Photothermal Spectroscopy for Planetary Sciences: A

Characterization of Planetary Materials in the Mid-IR

Christopher Coxa,b , Jakob Haynesa , Christopher Duffeya,b , Myles C. Hoskinsonb ,
arXiv:2411.13759v1 [astro-ph.EP] 21 Nov 2024

Christopher Bennettb , Julie Brisseta

a
Florida Space Institute, University of Central Florida, 12354 Research Pkwy, Orlando
FL-32826, USA
b
Department of Physics, University of Central Florida, 4111 Libra Dr, Orlando FL-32816, USA

Abstract
Understanding of the formation and evolution of the Solar System requires un-
derstanding key and common materials found on and in planetary bodies. Mineral
mixing and its implications on planetary body formation is a topic of high inter-
est to the planetary science community. Previous work establishes a case for the
use of Optical PhotoThermal InfraRed (O-PTIR) in planetary science and intro-
duces and demonstrates the technique’s capability to study planetary materials. In
this paper, we performed a measurement campaign on granular materials relevant
to planetary science, such as minerals found in lunar and martian soils. These
laboratory measurements serve to start a database of O-PTIR measurements. We
also present FTIR absorption measurements of the materials we observed in O-
PTIR for comparison purposes. We find that the O-PTIR technique suffers from
granular orientation effects similar to other IR techniques, but in most cases, is is
directly comparable to commonly used absorption spectroscopy techniques. We
conclude that O-PTIR would be an excellent tool for the purpose of planetary ma-
terial identification during in-situ investigations on regolith and bedrock surfaces.

1. Introduction
O-PTIR, or Optical-PhotoThermal InfraRed, is a relatively new form of spec-
troscopy (Spadea et al., 2021). It is a pump-probe technique that utilizes an IR
laser (pumping laser) and a visible laser (probing laser) (Bazin et al., 2022; Paulus
et al., 2022). The IR laser creates a photothermal effect causing thermal expan-
sion at the surface of a sample and causes a change in the index of refraction of the
sample (Olson et al., 2020). This method is described in detail in Cox et al. (2024),

Preprint submitted to Planetary and Space Science November 22, 2024

Bazin et al. (2022), Krafft (2022), and Spadea et al. (2021) among many others.
Cox et al. (2024) makes a case for the use of O-PTIR technology in the future
of planetary sciences. Cox et al. (2024) demonstrates the fascinating capabili-
ties of the O-PTIR technique. Particularly, the case is made for fast measurement
speeds while maintaining high spatial and spectral resolution. Additionally, they
demonstrate the lack of need for sample preparation and outline the benefits for
this technique in planetary science missions.
The present work will focus on the use of O-PTIR to study minerals relevant to
planetary science. Through analysis of returned Apollo samples and x-ray diffrac-
tion performed by the Curiosity rover on Mars, we know that materials exist on
the surfaces of the Moon and Mars. From these studies, we also know the materi-
als to be granular. For this reason, we decided to study specific minerals found on
the Moon and Mars, looking at granular samples in particular.
Detailed mineral composition analysis of regolith grains requires high spatial
resolution, ideally at the micron or sub-micron level. To achieve this level of
resolution, in-situ measurements or laboratory analysis (via a sample return mis-
sion) are usually required. Instrumentation involving Raman (e.g. SHERLOC
on Mars Perseverance Rover, Bhartia et al. (2021), Mass Spectrometry (Cassini’s
INMS, Waite et al. (2004)), and Scanning Electron Microscopes (SEM) (e.g. re-
turned Stardust Samples, Stroud et al. (2014)) can produce a spatial resolutions
in the micron or even sub-micron range. However, sometimes these methods are
destructive to the sample or may require additional information to identify a ma-
terial.
Mineral analysis, especially in situ, yields information regarding the evolu-
tion of our Solar System as well as individual planetary bodies (Andersen and
Haack, 2005; Gomes and Silva, 2007). This includes insight into how materials
segregated over time to their current locations (Cameron, 1988). Measurements
with sub-micron resolution allow for the ability to distinguish between differing
evolutionary histories (Ehrenfreund and Charnley, 2000; Tice et al., 2022).
Given this paper’s focus on mineralogical study, the wavenumber range used
here is 980 - 1800 cm−1 . This is primarily due to the key features found there that
help unambiguously identify minerals and compounds. Hu et al. (2016) defines
the ”fingerprint region” to be a region that contains many unique and complex
absorptions. It is typically accepted that it lies in the 500 - 1500 cm−1 (20.0-6.67
µm) range. Specifically, there are key features in the 805 - 1300 cm−1 (12.4-7.7
µm) range. An additional limitation is the range of the instrument used.
In this work, we focus on mid-IR wavelengths in order to detect the features
that exist within that range for materials relevant to planetary science. Miner-

2
als, such as olivine, basalt, anorthosite, etc., are often well characterized in IR
wavelengths and specifically in the wavelengths of interest for this work, as ev-
idenced by the existence of multiple openly available databases. These include
Infrared and Raman Users Group (IRUG)1 , SpectraBase2 , and Wiley’s Knowitall
database3 among many others. Databases such as these are sometimes open access
and they possess entries for materials such as the ones analyzed in this work. Ad-
ditionally, several minerals are discussed and described in The Infrared Spectra of
Minerals (Farmer, 1974). The textbook and each above database include data that
overlaps with our studied wavelength range, which allows us to include discus-
sions from the textbook into the results discussions. Further, we performed FTIR
absorbance (FTIR-A) measurements and utilized Wiley’s Knowitall database for
direct comparisons of IR absorption measurements to O-PTIR measurements of
our samples.
In this paper, we will describe how measurements were taken and analysis
of the data produced in Section 2. We will present our individual measurements
by material in Section 3. In Section 4 we will discuss the ability of O-PTIR to
identify and study planetary materials. Finally, in Section 5, we will summarize
our work and conclude the paper.

2. Methodology
This measurement campaign, using O-PTIR technology, produced mid-IR
data for several materials. In this section, we will describe the samples stud-
ied, how the measurements were performed, and how the data produced from
those measurements were analyzed. The FTIR-A data collection process is also
described here.

2.1. Samples
Samples in this study are materials relevant to planetary sciences. Below, Ta-
ble 1 lists each material that was characterized in this work. The most commonly
accepted chemical formula is listed next to the material in the table.
Samples were prepared in the same manner as Cox et al. (2024).

1
http://www.irug.org/search-spectral-database
2
https://spectrabase.com/
3
https://sciencesolutions.wiley.com/knowitall-analytical-edition-software/

3
Table 1: Chemical formulas of each material described and analyzed in this work. Anorthosite’s
and Basalt’s chemical formulas are left out since they are rocks and not a specific mineral.
Mineral Chemical Formula
Anorthosite –
Basalt –
Bronzite (M g, F e)2 Si2 O6
Fe-Carbonate (Siderite) F eCO3
Ferrihydrite F e10 O14 (OH)2
Gypsum CaSO4 · 2H2 0
Hematite F e2 O3
Hydrated Silica H10 O3 Si
Ilmenite (F e, T i)2 O3
Magnetite F e3 O4
Mg-Carbonate (Magnesite) M gCO3
Mg-Sulfate (Epsomite) M gSO4
Olivine (M g, F e)2 SiO4
Smectite Al2 H2 O6 Si

2.2. O-PTIR Measurements

The calibration and measurement processes are similar to the ones described
in Cox et al. (2024). Here, hyperspectral maps were chosen as the method for data
collection. This was decided to help reduce the significance of granular orientation
effects. The significance of these effects will be demonstrated in Section 3, where
the granular orientation effects for each material will be shown. A hyperspectral
map spectra is the average of all hyperspectral points in a given map. Some data
presented will be normalized. A normalization factor means every point in the
spectrum was multiplied by the factor.

2.2.1. Hyperspectral Maps

In Cox et al. (2024), it is shown that different materials produced various
SNRs. With this knowledge, we elected to measure each material’s SNR indi-

4
vidually. To do this, we measured each material using single hyperspectral point
measurements. One single point measurement is a single spectrum comprised of
three scans averaged together by the software of the instrument. Ten such single
hyperspectral point measurements were obtained and averaged together to pro-
duce a spectrum with error bars. The error bars here are a standard deviation of
the ten single hyperspectral point measurements. The spectrum with standard de-
viation error bars for each material is shown in the right panel of each subsection
figure in Section 3.

Figure 1: Mid-IR O-PTIR measurements of ilmenite. The solid black line is the measurement
discussed in Section 3.9. The red line is a larger hyperspectral map of the same sample. ”Larger”
here means a larger measurement area with a similar spacing of measurements. For the smaller
hyperspectral map (102 x 76 µm, 2 µm spacing, 2028 individual spectra), instrument settings
were: 77% IR Power, 3.5% Probe Power, and 20x Detector Gain. The spectrum was normalized
by a factor of 1.08 to be more visibly comparable to the larger map. For the larger hyperspectral
map (635 x 475 µm, 5 µm spacing, 12288 individual spectra), instrument settings were: 60% IR
Power, 3.5% Probe Power, and 20x Detector Gain. The spectrum was normalized by a factor of
100 to be more visibly comparable to the smaller hyperspectral map.

Figure 1 shows the mIRage® instrument measurements of two separate hyper-

spectral measurements of the same sample of ilmenite.

5
Figure 2: The same Mid-IR O-PTIR measurements of ilmenite shown in Figure 1. This plot
removes the normalization. The plot is to serve as a visual aid demonstrating the difference in the
measurements.

The difference between the measurements featured in both Figures 1 and 2 is

the size of the sample probed. The larger map shows a peak at 1016 wavenumbers
which is not seen in the spectrum of the smaller hyperspectral map. However,
the peaks at 1033, 1448, 1468, and 1481 wavenumbers are all seen in both the
spectrum produced by the small map and the spectrum produced by the large map.
Figure 1 shows the two measurements share very similar spectral peaks and the
smaller map spectrum contains most features seen in the larger measurement. The
peak of the spectrum of the larger hyperspectral map was much less intense (as
seen in Figure 2). This is likely due to the increased null space measured between
grains. These similarities and the reduction of null space gave us the confidence
to use smaller areas to make hyperspectral maps for the purpose of this work.

6
2.2.2. Granular Orientation
From Cox et al. (2024), we hypothesize that some of the variation is caused
by granular orientation effects. This issue is seen in other more traditional IR
methods (e.g. Serratosa and Bradley, 1958; Shuai and Yang, 2017). In order to
study the effect of grain orientation on the spectra we collect, we performed single
hyperspectral point measurements (described in Section 2.2.1) on several different
grains in a heterogeneously mixed sample. The found range of spectra from each
material is shown in the left panel of each subsection figure in Section 3.

2.3. FTIR Measurements

As stated above, we performed FTIR-A measurements in order to directly
compare our O-PTIR data to FTIR data. To collect FTIR measurements, a sim-
ilar method to Cox et al. (2024) is used. A pellet, prepared as described in Ap-
pendix Appendix A, is placed into an iS50R FTIR instrument in a way that the
sample is held in front of the beam. The OMNIC software was used to collect the
spectra. A pure KBr pellet is placed inside and allowed to purge for a minimum of
12 hours and then scanned as a background reference. We had collection settings
set to a gain of 2.0 and an aperture of 87. The measurements were 128 scans with
a resolution of 4. Three measurements were taken and their average is reported
here.
After background is collected, the pure KBr is removed and replaced with a
sample pellet. The mixture consists of the target sample and KBr. Each sample
measurement has the background subtracted from it. This subtraction is done by
the software. These measurements are used to compare an FTIR measurement
with an O-PTIR measurement of the same sample (red curve in the right panel of
subsection figures in Section 3.

2.4. Data Analysis

To analyze the data, we made hyperspectral maps (described in Section 2.2)
and averaged the spectral points to produce a single spectrum. Peaks were iden-
tified from this curve using the system’s software. Thresholds were set to distin-
guish peaks from noise in the measurement. Each material’s hyperspectral map
measurement along with error bars will be presented in their respective sections.
The error bars for these plots are the standard error of mean. In a separate plot,
the averaged spectrum from the hyperspectral map were normalized to one and
compared to absorbance data. The normalization was done by finding the most
intense peak and multiplying the entire spectrum by the value required to make
the max intensity equal to one.

7
3. Results
Using the measurement methods described in Section 2, we produced mid-IR
measurements for several materials relevant to planetary sciences (see a list in
Table 2). Spectra are presented as normalized spectra for an easier visual com-
parison. The ”normalization factor” listed in Table 2 indicates a spectrum was
multiplied by that number to normalize the spectrum to one, which summarizes
the run information for each measurement discussed in this work.

8
 Table 2: Run information for each material analyzed in this work. Each subsection will have this information presented, but it is sum-
marized in this table for ease of reference. ”Peak Height” refers to the maximum unnormalized O-PTIR peak measured in the sample,
”Normalization Factor” is what each spectrum value was multiplied by to normalize to one for visual comparison purposes, ”IR Power”
and ”Probe Power” indicate the power setting of the respective laser, ”HS Map Dimensions” refer to the size in microns of the hyper-
spectral map made of the sample, ”Spacing” refers to the distance in microns between each measurement that composed the map, ”Total
Spectra” is the total number of measurements in a given map.
Material Peak Height Normalization Factor Max Error IR Power (%) Probe Power (%) Gain (x) HS Map Dimensions (µm) Spacing (µm) Total Spectra
Anorthosite 1.113194865 0.898315319 0.07479 46 3.5 10 612 x 468 12 2080
Basalt 0.141632355 7.060533591 0.02635 21 3.5 20 225 x 210 5 1978
Bronzite 13.29053313 0.075241526 1.72984 5 3.5 10 78 x 78 2 1600
9

Fe-Carbonate (Siderite) 39.08304398 0.025586543 1.08248 10 0.43 50 120 x 84 3 1189

Ferrihydrite 0.015424165 64.83333133 0.00408 60 2 20 588 x 432 12 1850
Gypsum 0.033333541 29.9998131 0.00956 60 2 10 576 x 408 12 1715
Hematite 0.147469541 6.781061318 0.01356 46 2 20 600 x 444 12 1938
Hydrated Silica 15.14671771 0.066020904 0.77476 21 3.5 10 612 x 444 12 1976
Ilmenite 0.938251266 1.065812577 0.03073 77 3.5 20 102 x 76 2 2028
Magnetite 0.805184896 1.241950768 0.04704 46 2 20 588 x 432 12 1850
Mg-Carbonate (Magnesite) 2.506938405 0.398892928 0.12751 21 2 10 612 x 444 12 1976
Mg-Sulfate (Epsomite) 0.105215506 9.504302531 0.00912 21 3.5 20 564 x 432 12 1776
Olivine 3.062116513 0.326571506 0.17533 77 2 20 142 x 70 2 2592
Smectite 10.86572062 0.092032552 0.57973 21 3.5 5 540 x 380 10 2145
 In this section, we discuss the results of each material listed (in alphabetical
order). The findings are described and displayed in each subsection, but a sum-
mary is provided in Table 3. For each material we show a figure containing a left
and right plot. The left plot presents the granular orientation effects seen in the
measurements of individual material grains. Each line is the average of a set of
ten O-PTIR single point hyperspectral measurements (described in Section 2.2.1).
The spectra are recorded on different grains, which naturally have different orien-
tations in the sample, leading to the different features in their spectra. The gray
error bars represent the standard deviation of those ten measurements. In these
plots, the coloration does not matter other than to distinguish between the spectra.
Given that some of the spectra showed features with very low intensities, we chose
to show them in a log-lin plot to make these visible compared to more dominating
features.
The right plot of each materials’ figure presents a comparison between an O-
PTIR measurement (black line), an FTIR-A measurement (red line), and a com-
parable Wiley’s Knowitall database entry (cyan line). In this plot, the O-PTIR
measurement is an average over a hyperspectral map. The O-PTIR data utilizes
the left y-axis and the comparisons utilize the right y-axis.

10
Table 3: Peak wavenumbers and wavelengths for each material examined in this work. Dashes in
a cell indicate the lack of an O-PTIR peak measured in this wavenumber range for that material.
Mineral Peak(s) (Wavenumber, cm−1 ) Peak(s) (Wavelength, µm)
Anorthosite 986, 1027, 1150 10.1, 9.7, 8.7
Basalt 1004, 1447, 1465, 1483 10.0, 6.9, 6.8, 6.7
Bronzite 1023, 1465 9.8, 6.8
Fe-Carbonate (Siderite) 1039, 1444, 1465 9.6, 6.9, 6.8
Ferrihydrite – –
Gypsum 990, 1007, 1108, 1622 10.1, 9.9, 9.0, 6.2
Hematite 1013, 1020, 1087 9.9, 9.8, 9.2
Hydrated Silica 1082, 1632 9.2, 6.1
Ilmenite 1033, 1448, 1468, 1481 9.7, 6.9, 6.8, 6.7
Magnetite – –
Mg-Carbonate (Magnesite) 1021, 1468, 1482, 1514 9.8, 6.8, 6.7, 6.6
Mg-Sulfate (Epsomite) 1077, 1101, 1665 9.3, 9.1, 6.0
Olivine 1006, 1125 9.9, 8.9
Smectite 1028, 1447, 1466 9.7, 6.9, 6.8

11
3.1. Anorthosite

Figure 3: Mid-IR O-PTIR measurements of anorthosite (left) and the average spectrum from an
O-PTIR hyperspectral map of anorthosite (right).

Figure 3 (left) shows O-PTIR measurements of anorthosite at chosen locations

on the sample to show variations we hypothesize are due to grain orientation.
Figure 3 (right) shows an O-PTIR measurement of anorthosite with standard
error of mean error bars. Anorthosite shows O-PTIR peaks at 998, 1019, and 1095
wavenumbers. The double peak feature in the lower wavenumber range was seen
in two of the three spectra featured in Figure 3 (left). The O-PTIR measurements
compares quite well to both the FTIR measurement of the same sample (Figure 3,
right, red line) as it shares similar peak locations and a similar spectral shape. Ad-
ditionally, the O-PTIR spectrum matches the anorthite spectrum (Figure 3, right,
cyan line) quite well given the same double feature and matching peak locations.
We made the comparison between anorthosite and anorthite due to the similar
composition despite anorthosite being a plagioclase rock and anorthite being a
plagioclase mineral.
This also is consistent with Farmer (1974) which lists peaks associated with
anorthite in the IR. In the examined wavenumber range, anorthite peaks at 1020,
1085, and 1160. The peaks at 1020 and 1085 correspond closely with two of the
measured O-PTIR peaks.

12
3.2. Basalt (Glass-Rich)

Figure 4: Mid-IR O-PTIR measurements of basalt (left) and the average spectrum from an O-PTIR
hyperspectral map of basalt (right).

Figure 4 (left) shows O-PTIR measurements of basalt at chosen locations on

the sample to show variations we hypothesize are due to grain orientation.
Figure 4 (right) shows an O-PTIR measurement of basalt with standard error
of mean error bars. Basalt shows O-PTIR peaks at 1004, 1447, 1465, and 1483
wavenumbers. The O-PTIR spectrum shared a major peak with the FTIR mea-
surement of the same sample. Though the spectral shape is not a match, there
is significant decrease in signal for both spectra following the max peak. There
were no samples of close enough composition for basalt in the Wiley Knowitall
database.

13
3.3. Bronzite

Figure 5: Mid-IR O-PTIR measurements of bronzite (left) and the average spectrum from an O-
PTIR hyperspectral map of bronzite (right).

Figure 5 (left) shows O-PTIR measurements of bronzite at chosen locations

on the sample to show variations we hypothesize are due to grain orientation.
Figure 5 (right) shows an O-PTIR measurement of bronzite with standard error
of mean error bars. Bronzite shows O-PTIR peaks at 1023 and 1465 wavenum-
bers. The averaged hyperspectral map spectrum appears nearly identical to one of
the granular orientation effect images (Figure 5, left, cyan line). This spectrum had
the most intense peak among the whole set of granular orientation effect spectra
for bronzite. The O-PTIR spectrum shared a peak with both the FTIR measure-
ment of the same sample (red line) and a bronzite entry in the Wiley Knowitall
database. Even though the spectral shape between the O-PTIR measurement and
the FTIR measure of the same sample do not match, there is a consistent decrease
in both measurements following the major peak. Additionally, the O-PTIR mea-
surement closely resembles the database entry for bronzite despite the spectral
feature around the peak not being identical.

3.4. Fe-Carbonate (Siderite)

Figure 6 (left) shows O-PTIR measurements of siderite at chosen locations on
the sample to show variations we hypothesize are due to grain orientation.

14
Figure 6: Mid-IR O-PTIR measurements of siderite (left) and the average spectrum from an O-
PTIR hyperspectral map of siderite (right).

Figure 6 (right) shows an O-PTIR measurement of siderite with standard er-

ror of mean error bars. Siderite shows O-PTIR peaks at 1039, 1444, and 1465
wavenumbers. The O-PTIR measurement shares peaks with both the FTIR mea-
surement of the same sample and the Knowitall database entry for siderite. The
O-PTIR feature between 1000 and 1150 cm−1 is not identical to but does match
the FTIR feature on the same sample. This feature does not exist in the Know-
itall database entry. The O-PTIR feature between 1350 and 1550 cm−1 has sim-
ilar peaks in both the FTIR measurement of the same sample and the Knowitall
database entry. For the the FTIR measurement of the same sample, the peak is
shifted to a lower wavenumber, but is close. The Knowitall database entry peaks
at the same wavenumber but the feature is not nearly as wide. The FTIR mea-
surement of the same sample has a very similar spectral shape to the O-PTIR
measurement despite the shifted peak at the higher wavenumber.
According to Farmer (1974), siderite has absorption features at 1071, 1415,
and 1422. The siderite O-PTIR measured at 1039 cm−1 could be attributed to
the 1071 feature. It is 42 wavenumbers displaced, but the width of that feature is
approximately 200 wavenumbers, so it could cover that band. Additionally, the
peaks at 1444 and 1465 cm−1 are part of a feature with a width of almost 300
wavenumbers. This puts this peak well within the range of values listed for v3 for
siderite.
Farmer (1974) also lists peaks associated with siderite in the IR. In the exam-
ined wavenumber range, siderite peaks at 1412. This peak is only 32 wavenum-

15
bers displaced from one of the measured peaks in the O-PTIR. Additionally, the
feature around the peak at 1039 is similar in FTIR, but with differing intensities.
We hypothesize that this is potentially a polarization issue, but that kind of con-
sideration is beyond the scope of this work.

3.5. Ferrihydrite

Figure 7: Mid-IR O-PTIR measurements of ferrihydrite (left) and the average spectrum from an
O-PTIR hyperspectral map of ferrihydrite (right).

A single response was detected for ferrihydrite (shown in Figure 7, left). Cox
(2024) was unable to detect any signal from this material. The detected signal was
weak and the deviation was high at multiple points in the spectrum. This makes it
difficult to evaluate how granular orientation affects ferrihydrite.
Figure 7 (right) shows an O-PTIR measurement of ferrihydrite with standard
error of mean error bars. Ferrihydrite does not show significant O-PTIR spectral
features nor peaks in the chosen wavenumber range. Though a spectral response
was detected on a single grain, the signal was not strong enough to avoid being
dominated by the null measurements contained in the sample. The FTIR measure-
ment of the same sample was also devoid of features in this wavenumber range.
There were no found entries for ferrihydrite in the Wiley Knowitall database.

16
3.6. Gypsum

Figure 8: Mid-IR O-PTIR measurements of gypsum (left) and the average spectrum from an O-
PTIR hyperspectral map of gypsum (right).

Figure 8 (left) shows O-PTIR measurements of gypsum at chosen locations

on the sample to show variations we hypothesize are due to grain orientation. The
sudden increase at 1400 cm−1 in Figure 8 (left, black line) is caused by a switch
between the QCL chips.
The O-PTIR measurement shares peaks and features with both the FTIR mea-
surement of the same sample (Figure 8, right, red line) and the Knowitall database
comparison (Figure 8, right, cyan line). There is a double peak in the FTIR mea-
surement of the same sample which may also exist in the O-PTIR measurement,
though the second peak in the O-PTIR measurement is noisier. The Knowitall
database comparison is nearly identical though the peak at the higher wavenum-
ber is slightly shifted. Additionally, the feature near 1100 wavenumbers is not as
wide in the Knowitall entry as it is in the O-PTIR measurement.
According to Farmer (1974), anhydrite has absorption features at 1013, 1095,
1126, and 1149 and gypsum has absorption features at 1000, 1006, 1117, 1118,
1131, 1138, 1142, and 1144. The gypsum O-PTIR measured at 1007 cm−1 could
be attributed to the feature at 1006 as it is one wavenumber displaced. The peak at
1108 cm−1 is part of a feature with a width of approximately 100 wavenumbers.
This puts this peak well within the range of values listed for features for gypsum.
The feature around 1629 is seen in both O-PTIR and FTIR-A, but with dif-
fering intensities. This is potentially a polarization issue, but that is beyond the
scope of this work.

17
3.7. Hematite

Figure 9: Mid-IR O-PTIR measurements of hematite (left) and the average spectrum from an O-
PTIR hyperspectral map of hematite (right).

Figure 9 (left) shows O-PTIR measurements of hematite at chosen locations

on the sample to show variations we hypothesize are due to grain orientation.
Figure 9 (right) shows an O-PTIR measurement of hematite with standard er-
ror of mean error bars. Hematite shows O-PTIR peaks at 1013, 1020, and 1087
wavenumbers. The O-PTIR measurement shares peaks with both the FTIR mea-
surement of the same sample and the Knowitall database entry for hematite. The
Knowitall database entry (9, right, cyan line) does not contain the dip on the left
side of the feature that both the FTIR measurement of the same sample and the
O-PTIR measurement possess. However, the feature in the Knowitall database
entry does have the same width as the other two entries in the figure. The O-PTIR
measurement and the FTIR measurement of the same sample are nearly identical
in spectral shape though the intensities do vary slightly. All three measurements
are devoid of spectral features past 1100 cm−1 .

18
3.8. Hydrated Silica

Figure 10: Mid-IR O-PTIR measurements of hydrated silica (left) and the average spectrum from
an O-PTIR hyperspectral map of hydrated silica (right).

Figure 10 (left) shows O-PTIR measurements of hydrated silica at chosen lo-

cations on the sample to show variations we hypothesize are due to grain orienta-
tion.
Figure 10 (right) shows an O-PTIR measurement of hydrated silica with stan-
dard error of mean error bars. Hydrated Silica shows O-PTIR peaks at 1082 and
1632 wavenumbers. The spectral shape of the O-PTIR measurement is extremely
similar to both the FTIR measurement of the same sample and the Knowitall en-
try for hydrated silica. It is nearly identical to the FTIR measurement. In both
comparisons the peaks and features all line up.

19
3.9. Ilmenite

Figure 11: Mid-IR O-PTIR measurements of ilmenite (left) and the average spectrum from an
O-PTIR hyperspectral map of ilmenite (right).

Figure 11 (left) shows O-PTIR measurements of ilmenite at chosen locations

on the sample to show variations we hypothesize are due to grain orientation.
Figure 11 (right) shows an O-PTIR measurement of ilmenite with standard
error of mean error bars. Ilmenite shows O-PTIR peaks at 1033, 1448, 1468,
and 1481 wavenumbers. The averaged hyperspectral map spectrum appears to be
most similar to the spectrum represented by the red line in Figure 11 (left). The
comparable spectrum has a low intensity especially compared to the spectrum
represented by the cyan line in the granular orientation effect figure. The O-PTIR
spectrum is not extremely comparable to the available Knowitall database spec-
trum. However, the two spectra share a peak in the lower wavenumber range.
No FTIR measurement of the same sample is presented here because our FTIR-A
measurement for Ilmenite was null. This is not uncommon in FTIR-A and likely
means the sample was optically thick to the pellet size we produced.

20
3.10. Magnetite

Figure 12: Mid-IR O-PTIR measurements of magnetite (left) and the average spectrum from an
O-PTIR hyperspectral map of magnetite (right).

Figure 12 (left) shows O-PTIR measurements of magnetite at chosen locations

on the sample to show variations we hypothesize are due to grain orientation. Cox
(2024) was unable to detect any signal from this material. The detected signal was
weak and the deviation was high at multiple points in the spectrum. This makes it
difficult to evaluate how granular orientation affects magnetite.
Figure 12 (right) shows an O-PTIR measurement of magnetite with standard
error of mean error bars. Magnetite does not show significant O-PTIR spectral
features nor peaks in the chosen wavenumber range. Though a spectral response
was detected on a single grain, the signal was not strong enough to avoid being
dominated by the null measurements contained in the sample. The FTIR measure-
ment of the same sample was also devoid of features in this wavenumber range.
There were no found entries for magnetite in the Wiley Knowitall database. Cox
(2024) measured this same sample. In that work, using a similar method, there
were also no peaks nor features detected. Figure 12 (right) is similar to those
findings.

21
3.11. Mg-Carbonate (Magnesite)

Figure 13: Mid-IR O-PTIR measurements of magnesite (left) and the average spectrum from an
O-PTIR hyperspectral map of magnesite (right).

Figure 13 (left) shows O-PTIR measurements of magnesite at chosen locations

on the sample to show variations we hypothesize are due to grain orientation.
Figure 13 (right) shows an O-PTIR measurement of magnesite with standard
error of mean error bars. Magnesite shows O-PITR peaks at 1021, 1468, 1482,
and 1514 wavenumbers. The averaged hyperspectral map spectrum appears to be
most similar to the spectra represented by the black and cyan lines in Figure 13
(left). The O-PTIR measurement is comparable with both the FTIR measurement
of the same sample as well as the entry in the Knowitall database. All three
measurements peak in the higher wave number range at very similar positions.
However, the O-PTIR measurement has a wider feature at that peak than both the
FTIR measurement of the same sample and the Knowitall database entry. The
width of the spectral feature for the O-PTIR measurement is more comparable to
the Knowitall database entry.
According to Farmer (1974), magnesite has absorption features at 1436, 1450,
and 1599. The magnesite O-PTIR measured at 1468 cm−1 could be attributed to
the feature at 1450. It is less than 20 wavenumbers displaced, and the width of
that feature is approximately 200 wavenumbers, so it could cover that band. The
features located at 1436 and 1450 could help account for the concentration of ab-
sorbance peaks and the width of the O-PTIR feature at the wavenumber range. The
O-PTIR feature around 1021 does not seem to have a corresponding absorbance
peak.

22
3.12. Mg-Sulfate (Epsomite)

Figure 14: Mid-IR O-PTIR measurements of epsomite (left) and the average spectrum from an
O-PTIR hyperspectral map of epsomite (right).

Figure 14 (left) shows O-PTIR measurements of epsomite at chosen locations

on the sample to show variations we hypothesize are due to grain orientation.
Figure 14 (right) shows an O-PTIR measurement of epsomite with standard error
of mean error bars. Epsomite shows O-PTIR peaks at 1077, 1101, and 1665
wavenumbers. The O-PTIR measurement shares similarities with both the FTIR
measurement of the same sample and the Knowitall database entry, with an almost
perfect match to the Knowitall data and shared peak locations, if not shapes, with
the FTIR measurement.
According to Farmer (1974), epsomite has absorption features at 1020, 1085,
1110, 1155, 1175, and 1235. The epsomite O-PTIR measured at 1077 cm−1 could
be attributed to the 1085 feature as it is less than 10 wavenumbers displaced.
Additionally, the peak at 1101 cm−1 could be attributed to the 1110 feature as it
is also less than 10 wavenumbers displaced. The feature those peaks are a part of
is approximately 200 wavenumbers wide which could cover additional features.

23
3.13. Olivine

Figure 15: Mid-IR O-PTIR measurements of olivine (left) and the average spectrum from an O-
PTIR hyperspectral map of olivine (right).

Figure 15 (left) shows O-PTIR measurements of olivine at chosen locations on

the sample to show variations we hypothesize are due to grain orientation. Such
a dependence on grain orientation has also been seen by Taran and Koch-Müller
(2006) which specifically oriented grains of olivine and found differing spectral
shapes and amplitudes. We similarly found that as different grains were measured,
we saw differing spectral shapes and features as well as differing amplitudes in the
spectra. However, the O-PTIR spectral measurements showed greater variance
than those measurements seen in Taran and Koch-Müller (2006).
Figure 15 (right) shows an O-PTIR measurement of olivine with standard error
of mean error bars. Olivine shows O-PTIR peaks at 1006 and 1125 wavenumbers.
The O-PTIR measurement is not identical to the FTIR measurement of the same
sample nor the database entry in the Knowitall database. The only similarity of the
FTIR measurement of the same sample and the O-PTIR measurement is the peak
location. The similarity between the O-PTIR measurement and the Knowitall
database entry is a better match. They share a similar spectral shape with peak
locations slightly shifted. The less intense peak seems to be shifted more than the
intense peak.
Farmer (1974) lists peaks associated with olivine in the IR. In the examined
wavenumber range, olivine peaks at approximately 1000 cm−1 . This peak is only
6 wavenumbers displaced from one of the measured peaks in the O-PTIR.

24
3.14. Smectite

Figure 16: Mid-IR O-PTIR measurements of smectite (left) and the average spectrum from an
O-PTIR hyperspectral map of smectite (right).

Figure 16 (left) shows O-PTIR measurements of smectite at chosen locations

on the sample to show variations we hypothesize are due to grain orientation.
Figure 16 (right) shows an O-PTIR measurement of smectite with standard
error of mean error bars. Smectite shows O-PTIR peaks at 1028, 1447, and 1466
wavenumbers. The O-PTIR measurement has strong similarities with with both
the FTIR measurement of the same sample nor the database entry in the Knowitall
database. The FTIR measurement of the same sample shows a similar peak around
1050 cm−1 but is slightly shifted to a lower wavenumber. However, the feature
shows a similar width to the O-PTIR measurement. Additionally, the FTIR mea-
surement also shares the peak between 1350 and 1550 cm−1 . The feature peaks
very closely and had a nearly identical width. It also peaks between 1600 and
1700 cm−1 though that feature is not as intense and is less certain than the other
two features. The Knowitall database entry is missing the feature between 1350
and 1550 cm−1 , but does share the feature between 1600 and 1700 cm−1 with a
similar intensity as the O-PTIR measurement. The Knowitall database entry also
shares a peak location with the O-PTIR measurement around 1050 wavenumbers,
though the width of the feature is smaller there.

4. Discussion
Cox et al. (2024) made a case for the use of the O-PTIR technique in plane-

25
tary science space missions. We furthered the work by cataloguing and analyzing
specific materials relevant to planetary science. In this section we will discuss
the results of the measurements reported above. Primarily we will focus on the
outcomes of individual point measurements and the averaged spectra of the hy-
perspectral maps.

4.1. Granular Orientation Effects

Section 2.2.2 mentions the concern of granular orientation effects within our
measurements. It also describes the measurements we took to attempt to demon-
strate those effects in our samples. Section 3 shows the granular orientation effects
for each individual material as we were able to locate them. These effects are well
known and studied in literature (e.g. Serratosa and Bradley, 1958; Shuai and Yang,
2017). These studies typically utilize single grains and change the orientation of
the grain between each measurement. Sometimes these measurements have little
to no effect. Other times these orientation changes cause reduction in peaks or
even shifting peaks and changing spectral shapes.
In our measurements, we see that various grains in a bulk sample are capable
of producing differing spectra. This is likely due to the current configuration of
our system. The visible probing laser is circularly polarized, but the IR laser is
linearly polarized. Typically, one of these possible spectra dominate the average
over a large area. This could be due to random chance of which grains were in
the measurement area. However, in most cases (e.g. bronzite, gypsum, hematite,
etc.) certain spectral shapes and features peak at higher intensities. Those features
are typically dominating the average spectrum. In some cases of multiple spectra,
differing in orientation (e.g. anorthosite, magnesite, epsomite, etc.), it appears
that the average spectrum is influenced by each of the found spectra.
A future work will take individual grains of materials and will analyze specif-
ically how orientation effects spectra. This will better serve as a method to under-
stand how granular orientation effects specifically affect O-PTIR measurements
and clarify the role of light polarization in the instrument.

4.2. Planetary Material Identification

In Section 3 we report averaged spectra of hyperspectral maps for each mate-
rial discussed in this work. These spectra all contain standard error of mean error
bars. In all cases, there were no error bars that would significantly change the
shape of a spectrum. This is significant as it demonstrates the consistency of the
technique as it scans a sample.

26
 Further, the spectral shapes, spectral features, and spectral peaks are typically
unique. In some cases the individual spectral measurements may have resembled
other materials. However, in the case of bulk sample averages, the spectra appear
unique. In other works (such as Olson et al., 2020; Krafft, 2022), Raman is used
simultaneously with the O-PTIR technique. The simultaneous usage allows for
even more unique features for each material. This is feasible due to the green
visible laser causing fluorescence. Though the addition of Raman is out of the
scope of this work, it does speak to the increased ability to identify planetary
materials, which will be studied in future work.

4.3. O-PTIR Comparison to IR Absorption

The comparisons presented in Section 3 showed that O-PTIR measurements
consistently were comparable to IR absorption measurements. This was done by
comparing O-PTIR measurements to FTIR measurements of the same sample and
Wiley’s Knowitall database entries of the same material. In some cases, these
comparisons were nearly identical to the O-PTIR measurements. This is signifi-
cant in that it shows that O-PTIR measurements are comparable to IR absorption
measurements for minerals and granular materials. This shows that O-PTIR mea-
surements produced in future planetary science missions will be comparable to
existing IR absorption databases. Other works (e.g. Klementieva et al., 2020;
Bazin et al., 2022; Zhang et al., 2016) have made cases demonstrating the ability
of O-PTIR to compare to IR absorption (in some cases even producing identical
spectra). This work furthers this concept and demonstrates the ability to do the
same for granular samples.

5. Conclusion and Summary

In this work, we presented O-PTIR measurements of various materials rele-
vant to planetary science. We furthered the work presented in Cox et al. (2024)
and expanded the presentation of the capabilities of the O-PTIR technique. We
did so by: presenting the granular orientation effects on measurements through
single point measurements, presenting analysis of bulk measurements of mate-
rials, and made direct comparisons of O-PTIR measurements to IR absorption
measurements. We utilized the benefits of the O-PTIR technique described by
Cox et al. (2024) to produce a large catalogue of data for materials relevant to
planetary science. This further demonstrates the ability of O-PTIR to benefit the
planetary science community and the future of planetary science space missions.

27
 In upcoming work, we will utilize the ability to identify minerals to identify
constituent materials within a mixture. The ability to do this will be essential to
understanding the origins of materials and to understanding the evolution of these
materials over time. These mixtures will be known regolith simulant mixtures and
will be opaque materials. The ability to quantitatively analyze these materials will
represent a step forward in analysis of foreign planetary materials.

Acknowledgments
This work was supported by the NASA PICASSO grant #80NSSC22K1231.
Additionally, we would like to thank Dr. Kerri Donaldson Hanna for her fruitful
conversations.

References
Andersen, A.C., Haack, H., 2005. Carbonaceous chondrites: tracers of the prebi-
otic chemical evolution of the solar system. International Journal of Astrobiol-
ogy 4, 13–17.

Bazin, D., Bouderlique, E., Tang, E., Daudon, M., Haymann, J.P., Frochot, V.,
Letavernier, E., Van de Perre, E., Williams, J.C., Lingeman, J.E., et al., 2022.
Using mid infrared to perform investigations beyond the diffraction limits of
microcristalline pathologies: advantages and limitation of optical photothermal
ir spectroscopy. Comptes Rendus. Chimie 25, 1–27.

Bhartia, R., Beegle, L.W., DeFlores, L., Abbey, W., Razzell Hollis, J., Uckert,
K., Monacelli, B., Edgett, K.S., Kennedy, M.R., Sylvia, M., et al., 2021. Per-
severance’s scanning habitable environments with raman and luminescence for
organics and chemicals (sherloc) investigation. Space Science Reviews 217,
58.

Cameron, A.G., 1988. Origin of the solar system. IN: Annual review of astron-
omy and astrophysics. Volume 26 (A89-14601 03-90). Palo Alto, CA, Annual
Reviews, Inc., 1988, p. 441-472. 26, 441–472.

Cox, C., Haynes, J., Duffey, C., Bennett, C., Brisset, J., 2024. Photothermal
spectroscopy for planetary sciences: Mid-ir absorption made easy. URL:
https://arxiv.org/abs/2409.11626, arXiv:2409.11626.

28
Cox, C.T., 2024. The case for photothermal spectroscopy in the future of planetary
science missions .

Ehrenfreund, P., Charnley, S.B., 2000. Organic molecules in the interstellar

medium, comets, and meteorites: a voyage from dark clouds to the early earth.
Annual Review of Astronomy and Astrophysics 38, 427–483.

Farmer, V.C., 1974. The Infrared Spectra of Minerals. Mineralogical Society of

Great Britain and Ireland. URL: https://doi.org/10.1180/mono-4,
doi:10.1180/mono-4.

Gomes, C.d.S.F., Silva, J.B.P., 2007. Minerals and clay minerals in medical geol-
ogy. Applied Clay Science 36, 4–21.

Hu, H., Yang, X., Zhai, F., Hu, D., Liu, R., Liu, K., Sun, Z., Dai, Q., 2016. Far-
field nanoscale infrared spectroscopy of vibrational fingerprints of molecules
with graphene plasmons. Nature communications 7, 12334.

Klementieva, O., Sandt, C., Martinsson, I., Kansiz, M., Gouras, G.K., Borondics,
F., 2020. Super-resolution infrared imaging of polymorphic amyloid aggregates
directly in neurons. Advanced Science 7, 1903004.

Krafft, C., 2022. Optical photothermal infrared spectroscopic applications in

microplastics—comparison with fourier transform infrared and raman spec-
troscopy, in: Molecular and Laser Spectroscopy. Elsevier, pp. 305–336.

Olson, N.E., Xiao, Y., Lei, Z., Ault, A.P., 2020. Simultaneous optical photother-
mal infrared (o-ptir) and raman spectroscopy of submicrometer atmospheric
particles. Analytical Chemistry 92, 9932–9939.

Paulus, A., Yogarasa, S., Kansiz, M., Martinsson, I., Gouras, G.K., Deierborg,
T., Engdahl, A., Borondics, F., Klementieva, O., 2022. Correlative imaging to
resolve molecular structures in individual cells: Substrate validation study for
super-resolution infrared microspectroscopy. Nanomedicine: Nanotechnology,
Biology and Medicine 43, 102563.

Serratosa, J.M., Bradley, W., 1958. Determination of the orientation of oh bond

axes in layer silicates by infrared absorption. The Journal of Physical Chemistry
62, 1164–1167.

29
Shuai, K., Yang, X., 2017. Quantitative analysis of h-species in anisotropic miner-
als by polarized infrared spectroscopy along three orthogonal directions. Con-
tributions to Mineralogy and Petrology 172, 1–17.

Spadea, A., Denbigh, J., Lawrence, M.J., Kansiz, M., Gardner, P., 2021. Analysis
of fixed and live single cells using optical photothermal infrared with concomi-
tant raman spectroscopy. Analytical chemistry 93, 3938–3950.

Stroud, R.M., Allen, C., Ansari, A., Anderson, D., Bajt, S., Bassim, N., Bastien,
R.S., Bechtel, H.A., Borg, J., Brenker, F.E., et al., 2014. Stardust interstel-
lar preliminary examination xi: Identification and elemental analysis of impact
craters on al foils from the stardust interstellar dust collector. Meteoritics &
Planetary Science 49, 1698–1719.

Taran, M.N., Koch-Müller, M., 2006. Octahedral cation ordering in mg, fe 2+-
olivine: an optical absorption spectroscopic study. Physics and chemistry of
minerals 33, 511–518.

Tice, M.M., Hurowitz, J.A., Allwood, A.C., Jones, M.W., Orenstein, B.J., David-
off, S., Wright, A.P., Pedersen, D.A., Henneke, J., Tosca, N.J., et al., 2022.
Alteration history of séı́tah formation rocks inferred by pixl x-ray fluorescence,
x-ray diffraction, and multispectral imaging on mars. Science Advances 8,
eabp9084.

Waite, J.H., Lewis, W., Kasprzak, W., Anicich, V., Block, B., Cravens, T.E.,
Fletcher, G., Ip, W.H., Luhmann, J.G., McNutt, R., et al., 2004. The cassini
ion and neutral mass spectrometer (inms) investigation. The Cassini-Huygens
Mission: Orbiter In Situ Investigations Volume 2 , 113–231.

Zhang, D., Li, C., Zhang, C., Slipchenko, M.N., Eakins, G., Cheng, J.X., 2016.
Depth-resolved mid-infrared photothermal imaging of living cells and organ-
isms with submicrometer spatial resolution. Science advances 2, e1600521.

Appendix A. FTIR Pellet Making Method

Using PIKE Technologies’ recipe1 as a guide, a 200 mg pellet was made. The
pellets were made with 1-2% (2̃-4 mg) material of interest and 98-99% (1̃96-198

1
https://www.piketech.com/files/pdfs/PerfectPelletMakingAN.pdf

30
mg) KBr powder. The materials are ground into a fine powder and then mixed
and transferred into a pellet press. The top anvil of the press is rotated clockwise
with downward pressure to better level the sample before pressing the pellet. The
press was evacuated of air for 1 minute and 30 seconds. 3 tons of force was
applied to the pellet then immediately released. Then, 7 tons of force was applied
for 30 seconds then immediately released. The press was evacuated again for 1
minute. Finally, 10 tons of force was applied for 2 minutes. The top of the press is
disassembled from the base, the base is disconnected from the evacuation line, the
rubber o-rings are removed from the press, and then the top of the press is placed
back in the compactor. The compactor is used to push the pellet free of the press.
The pellet is removed from the press and then placed into a holder for analysis.
Finished pellets are placed in a dry box for storage until ready for analysis.

Appendix B. Max Peaks and Deviations

Below is a series of tables containing max peaks and deviations for the indi-
vidually measured spectra for each material.

Table B.4: A table featuring the max peaks and max deviations of each O-PTIR measurement of
anorthosite featured in Figure 3 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 7.322693 0.324179
Spectrum 2 7.977738 0.213787
Spectrum 3 7.357513 0.314597

Table B.5: A table featuring the max peaks and max deviations of each O-PTIR measurement of
basalt featured in Figure 4 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 5.455014 0.195806
Spectrum 2 59.42456 3.613043

31
Table B.6: A table featuring the max peaks and max deviations of each O-PTIR measurement of
bronzite featured in Figure 5 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 10.58403 0.693008
Spectrum 2 29.0392 1.997469
Spectrum 3 84.8512 5.504187
Spectrum 4 14.35081 0.534176

Table B.7: A table featuring the max peaks and max deviations of each O-PTIR measurement of
siderite featured in Figure 6 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 2.905344 0.137475
Spectrum 2 2.974705 0.157269
Spectrum 3 4.072176 0.541465

Table B.8: A table featuring the max peaks and max deviations of each O-PTIR measurement of
ferrihydrite featured in Figure 7 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 1.227325 0.234679

Table B.9: A table featuring the max peaks and max deviations of each O-PTIR measurement of
gypsum featured in Figure 8 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 2.681214 0.101136
Spectrum 2 2.453152 0.142794
Spectrum 3 1.9243 0.071491

32
Table B.10: A table featuring the max peaks and max deviations of each O-PTIR measurement of
hematite featured in Figure 9 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 9.201209 0.458004
Spectrum 2 0.609176 0.091371
Spectrum 3 3.38826 0.114438

Table B.11: A table featuring the max peaks and max deviations of each O-PTIR measurement of
hydrated silica featured in Figure 10 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 20.57146 0.598685
Spectrum 2 19.84694 3.233245
Spectrum 3 17.09244 2.493837

Table B.12: A table featuring the max peaks and max deviations of each O-PTIR measurement of
ilmenite featured in Figure 11 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 0.706231 0.114065
Spectrum 2 0.861958 0.188545
Spectrum 3 4.23806 0.157548

Table B.13: A table featuring the max peaks and max deviations of each O-PTIR measurement of
magnetite featured in Figure 12 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 12.28606 0.353978

33
Table B.14: A table featuring the max peaks and max deviations of each O-PTIR measurement of
magnesite featured in Figure 13 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 22.60104 2.37159
Spectrum 2 26.57293 0.742645
Spectrum 3 9.402586 0.405154
Spectrum 4 5.560586 0.844084

Table B.15: A table featuring the max peaks and max deviations of each O-PTIR measurement of
epsomite featured in Figure 14 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 1.113239 0.282312
Spectrum 2 2.566355 0.399092
Spectrum 3 6.026462 0.246884

Table B.16: A table featuring the max peaks and max deviations of each O-PTIR measurement of
olivine featured in Figure 15 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 95.26071 9.271133
Spectrum 2 4.663567 1.491738
Spectrum 3 5.732936 0.330583

34
Table B.17: A table featuring the max peaks and max deviations of each O-PTIR measurement of
smectite featured in Figure 16 (left).
Spectrum Max Peak Max Deviation
Spectrum 1 6.752547 0.922227
Spectrum 2 6.23368 1.048889
Spectrum 3 6.443592 1.591813
Spectrum 4 7.860179 2.305671
Spectrum 5 16.75025 4.600344

35