on spectral information
17
 Concrete Hypermap Example
Starting with AMICS

With the preloaded

database the concrete
sample was analyzed
and a according
database of minerals
with according spectra
was defined.

The analysis shows

the spatial distribution
of each of the
identified minerals
along the sample
surface.

18
Concrete Sample Mineralogy
Visualizing Results with AMICS

Resulting modal
composition of the
investigated area.

Relatively large
unknown portion due to
edge area around
grains, that cannot be
properly assigned to
either one phase.

Higher resolution scan

and longer exposure can
reduce the amount of
unknown material.

19
AMICS:
Relevant Mineralogical Information

Modal Mineralogy: How much of each mineral is present.

Elemental Assay: How much of each element is present (Whole Rock).

Elemental Distribution: How is the element of interest (EOI) distributed in

each mineral? E.g. Cr in Chromite vs Pyrope Garnet.

Particle (Grain) Size Distribution: Classify sizes of particles (grains).

Provide P-80, P-50 and P-20 values

Mineral Association: Identify how the minerals are associated with each-
other, e.g. Chromite and Garnet and Clinopyroxene etc.

Grain Shape Factor: The shape of the grain, i.e. euhedral, elongated.

Mineral Density Distribution: Classify densities of minerals. Identify how

the minerals are distributed among the densities.

20
 Academic Research
EDX and X-TRACE

As the spot in micro-XRF is in the range of 15 µm for higher

energies and about 30 µm for elements such as Si, to resolve
the micrometer and sub-micrometer scale an alternative
method is required. The common approach is an e-beam
excitation of the fluorescence in a SEM.

To correlate this information with micro-XRF results, X-

TRACE provides an X-ray excitation on SEM. This allows for
direct correlation of micro-XRF results with the sub-
micrometer features in heterogeneous samples.

This adds complimentary information

at a superior resolution,
its high sensitivity for light elements
down to boron. Where micro XRF
outperforms an EDX for elements with
Z > 20 in detection limits, below that
threshold the EDX has superior
sensitivity and detection limits.

https://www.understanding-
cement.com/alkali-silica.html
21
Analyzing Concrete
Measurement Conditions

• Map of of 56.4 mm by 10.8

mm with 20 µm pixel size
and 10 ms pixel time was
measured on a standard
M4 TORNADO utilizing
2x30 mm² detectors and a
Rhodium anode
• 4 h scan on a region from
left to right (thin green
frame)
• Gradient in Cl distribution
(orange) clearly visible

22
Analyzing Concrete
Enhanced Visualization and Dimensions

• False color representation of the chlorine concentration to enhance

dynamic to the human eye for small changes
• The chemical gradient of Cl is clearly visible by the color changing
accordingly to the element’s fluorescence intensity
• With the drawing tool based on the visualization, a line has been
drawn (shown above) and is 11.5 mm long, which gives an indication
of the depth of Cl diffusion into the concrete
23
 Analyzing Concrete
Cl Penetration Depth from Line Profile

• A line object can be drawn into the Hypermap, creating a line profile along the
sample.
• Additionally, the line was widened to average points perpendicular to the line.
The Cl signal is higher in the first ~ 11.5 mm (from the left).
• The strong modulation is a result of the large agglomerates in the matrix.
24
Analyzing Concrete
… a more Quantitative Approach: Objects

Objects with a width of 2 mm and almost the full height are created next to
each other. Their center position with respect to the sample surface are
determined with the help of the “draw tools”. The individual objects are saved
as new individual maps, using the “save all objects” function.

25
Analyzing Concrete
Phase Analysis and Concentrations

• Load each object

• Select only a signature element of the cement (here P)
• Press auto phase
• Select the cement phase (phases and sum them up if necessary)
• Create sum spectrum of the according (sum-)phase
• Quantify the according spectrum with appropriate method
• Move the spectrum to the project (or save to disk)
• Import (drag and drop) all sum spectra into point mode
• Export results table to Excel → plot

26
Measuring Concrete
Taking a Map and Visualizing Intensity

• Drill core of concrete

exposed to NaCl was
mapped
• The intensity in a
Region-Of-Interest
(ROI) for Cl-K-line
(as proxy for Na) is
displayed
• Intensity is found
virtually everywhere
27
 Measuring Concrete
Signal-to-Noise-Ratio (SNR)

• A look at the spectrum shows a strong overlap of the Cl

lines with backscattered Rh-L-line intensity
• Fast algorithm can perform deconvolution of 18
elements in 350 000 spectra in 50 sec, in order to
separate Cl from Rh intensity
• Rh-anode yields poor SNR due to overlap, suppressing
the Rh-L-line by a filter removes the need for
deconvolution at the cost of overall signal and
particularly for the light elements
Region of Interest (ROI) Deconvoluted signal

28
 Measuring Concrete
Alternative use of Ag-Anode

• To avoid the need for deconvolution, or requirement for a

filter, alternative anode material should be considered
• For the given analytical question silver offers a much
more efficient and potent excitation
• The backscattered Ag-L-line intensity no only does not
overlap with the Cl, but it provides a much more efficient
excitation of Cl by the 2.82 keV absorption edge
• No filter or deconvolution required
ROI of Cl with Rh-anode ROI of Cl with Ag-anode

29
Measuring Concrete
Is there an Ideal Source?

• A direct comparison shows that

the Ag excitation comes at a
cost as well, it overlaps
strongly with K
• As the Ag-L-lines are
transmitted more efficiently, it
provides better signal for
elements with 6 < Z < 18,
such as Cl, S, P, etc.
• It is possible to equip the M4
TORNADO with two sources,
but only one can be equipped
with a poly-capillary lens

From the above, it becomes

clear, there is no ideal anode for
all questions! In the decision for
an excitation source the priority
in terms of elements must be
considered!
30
Concrete
Why XRF? From Construction Site to Lab

• XRF allows for fast measurement of a surface on site using a handheld

XRF, such as the tracer

• On indication of contaminants a drill core can be taken for more in-

depth analysis

• An area of up to 19 cm x 16 cm can be mapped with an M4 TORNADO,

or an area of up to 80 cm x 60 cm utilizing an M6 JETSTREAM

• The penetration of catalysts, such as NaCl, into the concrete can be

visualized and semi-quantitatively analyzed

• For research purposes it can be analyzed even further utilizing EDX in

conjunction with X-TRACE, allowing to correlate micrometer and sub-
µm-scale element distributions

31
Live Demonstration
Phase Analysis and Concentrations

32
Concrete
Summary

• Micro-XRF can provide compositional and mineralogical information across a

range of scales

• Fundamental parameter (FP) based quantification allows for quick assessment

of composition in points

• Mapping of mineralogy using AMICS yields information about possible ASR and
ACR reactions as well as matrix and aggregate compositions

• Mapping of element intensities on a drill core sample can provide visualization

of alkaline hydroxide penetration into the volume

• Smart analysis of mapping allows for semi-quantitative analysis of Cl and K

concentration in the matrix

• Phase analysis allows for quick estimation of cement to aggregate ratio as well
as for semi-quantitative analysis

33
 Literature
M4 TORNADO in Concrete Research

Since in 2015 the first M4 was purchased for the purpose

of research in concrete more than 130 publications can be
found, and multiple Masters and PhD thesis utilizing the M4
in the field were published. Some example from various groups:
G. Bonifazi et al., "Hyperspectral imaging applied to the identification and classification of
asbestos fibers," 2015 IEEE SENSORS, Busan, 2015, pp. 1-4, doi: 10.1109/ICSENS.2015.7370458.
K. De Weerdt et al., “Comparing chloride ingress from seawater and NaCl solution in Portland
cement mortar,” Cement and Concrete Research, Volume 115, 2019, pp. 80-89, doi:
10.1016/j.cemconres.2018.09.014.

H. Morillas et al., “Multianalytical approach to evaluate deterioration products on cement used as

consolidant on lithic material: The case of Tello Obelisk, Lima (Peru),” Microchemical Journal, Volume
139, 2018, pp. 42-49, doi: 10.1016/j.microc.2018.02.017.

G. Plusquellec et al., “Determining the free alkali metal content in concrete – Case study of an
ASR-affected dam,” Cement and Concrete Research, Volume 105, 2018, Pages 111-125, doi:
10.1016/j.cemconres.2018.01.003.

A. Dehghan et al., “Application of X-ray microfluorescence for the determination of chloride

diffusion coefficients in concrete chloride penetration experiments,” Construction and Building
Materials, Volume 148, 2017, pp. 85-95, doi: 10.1016/j.conbuildmat.2017.05.072.

T. Danner et al., “Long-term Influence of Concrete Surface and Crack Orientation on Self-healing
and Ingress in Cracks – Field Observations,” Nordic Concrete Research, 58(1), 1-16. doi:
https://doi.org/10.2478/ncr-2018-0001

34
Webinar
Q&A

Questions, Thoughts or Comments?

If you have questions or want to contact us during

the Webinar, please type your questions, thoughts,
or comments in the Q&A box and press Submit.

We ask for your understanding if we do not have time to discuss all comments
and questions within the session.

Any unanswered questions or comments will be answered and discussed by

e-mail or in another Webex session.

35
More Information

For more information, please contact us:

Bruker Nano GmbH
info.bna@bruker.com

Or
Max.Buegler@bruker.com

Andrew.Menzies@bruker.com

36
Innovation with Integrity