ANALYSIS OF PFAS OVER TIME IN SOIL AT VENETUCCI FARM

ANALYSIS OF PFAS OVER TIME IN SOIL AT VENETUCCI FARM

by

ERIC PATRICK GAULKE

B.S. University of Colorado Colorado Springs, 2017

A thesis submitted to the Graduate Faculty of the

University of Colorado Colorado Springs

in partial fulfillment of the

requirements for the degree of

Master of Sciences

Department of Chemistry and Biochemistry

ii

This thesis for the Master of Sciences degree by

Eric Patrick Gaulke

has been approved for the

Department of Chemistry and Biochemistry

by

Janel Owens, Chair

Kevin Tvrdy

Wendy Haggren

iii

Gaulke, Eric Patrick (M.Sc., Chemistry)

Analysis of PFAS over Time in Soil at Venetucci Farm

Thesis directed by Associate Professor with Janel Owens

ABSTRACT

Per- and polyfluoroalkyl substances (PFAS, previously known as PFCs)

have been a concern as an environmental contaminant for decades. They are

resistant to degradation and commonly referred t

o as “forever chemicals.” Their

use has led to extensive groundwater pollution across the globe and have

polluted the aquifer below our very own local Venetucci Farm. This study used

high performance liquid chromatography coupled to tandem mass spectrometry

to look at PFAS concentrations in soil at Venetucci Farm over a one-year period

in multiple sample sites across the farm. Results have shown relatively stable

PFAS concentrations over time at sample sites with low total organic carbon

levels (p = 0.163, 0.213) and significant variability of PFAS concentrations over

time at sample sites with higher total organic carbon levels (p = 3.670×10

-3

_,

3.171×10

-8

_{, 5.273×10}

-11

_{). Overall, long chain PFAS such as PFOA and PFOS}

have been found with concentrations over 1.82 ng/g while other PFAS were

undetected or below the limit of quantitation of 0.04 ng/g.

iv

ACKNOWLEDGEMENTS

The Shimadzu Corporation and National Science Foundation

(CHE-1429567) are acknowledged for funds to purchase the Shimadzu LCMS-8030

system. The Pikes Peak Community Foundation, Sam Clarke, Tim Zant, and

Susan Gordon are acknowledged for their roles in providing access to the farm

and all of their help providing information about the farm. The Department of

Chemistry and Biochemistry are acknowledged for the tremendous support and

encouragement imparted onto me during my time at UCCS and for the Graduate

Teaching Assistantship opportunity which allowed me to develop leadership skills

and pay for my tuition. Dr. Tvrdy and Dr. Haggren are acknowledged for joining

me in my pursuit of a graduate degree as members of my thesis committee and

helping me throughout this process. I also want to acknowledge Dr. Owens

specifically for giving me the opportunity to work in her research lab as an

undergraduate student, which truly changed the course of my life.

v

Table of Contents

CHAPTER 1 ... 1

BACKGROUND AND LITERATURE REVIEW ... 1

1.1 Definition and History ... 1

1.2 Chemical Properties ... 3

1.3 Toxicity ... 5

1.3.1 Toxicity of PFOS and PFOA ... 5

1.3.2 Toxicity of Other PFAS ... 11

1.4 Use of PFAS on Military Bases in U.S.A. and Resulting Pollution ... 14

1.5 EPA Health Advisory Limit, Impact on Venetucci Farm, and Impact on

Surrounding Areas ... 16

1.6 PFAS Transport and Prediction for Venetucci Farm ... 17

1.7 Experiment Parameters and Hypothesis ... 19

CHAPTER 2 ... 24

MATERIALS AND METHODS ... 24

2.1 Chemicals and Supplies ... 24

2.2 Sample Collection and Preparation ... 25

2.3 Instrumentation... 31

vi

2.3.2 Tandem Mass Spectrometry ... 33

2.4 Quality Assurance and Quality Control ... 34

2.5 Statistical Analysis of Data ... 38

CHAPTER 3 ... 40

METHOD DEVELOPMENT AND OPTIMIZATION ... 40

3.1 Validated Changes ... 40

3.2 Invalidated Changes ... 44

CHAPTER 4 ... 46

RESULTS AND DISCUSSION ... 46

4.1 “Soil 1” Sample Site Data ... 46

4

.2 “Pinello” Sample Site Data ... 47

4.3 “Lower Field” Sample Site Data ... 48

4.5 “Raised Bed” Sample Site Data ... 50

4.6 Discussion ... 51

CHAPTER 5 ... 55

CONCLUSIONS AND FUTURE WORK ... 55

REFERENCES ... 57

Appendix ... 62

Condensed Sample Site Data with Associated Error ... 62

vii

Standards Raw Data ... 117

Spike Recovery Raw Data ... 219

viii

List of Figures

Figure 1

- Image showing generic stick structure of PFAS anions capped with

either a carboxylic acid or sulfonate group, n denotes that the hydrophobic C-F

chain can be longer or shorter depending on the specific PFAS molecule

……….1

Figure 2

- Map showing location of Venetucci Farm in greater Colorado area and

sample locations on the farm with Pinello (A), Soil 1 (B), Raised Bed (C), Hoop

House Basil (D), and Lower Field (E)

_{………..19}

Figure 3

- The predicted concentration of PFAS in soil at the five sample sites

chosen for this study

………..22

Figure 4

- A total ion chromatogram (TIC) of a 1.82 ng/mL analytical standard for

the PFAS compounds analyzed in this study

……….32

Figure 5

- Calibration curve for PFPeS generated from the instrument response

of standards in the sample batch

……….35

Figure 6

- Chromatogram of blank water for PFHxA with time along the x-axis

and instrument response measured by the y-axis

_………37

Figure 7

- Average spike and recovery data for not shaking Falcon tubes by

hand prior to sonication and shaking Falcon tubes by hand. Sample sets 1-3

were not shaken by hand and sample sets 4-6 were shaken by hand. Error bars

denote standard deviation

………..………..41

Figure 8

- Graph showing the ratio of a 0.77 ng/mL standard filtered using a

nylon GMF filter and unfiltered

……….42

Figure 9

- Graph showing the peak height ratio of a 0.77 ng/mL standard filtered

ix

Figure 10

- Graph depicting the ratio of peak heights for an unfiltered 0.77 ng/mL

standard and the same standard filtered with a RC 15 (0.45 μm) filter………….44

Figure 11

- Graphical representation of the data generated from the Soil 1

sample site over the duration of the study

………..47

Figure 12

- Graphical representation of the data generated from the Pinello

sample site over the duration of the study

………..48

Figure 13

- Graphical representation of the data generated from the LF sample

site over the duration of the study

………49

Figure 14

- Graphical representation of the data generated from the HHB sample

site over the duration of the study

………50

Figure 15

- Graphical representation of the data generated from the RB sample

x

List of Tables

Table 1

- PFAS compounds studied in this experiment with abbreviations,

chemical properties, and identifying information

……….…3

Table 2

- Mass spectrometry specifications, compounds are listed in order of

elution with PFBA being the first to elute

………33

Table 3

- Average percent recovery and percent relative standard deviation for

all spiked samples (n=2) in each batch throughout the entire duration of the

study

……….36

Table 4

- Data for PFOA concentrations at Soil 1 sample site with relative

percent difference (%RPD) for all months that Soil 1 samples were ran in

duplicate to show as an example for %RPD

………..38

Table 5

- Total organic carbon (TOC) in the soil at Venetucci Farm for the five

xi

List of Abbreviations

AFFF

aqueous film-forming foam

ALC

adult Leydig cell

ANOVA

analysis of variance

ATSDR

Agency for Toxic Substances and Disease Registry

CAS

Chemical Abstracts Service

CID

collision induced dissociation

C

solid

concentration in solid

CV

cartridge volume

C

water

concentration in water

DI

de-ionized

DoD

Department of Defense

dpf

days post fertilization

EDS

ethane dimethyl sulfonate

EPA

Environmental Protection Agency

ESI

electrospray ionization

ƒ

OM

fraction of organic matter

GMF

glass microfiber

HHB

hoop house basil

hpf

hours post fertilization

HPLC-MS/MS

high performance liquid chromatography coupled to tandem mass

spectrometry

ILC

immature Leydig cell

K

d

solid-water distribution coefficient

K

OM

organic matter partition coefficient

K

OW

octanol-water partition coefficient

LF

lower field

LHA

lifetime health advisory

LOAEC

lowest observed adverse effect concentration

LOQ

limit of quantitation

m/z

mass-to-charge ratio

MCT

microcentrifuge tube

MeHg

methyl mercury

NMDA

N-methyl-D-aspartate

PFAS

per- and polyfluoroalkyl substances

PFBA

perfluorobutanoic acid

PFBS

perfluorobutanesulfonic acid

PFC

perfluorinated chemical

xii

PFHpA

perfluoroheptanoic acid

PFHpS

perfluoroheptanesulfonic acid

PFHxA

perfluorohexanoic acid

PFHxS

perfluorohexanesulfonic acid

PFOA

perfluorooctanoic acid

PFOS

perfluorooctanesulfonic acid

PFPeA

perfluoropentanoic acid

PFPeS

perfluoropentanesulfonic acid

PPARα

peroxisome proliferator-activated receptor-alpha

ppb

parts per billion

ppt

parts per trillion

QA/QC

quality assurance and quality control

Q-RT-PCR

quantitative-real time-polymerase chain reaction

RB

raised bed

RC

regenerated cellulose

RCF

relative centrifugal force

ROS

reactive oxygen species

RPD

relative percent difference

RSD

relative standard deviation

SLC

stem Leydig cell

SPE

solid phase extraction

TOC

total organic carbon

USA

United States of America

1

CHAPTER 1

BACKGROUND AND LITERATURE REVIEW

1.1 Definition and History

Per- and polyfluoroalkyl substances (PFAS), previously known as

perfluorinated chemicals (PFCs), are a large class of anthropogenic chemicals.

They consist of an alkane chain of carbon atoms saturated with fluorine atoms

with one end of the carbon chain capped with either a carboxylic acid or a

sulfonate group. A generic structure for both the carboxylic acid and sulfonate

capped PFAS is shown in Figure 1. The carbon chains can range from very few

carbon atoms in the chain to very many, thus making the class of PFAS very

large. Owing to the physical properties of the C-C bonds, C-F bonds, and

–oic

acid/sulfonate functional groups, PFAS are amphiphilic in nature. The saturated

fluorine tail is highly hydrophobic and the functional head groups are hydrophilic.

This makes them resistant to both oils and water as well as being good

surfactants useful for aiding in extinguishing petroleum based fires.

1,2

_{Because of}

their wide array of useful properties and applications, PFAS have been heavily

Figure 1

. Image showing generic stick structure of PFAS anions capped with

either a carboxylic acid or sulfonate group, n denotes that the hydrophobic C-F

chain can be longer or shorter depending on the specific PFAS molecule

2

used worldwide in cookware, food packaging, clothing, carpets, firefighting foam,

and many other products. Annual worldwide production of these chemicals is on

the order of millions of kilograms.

2,3

_{PFAS have now been found in air, soil,}

water, rain, snow, biota, and human serum in many countries across the globe

and are now realized as ubiquitous pollutants.

4-11

PFAS have a fairly long history of use around the world. Perfluorooctanoic

acid (PFOA) and perfluorooctanesulfonic acid (PFOS) were first synthesized by a

company named 3M (makers of

Scotch™ tapes and Post-it

®

_{notes) in the}

1940s.

12

_{After being studied, the compounds were quickly realized as being}

resistant to both oils and water. These observations resulted in PFOS and PFOA

being mass produced. By the 1960s, PFOS was being used to develop aqueous

film-forming foam (AFFF), which was used primarily at airports and military bases

in the USA to extinguish petroleum based fires.

12

_{In the 1970s, both the National}

Academy of Sciences and 3M independently determined that PFOA was

completely resistant to environmental degradation and would likely persist in the

environment. In the early 2000s, this was re-determined by the EPA. Currently,

all American companies that have used PFAS in the past have phased out

legacy PFAS (7 or 8 carbons long) and have begun switching to shorter chain

PFAS (fewer than 7 carbons long). However, now that some legacy PFAS are

ubiquitous pollutants and resistant to degradation, switching to shorter chain

PFAS will likely do little to impact the general widespread environmental pollution

by PFAS chemicals.

3

1.2 Chemical Properties

As mentioned previously, the chemical properties of PFAS compounds

make them very unique with a large number of different commercial and

industrial applications. This section will briefly cover some of the typical chemical

properties as well as identifiers such as CAS registry numbers. Due to the nature

of the PFAS chemicals, some properties are difficult to determine experimentally

so models are used to estimate some chemical properties such as solubility in

water and the octanol-water partition coefficient, or K

OW

.

13

In order to estimate

Table 1

. PFAS compounds studied in this experiment with abbreviations,

chemical properties, and identifying information

Name (abbreviation) CAS Registry Chemical Formula Molecular Weight (g/mol) Aqueous Solubility Log KOW Perfluorooctanesulfonic acid (PFOS) 1763-23-1 C8HF17O3S 500.13 3.2 x 10 -3 mg/L @ 25o_C (estimated) 4.49 (estimated) Perfluorooctanoic acid (PFOA) 335-67-1 C 8HF15O2 414.07 3300 mg/L @ 25o_C14 4.81 _(estimated) Perfluoroheptanesulfonic acid (PFHpS) 375-92-8 C7HF15O3S 450.12 n/a 3.82 (estimated) Perfluoroheptanoic acid (PFHpA) 375-85-9 C 7HF13O2 364.06 4.37 x 105 mg/L @ 25o_C13 4.15 (estimated) Perfluorohexanesulfonic acid (PFHxS) 355-46-4 C6HF13O3S 400.12 6.2 mg/L @ 25o_C (estimated) 3.16 (estimated) Perfluorohexanoic acid (PFHxA) 307-24-4 C6HF11O2 314.05 15,700 mg/L @ ambient temperature15 3.48 (estimated) Perfluoropentanesulfonic

acid (PFPeS) 2706-91-4 C5HF11O3S 350.11 n/a n/a Perfluoropentanoic acid (PFPeA) 2706-90-3 C 5HF9O2 264.05 n/a n/a Perfluorobutanesulfonic acid (PFBS) 375-73-5 C4HF9O3S 300.10 510 mg/L @ temperature not specified16 1.82 (estimated) Perfluorobutanoic acid

4

the Log K

OW

values for these PFAS compounds, the Kowwin version 1.67

provided by the USEPA.

An interesting trend in PFAS molecules is the decreasing value of log K

OW

as tail length decreases. The octanol-water partition coefficient measures how a

chemical species behaves when introduced to a system of octanol and water. It

is used to show if the chemical species in question will congregate more in the

octanol layer or the water layer. This value is used commonly in environmental

chemistry to predict whether a chemical is likely to be retained in the fatty lipids

inside of someone’s body, or if the body is able to excrete the chemical as

waste.

17

_{The implications of the observed trend in decreasing log K}

OW

based on

tail length are that legacy PFAS like PFOS and PFOA are much more

bioaccumulative than their shorter chain counterparts, such as PFBS. This is

because the chemicals will be retained in fatty areas of the body and are much

more difficult for the body to excrete. Reported elimination half-lives in human

adults for PFOA and PFOS have shown to be around 3 years and 5 years

respectively, with some studies showing elimination half-lives on the scale of

decades.

13

_{Conversely, elimination half-lives in human adults for shorter chain}

PFAS like PFBA and PFBS have shown to be around 72 hours and 665 hours

respectively.

13

_{Clearly, the length of the hydrophobic tail has significant effects on}

how these compounds interact with the environment and living organisms. The

effect of how tail length affects PFAS transport through soil will be discussed

later in the chapter.

5

1.3 Toxicity

Toxicity of the PFAS compounds studied in this experiment will be broken

into sections based on the individual compounds, starting with the more common

legacy PFAS.

1.3.1 Toxicity of PFOS and PFOA

1.3.1a PFOS

Few data exist regarding the toxicity of PFAS in humans. There are,

however, many studies regarding the toxicity of various PFAS in different animals

and plants. Because PFOS was one of the most widely used PFAS for decades,

many toxicity studies concerning PFAS were done on PFOS specifically. The

range of these studies has been very wide, including studies on mice and rats,

marine biota, and even some plants. These studies have shown various effects

on these living organisms. This section on PFOS toxicity will highlight some of

these various studies and cover some key results, particularly from recently

published data.

The first study mentioned here is one reported by Reardon et al. (2019).

This team studied the neurodevelopmental and metabolomic responses in

Sprague-Dawley rats from prenatal co-exposure to PFOS and methyl mercury

(MeHg).

18

_{Although the study focused mainly on co-exposure with these two}

chemicals, part of the study included prenatal exposure to PFOS alone without

the addition of MeHg. Beginning on the first day of pregnancy, the mother rats

were dosed orally with gelatin capsules containing 1 mg/kg body weight of

PFOS. Dosage continued until weaning at 21 days after birth. After many

6

different behavioral tests and studies conducted on the offspring and

metabolomics tests performed on select regions of the brain, the results showed

that administration of PFOS alone did not affect the growth of the pups but did

affect the behavior of the pups as shown with increased activity or hyperactivity

when compared to the control group. In pups exposed to PFOS an increase in

aspartic acid, glycine, serine, and N-methyl-D-aspartate (NMDA) receptor

agonists was observed in the frontal cortex. It was hypothesized that the

observed hyperactivity resulted from stimulated NMDA receptors due to the

increased agonists for these receptors. In addition to these increased amino

acids and NMDA receptor agonists, the researchers observed lower levels of

cortical lipids such as phosphatidylcholine and sphingomyelins. The data from

this study suggest that prenatal exposure and pre-weaning exposure to PFOS

results in altered lipid metabolism in the frontal cortex, as well increased levels of

various amino acids and NMDA receptor agonists which manifests as

hyperactivity. More research needs to be done to better understand the exact

mechanism by which PFOS affects the levels of amino acids and NMDA receptor

agonists in the brain.

Another study done by Martínez et al. (2019) aimed to define the actual

mechanisms of PFOS toxicity in zebrafish embryos by combining morphological

and transcriptomic analyses.

19

_{This study first took zebra fish embryos at 2-5}

days post fertilization (dpf) and exposed them to varying levels of PFOS in order

to determine the lowest observed adverse effect concentration (LOAEC) by

observing any morphological changes in the embryos. Once the LOAEC was

7

determined, concentrations below the LOAEC were used for the transcriptomic

analyses. The results from the transcriptomic analyses showed that at low

aqueous concentrations of PFOS (similar to environmental concentrations at

ng/L scale), genes related to natural immunity and defense against infections

were significantly affected as well as other genomic regions. Specific regions of

the genome affected included ones that influence cell adhesion, cell-cell

signaling, apoptosis, and immune response. Essentially, this study confirmed that

PFOS exposure at a molecular level comparable to levels already found in

human serum can adversely affect the immune system of zebrafish. These

results led the authors to push for revised regulations on tolerable levels of PFOS

for human health.

Different from animal studies, a study done by Qian et al. tested PFOS

exposure on two different species of riparian plants, Acorus calamus and

Phragmites communis

in order to determine the toxic effects of PFOS on these

plants.

20

_{The study took these two different riparian plants and used a hydroponic}

system to expose the plants to natural lake water that had been spiked with

varying levels of PFOS over 48 days. Random plants were selected for testing at

0, 3, 6, 12, 24, and 48 days into the experiment. The team analyzed samples

from the leaf, stem, and root regions of the plants. The study looked at multiple

indicators of plant health including chlorophyll levels, soluble protein contents,

and oxidative stress response. The results showed that with higher

concentrations of PFOS (up to 50 mg/L) there were lower levels of chlorophyll

until day 24, at which point the chlorophyll levels seemed to stabilize and, in most

8

cases, began to rise again. This gave an indication that the plants were able to

adapt somehow to the PFOS exposure and work to regain their natural levels of

chlorophyll. The results from this part of the study showed that at lower levels of

PFOS (0.1 mg/L and 1 mg/L) the soluble protein content of the plants actually

rose. At higher levels of PFOS (50 mg/L), however, the soluble protein content

decreased significantly. PFOS was also shown to increase oxidative stress

resulting from higher levels of malonaldehyde and hydrogen peroxide and

actually decreased the activities of various antioxidant enzymes such as

superoxide dismutase, catalase, and peroxidase. Overall the study showed that

the levels of PFOS that have been found in the environment can have noticeable

effects on plant health such as reduced chlorophyll levels, increased oxidative

stress, and decreased levels of various soluble proteins.

Unfortunately, animal and plant studies are not always comparable to

human toxicity or effects from chemicals in humans. However, there have been

large-scale studies done on human populations that have looked for correlations

between PFOS exposure and health effects in various populations. A division of

the U.S Department of Health and Human Services called the Agency for Toxic

Substances and Disease Registry (ATSDR) has published a toxicological profile

for PFAS substances with information from studies done on PFAS up to June of

2018.

13

_{Due to the nature of these studies on various populations, only}

correlations can be drawn and direct causality cannot necessarily be proven.

Regardless, results from these large scale studies have shown that there is a

correlation between PFOS exposure and pregnancy-induced

hypertension/pre-9

eclampsia, liver damage characterized by increased serum enzyme levels and

decreased serum bilirubin, increases in serum lipids (mostly total cholesterol and

low-density lipoprotein cholesterol), increased risk of thyroid disease, decreased

antibody response to vaccines, greater risk of decreased fertility, small

decreases in birth weight, and greater risk of testicular and kidney cancer.

13

Between results from various animal and plant studies and results from studies

on human populations exposed to PFOS, it seems that PFOS is toxic to

biological systems and has the capacity to adversely affect human health and

homeostasis.

1.3.1b PFOA

Although the chemical structure of PFOA is very similar to PFOS, the

carboxylic acid functional group lends some unique characteristics to PFOA that

differentiate it from PFOS. This section looks at several studies focused on

PFOA for various toxicity effects in living organisms.

The first study highlighted here for PFOA toxicity was one done by Crebelli

et al. and focuses on the ability of PFOA to induce genotoxic effects in mice.

21

While PFOA has not been shown to directly cause genotoxic effects, it has been

thought to generate various reactive oxygen species (ROS) that can then lead to

subsequent genotoxicity through oxidative stress. Interestingly, the results from

this study showed that extended PFOA exposure in mice (up to 5 mg/kg body

weight) from drinking water did not lead to genotoxicity through oxidative stress.

However, the study did show that the PFOA exposure led to liver toxicity,

10

weight, and increased liver weight at both exposure levels of 1 mg/kg body

weight and 5 mg/kg body weight. The liver toxicity from PFOA exposure at 5

mg/kg body weight was manifested as less compact and less homogenous liver

parenchyma tissue which showed cytoplasmic vacoulation and hepatocellular

hypertrophy (thought to be the reason for increased liver weight).

A second study done by Lu et al. has observed the relation between

PFOA exposure and reproductive organ toxicity in male rats.

22

_{The goal of this}

study was to mimic pubertal development in 49 day old Sprague-Dawley rats by

using ethane dimethyl sulfonate (EDS) to eliminate adult Leydig cells (ALCs) in

the rats and then expose the rats to PFOA orally to study the progress of

transformation of stem Leydig cells (SLCs) into immature Leydig cells (ILCs) and

finally into a new generation of ALCs. Results from this study showed that PFOA

has the capacity to significantly delay the maturation of SLCs into ALCs resulting

in lower levels of testosterone in the rats. Overall, this study showed that PFOA

can be toxic to SLCs in testis and is of concern to pubertal boys exposed to

PFOA.

A third study by Du et al. intended to characterize the toxic effects of

PFOA as an endocrine disrupting chemical in zebrafish and steroidogenic effects

in H295R human cancer cells.

23

_{The study took zebra fish embryos that were}

forming normally at 4 hours post fertilization (hpf) and exposed them to PFOA at

concentrations of 100, 200, and 500 µg/L. After exposure to the PFOA solutions

for 120 hours, results from quantitative-real time-PCR (Q-RT-PCR) showed that

the gene expression levels for esr1 at 500 µg/L were significantly elevated and

11

expression levels for hhex and pax8 were significantly up-regulated in the

zebrafish. When the H259R cells were studied, they found that PFOA exposure

had decreased levels of testosterone

and induced gene expression of 17βHSD1,

CYP19, 3βHSD2, and CYP11B2—all of which are enzymes that play significant

roles in the endocrine system. PFOA exposure in the H295R cells also reduced

17βHSD4 expression and lowered levels of SF-1. The results from this study

show that PFOA has the capacity to affect estrogen receptor genes, early thyroid

development genes, and steroid synthesis genes showing toxicity to the

endocrine system.

The ATSDR also included PFOA in their toxicological profile report for

PFAS chemicals. Interestingly, PFOA showed many of the same results as

PFOS when exposed populations were studied.

13

_{This may be because both}

PFOS and PFOA are considered ubiquitous pollutants, so many of the

individuals included in these studies may have had both PFOS and PFOA

present in their bodies. Nonetheless, results from these studies summarized by

the ATSDR showed that PFOA is correlated with pregnancy-induced

hypertension/pre-eclampsia, liver damage, increases in serum lipids, increased

risk of thyroid disease, decreased antibody response to vaccines, increased risk

of asthma diagnosis, increased risk of decreased fertility, and small decreases in

birth weight.

13

1.3.2 Toxicity of Other PFAS

The vast majority of health and toxicity studies done on PFAS compounds

are focused on PFOS and PFOA which have historically been the most widely

12

used PFAS compounds. The focus on PFOA and PFOS has led to very little data

generated on the toxicity of other PFAS. For example, the ATSDR toxicological

review on PFAS compounds included 187 animal studies of which only seven

included PFHxS and PFHxA and only six included PFBA. None of the studies

included in the ATSDR review (animal or human) include PFBS, PFPeS, PFPeA,

or PFHpS. The toxicity of these other PFAS compounds should be of great

concern because now that legacy PFAS have been phased out, the shorter chain

PFAS compounds are now being used as replacements in the products that used

to contain legacy PFAS. This increased use of short chain PFAS results from an

effort to replace legacy PFAS in products that still have a high demand for use,

such as the AFFFs at airports and military bases.

1.3.2a PFHpS and PFHpA

No significant toxicology data exists for PFHpS. PFHpA has been shown

to activate mouse and human peroxisome proliferator-activated receptor-alpha

(PPARα) in COS-1 cells, suggesting there are some hepatotoxic effects.

24

1.3.2b PFHxS and PFHxA

According to the ATSDR review, PFHxS in humans may cause liver

damage marked by increase in serum enzymes and decreases in serum bilirubin

levels and can also affect response to vaccines by decreased antibody levels.

13

PFHxS and PFHxA have shown to activate mouse and human

PPARα in COS-1

cells, suggesting there are some hepatotoxic effects.

13

13

1.3.2c PFPeS and PFPeA

No significant toxicology data exists for PFPeS. PFPeA has been shown

to activate mouse and

human PPARα in COS-1 cells, suggesting there are some

hepatotoxic effects.

24

1.3.2d PFBS and PFBA

Since legacy PFAS have been phased out, the main replacement for

PFOS has been PFBS. Because of this, there has been more focus on the

toxicity of PFBS than there has been on other short chain PFAS. A study done

on zebrafish embryos by Sant et al. found that when zebrafish embryos were

exposed to up to 32 µM PFBS, the embryos showed a slight increase in tail

deformities, cranial malformations, and failure to inflate the swim bladder. The

PFBS exposure also affected several genes related to lipid metabolism.

25

_A

different study done by Chen et al. on how PFBS exposure at environmentally

relevant concentrations affects marine medaka, O. melastigma, showed that

PFBS has the ability to accumulate in the eyes and resulted in potential adverse

effects to visual function through neural signaling pathways.

26

_{A third study by}

Chen et al. compared the toxic effects of PFBS to PFOS by exposing

Caenorhabditis elegans

to high levels of PFBS that resulted in similar internal

concentrations as PFOS.

27

_{The results from this study showed reproductive}

toxicity marked by decreased egg production and brood number, increased

capacity of germ cell apoptosis through upregulated expression of pro-apototic

genes egl-1 and ced-13, and increased ROS levels. So far, many of the studies

done on PFBS toxicity have concluded that PFBS is much less bioaccumulative

14

than PFOS but if high enough concentrations are reached in the body, it can

have similar toxicities to PFOS, the compound it has been intended to replace.

The study done by Wolf et al. on PPARα activation also included PFBA,

and similar to other PFAS carboxylates, showed that PFBA can activate mouse

and human

PPARα in COS-1 cells, suggesting there are some hepatotoxic

effects.

13

_{Other than this, studies included in the toxicological review by the}

ATSDR showed that PFBA is not toxic developmentally, immunologically, to the

respiratory system, cardiovascular system, gastrointestinal system, renal system,

bone and skeletal muscle system, endocrine system, brain, spinal cord, or

reproductive system. Th

e only finding other than PPARα activation was that

PFBA can affect the hematological indices manifested as reduced red blood cell

count and hemoglobin.

13

1.4 Use of PFAS on Military Bases in U.S.A. and Resulting Pollution

One of the major uses of legacy PFAS has been in AFFFs at military

bases. As stated earlier, these AFFFs containing legacy PFAS are used to

extinguish petroleum-based fires. AFFFs are effective at rapidly extinguishing

petroleum-based fires because they are able to spread across the surface of the

fire separating the fuel source from the flames. A secondary characteristic of

AFFFs containing fluorosurfactants is that after they separate the fuel source

from the flames, the film that forms across the surface of the fuel prevents

flammable vapors from escaping and spreading. This makes these AFFFs

effective at both extinguishing a fire and preventing re-ignition of the fuel, which

is a serious concern in any petroleum-based fire. This multi-action effectiveness

15

of these foams has vast use across the country in very large quantities for either

training or extinguishing actual fires that occur, particularly on military bases and

airports.

Since their implementation in AFFFs in the 1960s and the beginning of

their phase out by American manufacturers in 2000, it has been shown that the

decades of use has led to extensive PFAS pollution in drinking water and

groundwater within and around military bases.

28

_{This water pollution has affected}

many communities near military bases across the country and at/near American

military bases in other countries as well.

28

_{In March of 2018, the Department of}

Defense (DoD) published a document outlining PFAS levels in drinking water

sourced by any branch of the DoD (Air Force, Army, Navy, and Defense Logistics

Agency) as well as some private drinking water sources located nearby military

bases. The document also detailed any actions that might have been taken by

the military in attempts to sever the exposure pathway of PFAS into humans from

drinking water

—for example, providing bottled water to residents of an affected

community.

28

_{The published results showed that out of 2,668 private groundwater}

wells located near military bases, 1,621 tested above the EPA Lifetime Health

Advisory (LHA) for PFOS/PFOA set in 2016.

28

_{Results from the testing of military}

supplied drinking water showed that out of 524 drinking water systems supplied

by the DoD, 24 tested above the EPA LHA for PFOS/PFOA.

28

16

1.5 EPA Health Advisory Limit, Impact on Venetucci Farm, and Impact on

Surrounding Areas

In 2016 the EPA set new LHAs on two PFAS compounds, PFOS and

PFOA.

28

_{These LHAs were set at 70 parts per trillion (ppt or ng/L) individually or}

combined for PFOS and PFOA. Soon after the new LHA was set by the EPA, it

was found that the water supplied by Venetucci Farm from the Widefield Aquifer

in Colorado Springs exceeded the new LHA. Owners of the farm quickly stopped

all food production since PFAS were being found in the water above the LHA

limit. The caretaker of the farm also privately sent samples of various crops and

food from the farm to a private lab, which reported that the samples had been

contaminated by PFAS compounds. This has been a significant point in time for

the farms history as being the last historical farm in El Paso County, and now all

food production on the farm has been stopped. This pollution by PFAS chemicals

has essentially shut the farm down.

Venetucci Farm sits on top of a very shallow and long aquifer called the

Widefield Aquifer. For decades, the farm watered all of it crops using water

pumped up from this aquifer. The owners of the farm also leased water to the

Colorado Springs Utilities company, which was then used to supply many homes

in the Census-Designated Place of Security-Widefield. Water from these homes

also tested above the EPA LHAs for PFOS and PFOA. Results from an official

investigation by the Air Force concluded that the use of AFFF and water

containing PFAS at a nearby Air Force base was the leading cause of

groundwater pollution with PFAS in the Widefield Aquifer.

29

17

1.6 PFAS Transport and Prediction for Venetucci Farm

Over the years, many locations and methods of PFAS release into the

environment have been identified. As previously stated, one of the major

methods of release has been from the use of AFFFs on military bases and

airports. However, there are many other pathways for PFAS to enter the

environment, such as through the waste of consumer products containing PFAS

or wastewater treatment plants discarding sludge containing PFAS.

30

_{A typical}

transport path of PFAS through the environment starts with the pollution of

groundwater, and from there makes its way through soils, sediments, surface

water, and even through the atmosphere.

31

_{An important consideration of how}

PFAS move through the environment is their characteristic hydrophobic tails and

polar heads. Because of this amphiphilic property, PFAS typically gather at

media interfaces, for example a water/soil interface or a water/air interface.

30

_In

addition to media interfaces, there are also factors within soil that also affect

PFAS transport and adsorption to soils, such as the organic carbon content of

the soil.

32-34

_{Furthermore, different PFAS will have different adsorption behavior}

in soils based on the tail length of the fluorinated carbon chain or whether the

head group is a carboxylate or sulfonate.

32

_{Overall, the research that has been}

done on PFAS transport and adsorption to sediments has thus far shown that the

amphiphilic nature of PFAS molecules and their broad range of structures makes

PFAS transport very complicated, and is still not well understood.

Venetucci Farm has been an agricultural location for many decades,

which means the soil is rich with organic carbon. Furthermore, some soils on the

18

farm have been more heavily farmed than other parts of the farm. These

attributes of the soil and farming patterns over decades can have several

implications for PFAS transport through the soils at the farm. It has been shown

that longer chain PFAS will sorb more strongly to soils with higher amounts of

organic carbon, implying that organic carbon content in the soil is a large factor of

influence on PFAS transport and adsorption.

34

_{The reason for this is that organic}

carbon in soil is generally a product of natural decay from once living organisms.

A lot of the organic carbon from this decay exists as nonpolar carbon chains that

can interact with the nonpolar tails of PFAS more strongly based on tail length.

When a PFAS has a longer tail, there is more opportunity for interaction with the

organic carbon found in a soil, leading to stronger sorption. For Venetucci Farm,

this means that PFAS are more likely to strongly adsorb to organic carbon in the

soil on the farm rather than travelling with water flow and being transported off

the farm. Also, this means that some soils on the farm are more likely to have

higher concentrations of legacy PFAS relating to how heavily the soil was farmed

and amended compared to other parts of the farm. Shorter chain PFAS

molecules do not follow the same type of adsorption chemistry as the longer

chain PFAS.

34

_{The shorter hydrophobic chain leads to a weaker hydrophobic}

interaction with any surrounding material, mainly organic carbon in soil.

Therefore, adsorption and transport of shorter chain PFAS in soils or sediments

is more strongly driven by things like electrostatic potential, the mineral

composition of the soil, and the actual chemistry of the water travelling through

the soil.

19

1.7 Experiment Parameters and Hypothesis

For this experiment, five locations spread across Venetucci Farm were

chosen to measure the concentration of several PFAS once a month over the

course of one year. The molecules listed in Table 1 are the PFAS compounds

that were monitored in this study. The sample locations ranged from a heavily

farmed location at a higher elevation on the property (Soil 1, see Figure 2), a

heavily farmed location at a higher elevation on the property further away from

Figure 2

. Map showing location of Venetucci Farm in greater Colorado area and

sample locations on the farm with Pinello (A), Soil 1 (B), Raised Bed (C), Hoop

House Basil (D), and Lower Field (E)

20

the main house (Pinello, Figure 2), raised bed soil (RB, see Figure 2), soil from

inside a small hoop house not exposed to natural precipitation (HHB, see Figure

2

), and soil that was very lightly farmed at a lower elevation on the property (LF,

see Figure 2). To obtain the soil samples, a trowel was used to dig down

approximately 3 inches beneath the surface and soil was scraped from the side

of the small hole into a 50-mL Falcon tube. Samples were then labeled and

frozen until analysis. Extraction of target PFAS was carried out using a simple

liquid extraction and clean up was carried out using solid phase extraction (SPE)

in order to isolate the analytes into a clean solvent in order to protect the

instrumentation. Quantitation of PFAS in samples was carried out using high

performance liquid chromatography coupled to tandem mass spectrometry

(HPLC-MS/MS) detector using a triple quadrupole mass analyzer.

Due to the nature of the different sample locations and history of the farm,

there are several hypotheses for PFAS concentrations in the soil over time at

Venetucci Farm. First, it is hypothesized that soil from the raised bed will have

some of the highest concentrations of legacy PFAS due to the organic carbon

rich soil, while the in-ground locations of Soil 1 and Pinello will have slightly lower

concentrations of legacy PFAS because of natural water flow through these

locations. Second, LF soil will likely have the lowest concentrations of legacy

PFAS due to having less organic carbon in the soil from less intensive farming

over the years. Third, Concentrations of shorter chain PFAS will most likely be

highest in the HHB samples because of the protection from natural precipitation

and weathering. Finally, over time, it is expected that legacy PFAS and shorter

21

chain PFAS concentrations in all sample sites will remain relatively steady over

the course of a single year, with only minor fluctuations. This predicted resistance

to changes in concentration are because of the persistent nature of PFAS

molecules mentioned earlier (the resistance to degradation in a natural

environment and their ubiquitous presence).

In order to get a more quantized prediction on the levels of PFAS that will

be found during this study, data regarding the concentration of PFAS in the well

water from Venetucci Farm and several equations can be used.

17

_{By using the}

water data gathered by UCCS previously, the estimated log K

OW

values for the

PFAS compounds, and the total organic carbon in the different sample sites,

actual concentrations can be predicted. First using the estimated log K

OW

of a

PFAS and Equation 1, the organic matter partition coefficient (K

OM

) can be

calculated.

𝐿𝑜𝑔 𝐾

𝑂𝑀

= +0.82𝐿𝑜𝑔𝐾

𝑂𝑊

+ 0.14

Equation 1

Once KOM is calculated, the relationship between fraction of organic matter

(ƒ

OM

), the solid-water distribution coefficient (K

d

), and K

OM

can be used to

calculate K

d

using Equation 2. Soil from each sample site was sent to the

Colorado State University Soil, Water and Plant Testing Laboratory in Fort

Collins, Colorado to be tested for Total Organic Carbon (TOC). This TOC is

converted to decimal form and used for

ƒ

OM

.

𝐾

_𝑂𝑀

=

𝐾𝑑

22

Finally, Equation 3 can be used to calculate the predicted concentration

that will be found in the soil at Venetucci Farm. The value for C

water

was used

from the UCCS data of PFAS concentration in well water at Venetucci Farm.

_𝐾

_𝑑

₌

𝐶𝑠𝑜𝑙𝑖𝑑

𝐶𝑤𝑎𝑡𝑒𝑟

Equation 3

Figure 3

shows the predicted concentrations of PFAS in soil at the five

sample sites chosen using equations 1-3 along with the PFAS concentration in

the well water, the estimated log K

OW

, and the fraction of organic matter in the

soil at Venetucci Farm. Because of the high TOC content of the HHB sample

site, it is estimated that this site will have significantly higher PFAS

concentrations than the other sites. It is also important to note that the predicted

Figure 3

. The predicted concentration of PFAS in soil at the five sample sites

23

PFAS concentrations in the soil is on the scale of tens of parts per billion (ppb),

or ng/g.

24

CHAPTER 2

MATERIALS AND METHODS

2.1 Chemicals and Supplies

All water used for this experiment was organic-free 18 M

Ω DI water

filtered in house using a Barnstead filtration system from Fisher Scientific. LC-MS

optima grade methanol was purchased from Fisher Scientific (A456-4). One

molar ammonium formate was prepared from LC-MS grade ammonium formate

purchased from Sigma-aldrich (70221-25G-F). LC-MS grade ammonium

hydroxide was acquired from Fluka (44273-100ML-F) and stored in the

refrigerator throughout the experiment. The ethyl acetate used for the experiment

was HPLC grade. The native PFAS stock solution used for standard preparation

and spike/recovery studies was purchased from Wellington Labs, Inc. (2 ng/µL,

913-772-4914, PFAC-MXC).

All supplies used required careful attention to the presence of any Teflon

containing parts, making sure that all supplies used contained no Teflon. This is

because Teflon is a fluoropolymer of tetrafluoroethylene. Therefore, Teflon has

the potential to strongly interact with any PFAS analytes and alter their final

concentration in the sample injected into the LC-MS/MS instrument. This can

lead to quantified concentrations of PFAS that are not actually representative of

their actual concentration in the soil sample. Polypropylene microcentrifuge tubes

(1.5 mL) were acquired from Fisher Scientific (05-408-129). Polypropylene

centrifuge tubes (15 mL), referred to as Falcon tubes, were acquired from Fisher

Scientific (05-539-12). Polypropylene centrifuge tubes (50 mL), also referred to

25

as Falcon tubes, were acquired from Fisher Scientific (05-539-8). Vials for final

sample preparation steps to store samples for injection into the LC-MS/MS

system were Wheaton 12 x 32 mm polypropylene vials acquired from Fisher

Scientific (03-336-1). Caps used for final sample vials were 11 mm polyethylene

olefin snap caps acquired from Agilent Technologies (5181-1513). Syringes used

for the experiment were various microliter scale glass syringes from Hamilton

with stainless steel plungers containing no Teflon plugs. P1000 and P300

micropipettes were also used to measure out 1 mL and 300 µL volumes of water.

Cartridges used for SPE were Oasis HLB SPE cartridges 3 cc volume with 60 mg

sorbent bed of 30 µm particle size acquired from Waters Corporation

(WAT094226). Reservoirs used to attach to SPE cartridges were 70 mL solvent

reservoirs acquired from Fisher Scientific (06-713-085). Adapters used to attach

the reservoirs to the SPE cartridges were SPE cartridge/reservoir adapters

acquired from Restek (26007). SPE was carried out using an SPE vacuum

manifold. Sonication steps were carried out using a water bath sonicator.

Centrifuge steps were carried out using a benchtop swinging bucket centrifuge

(Thermo Fisher Sorvall ST16) capable of 4000 rpm and large enough to hold the

15 mL Falcon tubes. A nitrogen gas manifold connected to a nitrogen gas tank

was used for evaporating steps.

2.2 Sample Collection and Preparation

Sample collection was carried out once a month with the following

procedure. First, sample collection sites were chosen based on physical location

on the farm to represent many different soil types across the farm. Refer to

26

Chapter 1

to see details relating to sample locations and soil types. The same

sample sites were used to collect samples for the entire duration of the study.

Samples were collected by digging down ~3 inches below the surface to

eliminate possible variability of data caused by natural weathering on the soil

surface. Then, soil was scraped from the side of the small hole into a 50-mL

Falcon tube using the cap. The Falcon tube was labeled with the site name and

date collected and then stored in a freezer until preparation and analysis.

Prior to analysis, samples were removed from the freezer and transferred

to a small Pyrex beaker. The beaker was covered with a chemwipe and allowed

to dry for several days on a lab benchtop. Two days into the drying process, any

large clumps of soil were broken apart using a stainless-steel transfer spatula to

ensure complete and even drying, and then re-covered. After samples were

completely dry, the soil was filtered using a double-layer steel mesh screen to

remove larger rocks and debris. The soil was then transferred to a 15-mL Falcon

tube and stored in a refrigerator until preparation and analysis.

Sample preparation started by making a fresh batch of calibration

standards. This was done by first creating an intermediate stock solution with a

concentration of 0.020 ng/µL from the native PFAS stock solution. To make the

intermediate stock solution 0.020 ng/µL, 10 µL of the native stock solution was

added to 1.000 mL of H2O in a 1.5 mL microcentrifuge tube

using a 1000 μL

Thermoscientific autopipette. The microcentrifuge tube was then vortexed for 10

seconds and shaken by hand, making sure the lid was tightly secured on all

sides. Once the intermediate stock solution was made, six analytical standards

27

were made with concentrations of 0.04 ng/mL, 0.10 ng/mL, 0.20 ng/mL, 0.39

ng/mL, 0.77 ng/mL, and 1.82 ng/mL. To make the calibration standards 1.000 mL

of H2O was added to six 1.5 mL microcentrifuge tubes each. Then, in order of

increasing concentration, 2 µL, 5 µL, 10 µL, 20 µL, 40 µL, and 100 µL of the

intermediate stock solution was added to the 1.5 mL microcentrifuge tubes to

make the 0.04 ng/mL, 0.10 ng/mL, 0.20 ng/mL, 0.39 ng/mL, 0.77 ng/mL, and

1.82 ng/mL standards respectively.

To begin work with the actual soil, a composite soil sample was prepared

using soil from the five different sample sites for an individual month. To make

the composite soil, two grams of soil were weighed out from each sample site

and added to a small Pyrex beaker. Once the soil had been added to the beaker

from each sample site, the soil was mixed thoroughly with a stainless steel

spatula. The composite soil was then transferred to a 15 mL Falcon tube. One

gram (measured to the ten thousandths place) of composite soil was added to

three 15 mL Falcon tubes each, labeled as “Comp_sample collection date”,

“S1_sample collection date”, and “S2_sample collection date”. S1 and S2 were

spiked with the intermediate stock solution by injecting 20 µL under the surface of

the soil in the Falcon tube and used for a spike/recovery study for the sample

batch. One gram of soil from each sample site was then weighed out individually

and added to separate 15 mL Falcon tubes

and labelled as “sample site

abbreviation_sample collection date”. One of the five sample sites was chosen

randomly using a random number generator function in Microsoft Excel to be ran

28

in the sample batch in duplicate (having two 1-gram aliquots weighed out and

included in the same sample batch).

After all the soil samples for the batch were weighed out (composite, two

spikes, five sample sites with one random sample site extracted in duplicate) the

liquid extraction phase of the method was carried out. To begin extraction, 3 mL

of a 75/25 mixture of MeOH/H2O, respectively, was added to the 15 mL Falcon

tube using a 3.00 mL volumetric pipette. The Falcon tubes were then shaken by

hand until the MeOH/H2O mixture and soil are homogenous. The Falcon tubes

were then placed into a water sonication bath that had been preheated to 55°C

and were sonicated for 15 minutes. When finished sonicating, the Falcon tubes

were transferred to a swinging bucket centrifuge and centrifuged for 5 minutes at

a relative centrifugal force (RCF) of 3011 x g. The resulting supernatant was then

transferred to a clean 15 mL Falcon tube and set aside.

The second phase of liquid extraction was then performed in the original

Falcon tubes containing soil samples. This phase began with the addition of 3 mL

of a 50/50 MeOH/H2O mixture to the soil sample. The pH of the soil/solvent

mixture was then adjusted to pH 8.5 with the addition of 15 µL of NH4OH. The

addition of NH4OH was to ensure that any remaining PFAS analytes were

deprotonated so they migrated to the extraction solvent mixture instead of

remaining in the soil. The samples were then shaken by hand again to ensure a

homogenous mixture and then added to the water sonication bath (maintained at

55°C) and sonicated again for 15 minutes. After sonication, the samples were

centrifuged again for 5 minutes at an RCF of 3011 x g. This second supernatant

29

generated was then decanted into the clean 15 mL Falcon tube containing the

supernatant from the first phase of extraction. The Falcon tubes were then

placed into an aluminum block on top of a hot plate that was maintained at 60°C

and a N2 gas evaporating manifold was used to evaporate the combined

supernatant to 2 mL (evaporating the MeOH and leaving only H2O). Before

moving on to SPE, evaporated samples were decanted into a 2-mL MCT and

centrifuged for 20 minutes at an RCF of 17000 x g. This extra centrifugation step

was to force any colloidal soil to the bottom of the MCT so a clean extraction

mixture remains.

The third phase of the extraction procedure used SPE to isolate the PFAS

analytes, another centrifugation step, and LC syringe filters to ensure the

LC-MS/MS instrument stays in good working condition. The SPE manifold was

loaded with cartridges for PFAS collection, connected to a vacuum and held at a

pressure of 5 in Hg. The cartridges were conditioned with one cartridge volume

(CV) of MeOH followed by one CV H2O. During the conditioning and actual PFAS

extraction, the drip rate was set to about one drip per 2 seconds. Once the

cartridges were conditioned, 2 mL of H2O were added directly in the cartridges

and then the cartridges were topped with 70 mL SPE reservoirs. The SPE

reservoirs were filled with 25 mL of H2O using a 25 mL volumetric pipette and the

2 mL of centrifuged sample were added to the SPE reservoir as well. The 2 mL

MCT that held the 2 mL of sample, was rinsed with roughly 2 mL of H2O and the

rinse was then added to the SPE reservoir. The sides of the reservoir were

rinsed with H2O using a Pasteur pipette to ensure any sample that had splashed

30

onto the wall of the reservoir was washed into the 25 mL of H2O. While held at a

pressure of 5 in Hg, the reservoirs and cartridges were completely drained into

the SPE manifold. After elution of all the H2O, the cartridges were dried under a

pressure of 10 in Hg for 10 minutes by removing the reservoirs, fully opening all

stopcocks with cartridges, and adjusting the vacuum line valve. SPE was finished

by eluting the PFAS analytes for collection into a clean 15 mL Falcon containing

300 µL of H2O. The H2O was used as a keeper solvent to hold the PFAS

analytes when the elution solvents were evaporated. To elute the PFAS analytes,

1 CV of MeOH was eluted followed by 1 mL of ethyl acetate.

The last phase of sample clean-up entailed evaporating the elution

solvents, normalizing the sample volumes, centrifuging again, and finally filtering

the sample. Evaporation was carried out using a N2 gas evaporation manifold

until there was roughly 300 µL left in the Falcon tube. Then, using a glass

Hamilton syringe, the total sample was drawn up (should be close to 300 µL), the

needle was wiped with a chemwipe, and then H2O was drawn up to the 500 µL

mark. This allowed us to easily normalize the sample volume to exactly 1 mL.

The 500 µL was ejected into a 2-mL MCT and the syringe was rinsed 10 times

with MeOH. Five hundred µL of H2O was then added to the same 2-mL MCT

using the same 500 µL Hamilton syringe, normalizing the sample volume to 1

mL. Once the volume had been normalized to 1 mL, the sample was centrifuged

for 20 minutes at an RCF of 17000 x g, filtered using a Sartorius Minisart RC-15

filter

(0.45 μm), and then added to a sample vial. This filtration step is paramount

to maintaining the LC-MS/MS instrument. Analytical standards were also filtered

31

and added to sample vials. An H2O blank was also filtered using the same

method and added to a sample vial to complete the full sample batch.

2.3 Instrumentation

For this study, qualitative and quantitative data were obtained using a

Shimadzu LCMS-8030 high performance liquid chromatograph with a triple

quadrupole mass spectrometer. A binary mobile phase system was used for

chromatographic separation of PFAS analytes prior to ionization and analysis

with tandem mass spectrometry. This section will cover all of the instrument

parameters as well as mobile phases used, LC column, and information about

the mass-to-charge ratios (m/z) for qualitative identification and quantitative data

for each PFAS molecule in this study.

2.3.1 Liquid Chromatography

The reverse phase liquid chromatography for this study started with

preparation of mobile phases. Mobile phases for this study were polar and

categorized as Mobile Phase A and Mobile Phase B. Mobile Phase A consisted

of 1 L of H2O with 2.5 mL of 1 M ammonium formate and 200 µL of ammonium

hydroxide. Mobile Phase B consisted of 1 L of MeOH with 2.5 mL of ammonium

formate and 200 µL of ammonium hydroxide. If there was enough mobile phase

left over from a previous sample run, then instead of making a fresh mobile

phase, ammonium hydroxide was added relative to the volume of mobile phase

remaining (for example, if there was 500 mL of a mobile phase, then 100 µL of

ammonium hydroxide was added prior to running a sample batch) and the mobile

phase was used again for the current sample batch. Once the mobile phases

32

were connected to the two mobile phase pumps, transfer lines were purged by

opening the divergence valves for each pump, and then mobile phases were

pumped through the transfer lines. Once the transfer lines were purged, a

sample injection using the high concentration analytical standard is injected to

ensure instrument is working properly, acquisition windows are accurate, and

peak shapes are well resolved. Figure 4 shows a total ion chromatogram of a

1.82 ng/mL standard for the PFAS compounds analyzed in this study.

Parameters for LC were as follows. Autosampler chamber was held at

15°C, sample was injected into the instrument in 40 µL volumes at 15 µL/sec with

a needle stroke of 52 mm. The autosampler needle was rinsed between

injections with 500 µL of MeOH at a rinsing speed of 35 µL/sec. The column

used for this study was a Waters Xbridge C18 3 µm x 5.0 x 4.6 mm (186003021)

Figure 4

. A total ion chromatogram (TIC) of a 1.82 ng/mL analytical standard for

33

and served as the non-polar stationary phase, and the column oven was held at

40°C. Mobile phases were set to a combined total flow rate of 0.4 mL/min starting

at 35% B, ramping to 95% B over 6 minutes, held at 95% for 0.1 minutes, and

ramping back to 35% B for 3.9 minutes making the total runtime 11 minutes.

Total pressure was set to a maximum of 4500 psi but changes over time, not

usually exceeding 3000 psi.

2.3.2 Tandem Mass Spectrometry

Table 2

. Mass spectrometry specifications, compounds are listed in

order of elution with PFBA being the first to elute

Compound Precursor Ion (m/z) Q1 Pre Bias (V) Product Ion (m/z) Q3 Pre Bias (V) Collision Energy (V) PFBA 212.80 18.0 169.15 18.0 10.0 PFPeA 262.90 13.0 219.05 13.0 10.0 PFBS 298.80 12.0 80.00, 99.20 12.0, 12.0 30.0, 30.0 PFHxA 312.90 12.0 269.00 15.0 10.0 PFPeS 348.90 14.0 80.00, 99.00 18.0, 18.0 35.0, 35.0 PFHpA 362.90 11.0 319.10, 169.10 10.0, 20.0 16.0, 20.0 PFHxS 399.00 12.0 80.00, 99.05 18.0, 18.0 35.0, 35.0 PFOA 412.90 17.0 369.05, 169.10 20.0, 20.0 10.0, 20.0 PFHpS 448.90 18.0 80.00, 99.05 20.0, 20.0 40.0, 40.0 PFOS 498.90 12.0 80.00, 99.05 20.0, 20.0 45.0, 45.0

The interface between HPLC and MS for this instrument was electrospray

ionization (ESI). Nebulizing gas flow rate was set to 1.5 L/min, with desolvation

line temperature set to 250°C, heat block temperature set to 400°C, and drying

gas flow rate set to 15 L/min. Argon gas was used for collision induced

dissociation (CID) and the flow rate was maintained at 230 kPa. Multiple reaction

monitoring (MRM) was used for this study with the instrument being ran in