Wm] A m. “—

 

 

1
. .
hjfiu Xvi-3‘
T-Hu‘ '
~. E ,

f

’ yr

.H‘Zlyn my;
«v :x W
J. . (¢§*k§:31, ‘5;

1'5;

E3312

 

.2%

~. .g;-‘

«

 

1;

afififiiw

:2? “:4;

.

.;. A ,

1 . “an": a
9 A‘l'

 

THESIS

This is to certify that the

thesis entitled

Fundamental Studies of Chromatographic Stationary Phases
As Selective Coatings for Mass Based Chemical Sensors

presented by

Jeffrey Paul Rasimas

has been accepted towards fulfillment
of the requirements for

M. S . degree in Chen.” Sth

Maw

flay ”Major professor”

 

Date %/h(/cl%

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

iwillilliillmiim

9300

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before data duo.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Io An Affirmative ActiorVEquol Opportunity Institution
6min

 

 

FUNDAMENTAL STUDIES OF CHROMATOGRAPHIC STATIONARY PHASES AS
SELECTIVE COATINGS FOR MASS BASED CHEMICAL SENSORS

By

Jeffrey Paul Rasimas

A THESIS
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemistry

1994

ABSTRACT

FUNDAMENTAL STUDIES OF CHROMATOGRAPHIC STATIONARY PHASES AS
SELECTIVE COATINGS FOR MASS BASED CHEMICAL SENSORS

By

Jeffrey Paul Rasimas

The selectivity of surface acoustic wave (SAW) device sensors coated with
chromatographic stationary phases has been examined. SAW responses for uncoated
(blank) and coated devices were compared with film/analyte interactions studied with in-
situ Fourier transform infrared spectroscopy (FTIR) using attenuated total reflectance
(ATR) sampling. Enhanced sensor responses were found to correlate with changes in
polymer morphology observed by FI'IR/ATR.

SAW sensors coated with chromatographic stationary phases acquire selective
properties similar to those predicted from chromatographic properties. Non-selective
stationary phases were found to concentrate analyte on the sensor and enhance its overall
response without selective behavior. Selective phases responded to all test analytes,

however enhanced responses correlating to selective interactions were observed.

To my wife Elizabeth
My parents, Janet and Carl, Leon and Carol
My grandparents, Sam, Regina, Louise, and Bill

and Mr. Eugene Overton

ACKNOWLEDGMENTS

I would like to thank my research preceptor, Dr. Jeffrey Ledford, for his unending
patience and guidance during the work that led to this publication. Our discussions were
always insightful, and the knowledge and experiences he has given me have enriched me
greatly. I must also thank my sensor research group colleagues, Greg Noonan and Ed
Townsend for their insights and assistance in this research. The remaining Ledford group
members (Per, Mike, Kathy, Mark, Paul, and Radu) also helped me with areas relevant to
this research.

This research was sponsored in part by MARTIN MARIETI‘A ENERGY
SYSTEMS, INC., and performed under subcontract 19K-MP607C for the Oak Ridge K-
25 Site, which is managed for the US. DEPARTMENT OF ENERGY under Contract
DE-AC05-84OR21400.

iv

TABLE OF CONTENTS

Page
List of Tables ............................................................................................................... vii
List of Figures ............................................................................................................. viii
Chapter 1: Introduction to the Study of Mass Sensors ................................................. 1
Overview of Chemical Sensors .............................................................. 1
Introduction to Mass Sensors ............................................................... 2
The Piezoelectric Effect
Surface Acoustic Wave (SAW) Devices
1.3. Response Modes of Polymer Coated Acoustic Wave Sensors ............... 6
Response Due to Mass Changes
Response Due to Viscoelastic Changes
1.4. Predictive Models of Sensor Response Due to Partitioning ................... 11

Boiling Point Model
Solubility Parameter Model.
Linear Solvation Energy Relationships

1.5. Overview of Chromatographic Stationary Phases as Selective

Coatings ............................................................................................... l4

Classification of Chromatographic Stationary Phases
Previous Work with Chromatographic Stationary Phases as
Selective Coatings for SAW Sensors

1.6. Experimental Methods .......................................................................... 18
Sensor Test Apparatus
Fourier Transform Infrared (FTIR) / Attenuated Total
Reflectance (ATR) Spectroscopy

1.7. Focus of Research ................................................................................ 21

1.8. References ............................................................................................ 23
Chapter 2: The Effect of Elastic Properties on the Response of Carbowax 20M Coated

Surface Acoustic Wave (SAW) Devices .................................................... 26

2. 1. Introduction .......................................................................................... 26

2.2. Experimental Methods .......................................................................... 28

2.3. Results ................................................................................................. 34

2.4. Discussion ............................................................................................ 44
2.5. Conclusions .......................................................................................... 49
2.6. References ............................................................................................ 50

Chapter 3: Studies of the Response of Tenax GC Coated Surface Acoustic Wave

(SAW) Devices ......................................................................................... 52
3. 1. Introduction .......................................................................................... 52
3.2. Experimental Methods .......................................................................... 55
3.3. Results and Discussion .......................................................................... 61
3.4. Conclusions .......................................................................................... 74
3.5. References ............................................................................................ 75
Chapter 4: Future Work .............................................................................................. 77
4.1. Comparison of Sample Introduction Methods ....................................... 77
4.2. Analyte Induced Viscoelastic Effects on SAW Responses ..................... 78
Background

Proposed Research

4.3. References ............................................................................................ 84

Table

Table 2.1.

Table 2.2.

Table 2.3.

Table 2.4.

Table 3.1.

Table 3.2.

List of Tables

Page
Probe Molecules Used for Studies of Carbowax 20M
Coated SAW Devices .............................................................................. 3O
SAW Device Response Ratios (Carbowax 20M/Blank) ........................... 36
Summary of FT IR/ATR Spectral Profiles ............................................... 43
McReynolds Numbers for Carbowax 20M ............................................. 47
Probe Molecules Used for studies of Tenax GC
Coated SAW Devices .............................................................................. 57
SAW Device Response Ratios (Tenax GC/Blank) ................................... 64

List of Figures

Figure Page
Figure 1.1. A Surface Acoustic Wave (SAW) Device. ................................................ 4
Figure 1.2. The Maxwell Model of Shear Viscoelastic Behavior. ................................ 9
Figure 1.3. Sensor Gas Handling System .................................................................... 19
Figure 2.1. Schematic Diagram of SAW Experimental Apparatus ............................... 32
Figure 2.2. Response of Carbowax 20M Coated SAW Devices at 60°C. .................... 35
Figure 2.3. FTIR/ATR Spectra for n-octane over Carbowax 20M at 60°C .................. 38
Figure 2.4. FTIR/ATR Spectra for N itropropane over Carbowax 20M at 60°C. ......... 39
Figure 2.5. FIIR/ATR Spectra for Pyrrole over Carbowax 20M at 60°C .................... 41
Figure 3.1. Schematic Diagram of SAW Experimental Apparatus ............................... 59
Figure 3.2. Response of Tenax GC Coated SAW Devices at 60°C .............................. 62
Figure 3.3. FI‘IR/ATR Spectra for Acetone over Tenax GC at 60°C .......................... 66
Figure 3.4. FI'IR/ATR Spectra for Benzene over Tenax GC at 60°C .......................... 67
Figure 3.5. FTIR/ATR Spectra for n-heptane over Tenax GC at 60°C ........................ 68
Figure 3.6. FTIR/ATR Spectra for Toluene over Tenax GC at 60°C .......................... 69

Figure 3.7. FTIR/ATR Spectra for Tetrachloroethylene over Tenax GC at 60°C. ....... 70

Figure 3.8. FTIR/ATR Spectra for n-octane over Tenax GC at 60°C. ........................ 71
Figure 3.9. FI‘IR/ATR Spectra for Thiophene over Tenax GC at 60°C. ..................... 73
Figure 4.1. [3-cyclodextrin Structure and Dimensions. ................................................ 88

Chapter 1
Introduction to the Study of Mass Sensors

1.1. Overview of Chemical Sensors.

Chemical sensors are rapidly gaining in importance and acceptance for use as in-
situ devices capable of providing real-time information about the chemical composition of
various environments. Industrial process streams, hazardous waste sites, or personal
exposure may all be monitored using chemical sensors. The ideal chemical sensor
provides chemical specificity, high sensitivity, low detection limits, reproducibility,
reversibility, durability, and rapid response times. Unfortunately, available technology has
been unable to produce a sensor that meets all these requirements. The most challenging
aspect of sensor research is the design of a species that can provide chemical specificity to
the sensor. Chemical specificity is defined as the ability of the sensor to discriminate
selectively for a single target analyte in the presence of other similar analytes. Selectivity
in chemical sensors generally arises through the use of selective coatings that interact with
analytes in specific ways.

The process of molecular recognition is considered to be the ideal level of chemical
selectivity. In a review by Thompson et al. [1], the recognition process is described as an
associative one-to-one reaction between a selective binding site and a molecule of interest.
Chemical sensor based molecular recognition requires an adsorption process taking place
on the sensor where the recognition event is then transduced into a useful signal. Since
the interactions in the recognition process occur at both the sensor surface and within the
selective coating, the transduction mechanism must be sensitive to both these events.

Chemical sensors based on optical, thermal, electrical, and mass properties have

2
been described [2-3]. Perhaps the most general transduction mechanism is based on mass.
However, the selectivity of mass-based chemical sensors remains, in the view of

Nieuwenhuizen and Venema [4], poorly studied or not studied at all."

1.2. Introduction to Mass Sensors

1.2.1. The Piezoelectric Effect

The piezoelectric effect was first reported by Jacques and Pierre Curie [5] in 1880.
They reported that when certain materials were stressed by a mechanical force an
electrical signal was produced. The converse, the application of an electrical signal to a
piezoelectric material to produce a deformation in the crystal, was also described. Since
the initial discovery of the piezoelectric effect, numerous materials that are piezoelectric
have been described. Some of the more common piezoelectric materials are lithium
niobate (LiNbO3), lead titanate (PbTiO3), lead zirconate (PbZrO3), and quartz (a-SiOz).

Materials which display piezoelectric characteristics do so because of the lack of a
center of inversion. Of the 32 known point groups, 20 lack inversion centers, however
not all 20 exhibit appreciable piezoelectric characteristics. Because of the lack of center of
inversion symmetry, application of a strain distortion leads to a macroscopic charge
separation within the solid. The strain and displacement can be equated to the stress and

electrical forces acting on the crystal by:
D = eE + dT (1.1)

S=d'e+sT (1.2)

where E is the electrical field gradient, 2 is the electrical permittivity matrix, and d' or d are

piezoelectric strain coefficient matrices. Equations 1.1 and 1.2 show that the magnitude

3

of the measurable piezoelectric effect is determined by the values of the piezoelectric
coefficients.

a-Si02 is the compound most often used for mass sensor applications. Quartz
exhibits many desirable properties for chemical sensing applications. It is physically
rugged, chemically inert, and inexpensive. Chemical mass sensors utilize both quartz
crystal microbalances (QCM) and surface acoustic wave (SAW) devices. QCMs
propagate a bulk acoustic wave through a quartz substrate, and the frequency shift
observed upon sorption of material can be related to the mass of adsorbed material. SAW
devices use surface acoustic waves propagated across the surface of a piezoelectric
material in an analogous manner. Mass sensor data presented in this thesis were collected
using SAW devices. Thus, only the fundamentals of SAW operation will be described
below.

1.2.2. Surface Acoustic Wave (SAW) Devices

Surface acoustic waves (SAWS) were first described by Rayleigh in 1855 [6].
Figure 1.1 shows a schematic diagram of a SAW device. The generation of SAWs occur
in a piezoelectric substrate when an oscillating potential is applied to a series of
interdigitated fingers (IDTs) that have been photolithographically deposited directly on the
piezoelectric surface. Deformation of the crystal lattice by the potential applied to the
fingers of the IDT causes a strain in the piezoelectric crystal, which is relieved by the
launching of a surface acoustic wave. At a distance away from the input IDTs, an
opposing set of IDTs receives the wave and converts it back into an electrical signal. The
acoustic energy of the SAW is nominally confined within one acoustic wavelength of the
piezoelectric surface, and will be affected by the surface of the propagating medium.

Therefore surface acoustic waves interact strongly with thin films deposited on the surface

. Selective
Surface Acoustic Wave Film

fl.

/g
A, /
In ut m
that.
/ /

 

 

 

 

 

 

 

 

. Piezoelectric
RF Amp lrfier Substrate
Coupler—yfi
Frequenc
Counter

 

 

 

Adapted from Frye, G. C.; Martin, S. J. Appl. Spec. Rev. 1991, 26(1-3), 75.

Figure 1.1. A Surface Acoustic Wave (SAW) Device.

5

of the device. The frequency of the surface wave is not altered by interaction with the
surface of the crystal, however changes in wave amplitude occur from the coupling of
acoustic energy from the surface into the coated layer. The phase of the wave is also
affected by a change in the wave velocity. Normal operation of SAW sensors incorporate
the device in an oscillator feedback circuit that relates the decrease in frequency to a
frequency change. SAW device resonant frequencies are a function of the spacing of the
IDT fingers. Current photolithographic techniques allow commercially available SAW
based mass sensor device operating frequencies in the 58 to 400 MHz range.

An expression that describes the fractional change in wave velocity when a thin
film is coated on a SAW device was derived by Auld [7] and modified by Wohltjen [8].

The equation is :

 

. . 4r X+u'
Af=k+k,f‘h —kf2h . , 1.3
(l -)o p 20 [Vii (X‘qu J] ( )

where Af is the change in resonant frequency due to a perturbation in the wave velocity

by the thin film, kl and R2 are substrate constants, fo is the uncoated resonant frequency,
h is the film thickness in meters, p. is the film coating density, X is the Lamé constant, it
is the film shear modulus, and VR is the Rayleigh wave velocity. If the term
(X + u. ) / (X + 211') is in the range of 0.5 to 1.0 for polymeric coatings above their glass
transition temperature, the second term can be dropped since the modulus terms are small
compared to the square of the Rayleigh wave velocity. Equation 1.3 can then be
simplified to:

Af = (k1 + k,)f§hp' (1.4)

From equation 1.4, it can be seen that Af is proportional to hp', and mass sensitivity

scales with the square of the resonant frequency.

6
Wohltjen and Dessy introduced the SAW device as a chemical sensor in 1979 [9-
11], and Wohltjen [8] described design concerns for SAW device based chemical sensors
in 1984. The operation of a SAW device was detailed, and changes in SAW frequency,
amplitude, and phases were measured. Measurement of frequency was proposed to be the

most useful for a practical device because it provides the best signal to noise ratios.

1.3. Response Modes of Polymer Coated Acoustic Wave Sensors

Acoustic wave devices are capable of providing the sensitivity necessary to be
useful transducers for chemical sensing applications. However, the bare (uncoated) quartz
surface is generally non-selective. A selective chemical sensor can be created by
modifying the surface of the acoustic wave device with a selective coating. This
modification commonly consists of covalently bound recognition sites, Langmuir-Blodgett
films, and organic or inorganic polymeric films.

In the case of polymer films, the response of the device is a function of two distinct
events. First, the added mass upon adsorption of the analyte affects the propagation of the
acoustic wave as discussed above. Secondly, the adsorbed analyte can cause the polymer
to swell or contract, which can induce changes in the Viscoelastic properties of the film.
These Viscoelastic changes also affect the propagation of the wave and lead to changes in

the amplitude, or velocity of the wave.

1.3.1. Response Due to Mass Changes

King [12] described the response of a coated quartz crystal microbalance (QCM)

in terms of the partitioning of the analyte between the gas phase and the polymer coating.

7
The partition coefficient (K) is a measure of the equilibrium between analyte in the vapor

phase and that sorbed in the stationary phase and is given by

r<=C%V (1.5)

Where Cs is the analyte concentration in the stationary phase and CV is its concentration
in the vapor phase. KGUC values are conventionally determined through gas-liquid
chromatography (GLC) specific retention volumes using a packed chromatographic
column that contains a solid support coated with a thin film of the polymer of interest
Research by Grate et al. [13] was targeted at developing correlations between
SAW sensor responses with gas-liquid chromatographic (GLC) parameters. An

expression for the partitioning occuning on the coating of a SAW device (KSAW) was

given as

M on
K =———” ”n 1.6
5“” Ach, ( )

where Afv is the frequency shift (Hz) due to analyte adsorption, Afc is the frequency shift
(Hz) due to the film coating, pfilm is the film density (g/mL), and Cv is the concentration of
analyte in the mobile phase (g/mL). A strong correlation was found between the KSAW

and KGDC data, but it was found that the KGbC values were consistently lower than stw-

8
1.3.2. Response Due to Viscoelastic Changes

The effect of changes in polymer Viscoelastic properties on the frequency response
of SAW devices has been investigated by Ballantine and Wohltjen [14]. Based on
equation 1.4, it was estimated that a 0.1 pm film of density of l g/mL should cause a
frequency shift of 337 kHz on a 158 MHz SAW. It was estimated that if polymer film had
a shear modulus (u) on the order of 107 dyne/cm2 the contribution of the second term of
equation 1.3 would be ~35 Hz. However, it was estimated that the shear modulus of films
on a SAW device could increase several orders of magnitude because the glass transition
temperature (Tg) increased due to the high frequency of the piezoelectric device. If the
shear modulus increased to 1010 dyne/cmz, the contribution of this change in Viscoelastic
properties increases to ~35 kHz. Because of this work, it was proposed that the
contribution of changes in Viscoelastic properties could contribute up to 10% of the
frequency shift expected only from mass loading effects.

Martin and Frye [15] described the use of quartz crystal microbalances (QCMs)
for use in polymer film characterization. Rigid films oscillate synchronously with the
resonator and respond solely to changes in the film mass. Viscous films oscillate
asynchronously with respect to the quartz substrate. This causes the upper regions of the
film to lag behind the film/quartz interface oscillation, which in turn induces shear
deformations in the film. These deformations, termed resonant damping, cause elastic
energy to be stored and dissipated in the film. In this case the resonator no longer
functions as a simple microbalance because the resonant damping is a function of film

thickness, density, and shear elastic properties.

9
Martin et a1. [16] found that film Viscoelastic preperties also affect SAWs. A
simple Maxwell model (Figure 1.2) was used to describe the shear Viscoelastic behavior of
polymer films. The shear viscosity (1]) was the viscous element that represented the
frictional resistance of polymer chains flowing past each other. The shear stiffness (u)
represented restoring forces arising from polymer chains seeking the most probable

configuration.

 

 

..........

Figure 1.2. The Maxwell Model of Shear Viscoelastic Behavior.

Solvent adsorption causes plasticization, which reduces the effective viscosity 1] so that
Viscoelastic properties change from elastic to viscous behavior. This transition was also
observed for the absorption of gas-phase species.

Grate et al. [17] have suggested that response of a coated SAW device is
dominated by swelling-induced modulus changes in the film. This claim developed from

data compiled in comparing KSAW and KGIJC measurements on fluoropolyol,

poly(isobutylene), and poly(epichlorohydrin) films.

10

It was observed that KSAW values were consistently higher than KGm values. Equation

1.6 was modified to give a new relationship for the calculation of KSAW values

(1.7)

The empirical factor of 4 was included in the denominator since the observed sensor
response due to analyte induced film swelling was at least three times greater than the
expected response to gravirnetric effects. By using the factor of 4 the response would be
most accurately predicted. This equation neglects the effect of varying vapor densities on
sensor response. It was proposed that polymer swelling affects the response because of a
reduction of the film modulus. It was estimated that the modulus of a rubbery polymer
could increase from 106 N/m2 to 109 N/m2 when placed on a high frequency SAW device.
The work resulted in a model that proposed that the response of polymer coated SAW
devices can be dominated by analyte induced swelling of the coated film.

Ballantine [18] has investigated the effect of film morphology on the frequency
response of polymer coated SAW devices. His work was based on a model proposed by

Grate et al. [17] where the frequency shift due to vapor sorption was represented as

 

Afv ___ (AfszKGLC J+ (MSCVKGLC )( ASAW) (18)
ps pl. 0'

where Afv is the frequency shift of the SAW exposed to a vapor of concentration CV,
KGLC is the partition coefficient determined from gas-liquid chromatography, Af‘ is the
frequency shift caused by deposition of the coating, p‘ is the density of the polymer
coating, pL is the density of the vapor, (1 is the polymer thermal expansion coefficient, and
AS AW is the frequency change of a polymer coated SAW due to thermal swelling. The

first term of equation 1.8 estimates the increase in mass due to sorption of analyte into the

1 1

film, while the second term estimates a change in frequency due to swelling of the film. It
should be noted that this is a qualitative estimate because the effects of thermal swelling
are not well established. It was found that increased film thicknesses resulted in more
dramatic Viscoelastic effects. It was calculated that a PIB film of a thickness of 1.6 mm
was sufficient to induce changes in film resonance as described by Martin and Frye [15].
Typical film thicknesses coated on 158 MHz SAWs were on the order of 0.1 pm, which
were predicted to cause a frequency increase of 27 kHz. It was found that the effect of
the change in modulus upon analyte sorption are most dramatic in thick (~0.076 um) PIB
films and less dramatic for thin («0.026 um) films.

In summary, sensor response is a convolution of effects from mass loading and
from changes in the Viscoelastic properties of the coating. Unfortunately, there is not a

model widely accepted by workers in this field.

1.4. Predictive Models of Sensor Response Due to Partitioning

Three models describing the interaction of vapor with coating material presently
exist [19]. One model was based on the frequency response of a SAW sensor being a
function of analyte boiling point. A second model attempted to correlate the SAW
response with the analyte Hildebrand solubility parameter. The third model conelated
responses with the magnitude of interaction coefficients determined from linear solvation

energy relationships (LSER).

1 2
1.4.1. Boiling Point Model

The boiling point model predicts mass sensor response from the boiling point of
the analyte. This model is the most basic of the three, and uses molecular weight, density,
vapor pressure, and vapor activity coefficients to describe interaction of analytes with
films. The model does not consider deviations from ideal vapor-polymer behavior or any
selective properties of the film or selective interactions that can occur between a film and
an analyte. This work was similar to a study described by King [20], where it was
observed that the frequency of coated QCM devices decreased with increasing analyte
boiling point. Patrash and Zellers also found that SAW response could be predicted

reasonably well using analyte boiling point [19].

1.4.2. Solubility Parameter Model

The solubility parameter model uses analyte Hildebrand solubility parameters to
predict sensor response. Hildebrand solubility parameters (8) are derived for specific

analytes through a relationship detailed in equation 1.9:

8 = (,ABH- RT)“2 / v“2 (1.9)

where 1AgH is the analyte molar vaporization enthalpy and V is the analyte molar volume.

The solubility parameter can be thought of as a qualitative measure of the cohesive energy
of the analyte. Generally, species with high solubility parameters are miscible with
molecules of similar solubility parameters. There are extensive tables of Hildebrand
solubility parameters for numerous analytes [21], which makes a model based on these
analytes attractive. However, this method also does not adequately describe the

interactions between the analyte and polymer coating.

13

1.4.3. Linear Salvation Energy Relationships

Grate and Abraham [22] described the use of linear solvation energy relationships
(LSERs) to quantify coating material solubility properties. The LSER model is based on 5
distinct intermolecular interactions: dispersion forces, dipole induction interactions, dipole

orientation interactions, and hydrogen-bonding interactions and is given as

logK = c + aa§+b B§+llogL16 +rR2+snf (1.10)

In these studies, the solute is considered to be the analyte, and the solvent is the film into
which the solute partitions. The (at;{ parameter is a measure of solute hydrogen-bond acid

strength and is derived using values of equilibrium constants for the complexation of acids

by reference bases in an inert solvent. The [31; parameter is a measure of solute hydrogen-

bond base strength and is determined using values of equilibrium constants for the
complexation of bases by reference acids in an inert solvent. The logL16 parameter
describes dispersion interactions and is a combination of the energy required to form a
cavity in the film, the energy when the vapor fills the cavity, the subsequent film
reorganization energy around the vapor, and finally, the resultant attractive interactions
between the vapor and the film. R2 is a measure of the ability of a solute to interact with a
solvent through It and 1t electron pairs. it? measures the ability of a compound to stabilize
a neighboring charge or dipole. A more detailed description of each variable and
collections of available solute parameters may be found in the extensive work of Abraham
et at. [23-33].

From expression 1.6, the KSAw values can be calculated for each analyte and film.
Once these values and the appropriate analyte constants are known, linear solvation

energy relationships (LSERs) can be developed by multiple linear regression modeling to

l4

determine the values of the constants a, b, c, l, r, and s which describe the solubility
interactions taking place between the vapor molecules and the polymer phase on the SAW
device. The magnitude of these coefficients is believed to describe the relative
contribution of each possible interaction mechanisms between a film and analyte vapor.

LSER relationships have been determined for the commonly used gas
chromatographic stationary phases [23.24.34]. The use of this body of work as a tool for
the selection of coatings for mass sensors has not been described. However, Patrash and
Zellers [19] used partition coefficients determined by SAW analysis to determine LSER
values for poly[bis(cyanoallyl)siloxane], poly(methylphenylsiloxane), poly(phenyl ether),
and poly(isobutylene) films. They also showed that the response of a test analyte could be
estimated within 25% from the LSER data.

1.5. Overview of Chromatographic Stationary Phases as Selective Coatings

1.5.1. Classification of Chromatographic Stationary Phases

GC stationary phases are typically polymeric materials with low vapor pressures,
high thermal stability, and functional groups that impart selectivity towards the gas-phase
analytes. Traditional stationary phase characterization methods involve observation of
analyte retention and peak elution profiles. Data for numerous chromatographic
stationary phases have been compiled into various tables of chromatographic indices. The

first system was proposed by Kovats [35] and discussed by Ettre [36].

15

The Kovats retention index (I) is defined as

 

1:100 [ log V... - log V...

+ 100n (1.11)
log VN‘ - log VN‘n

n+1

where VM is the net retention volume of the sample component, and VN,n+l Vs... are the

net retention volumes for two n-alkanes with total carbon numbers n and n+1 which
bracket the component. Based on equation 1.11, retention for n-alkanes scaled
logarithmically with total carbon number, and a change in 1 carbon in a hydrocarbon
homologous series results in a increase of 100 units in I. The broad acceptance of the
Kovats retention indices resulted in the characterization and tabulation of numerous
stationary phases [37-38].

Rohrschneider [39] also developed an empirical solvent scale that was based on
the splitting of polarity into various interaction forces. The scale is based on Al , which is
the retention index difference of a solute between a polar liquid phase and a standard
nonpolar phase. Squalene was used as the non-polar phase in this work. A first
approximation of AI represents the specific interaction forces that occur on the polar

phase as follows:

N=Za on)

where Bi represents the interaction energy for dipole-dipole, induction, and other forces.
A detailed derivation of the index can be found elsewhere [40].

Using the AI scheme, McReynolds [41] characterized more than 200 stationary
phases using the analytes benzene, 1-butanol, 2-pentanone, nitropropane, and pyridine. A
large McReynolds number is indicative of strong retention on a stationary phase, while

smaller McReynolds numbers suggest less retention.

”LEI

16

1.5.2. Previous Work with Chromatographic Stationary Phases as Selective

Coatings for SAW Sensors

King [12] showed that a QCM crystal coated with gas chromatographic stationary
phases became a sensitive and selective sorption detector with detection limits of 0.1 ppm
for water and 1 ppm for xylene. The QCM sorption detector at room temperature had a
linear sensor response, and the response for a squalene coated device increased with
analyte boiling point. King defined selectivity as the ability of the sensor to discriminate
between xylene and water in a stream of nitrogen.

Carbowax 20M, Apiezon grease, SE-30, and QF-l stationary phases were studied
as selective coatings for 802 [42-43]. Carbowax coated mass sensors have been reported
to be selective for aromatics, n-alkanes, alcohols, and ketones [44-46], chloroform and
toluene [47], mononitrotoluene [48], toluene in air [49], toluene and acetone in air [50]
nitrobenzene [51], and water [52]. These experiments studied the interaction of neat
analytes. It was found that mixtures of analytes caused increased sensor responses, and it
was difficult to distinguish between species using sensor responses.

Edmonds and West [47] found that sensor sensitivity and selectivity for
chloroform was related to the polarity of the coating, with Carbowax 20M being the most
sensitive, followed by dionylphthalate and Apeizon—M. Carbowaxes and poly(ethylene
glycols) are considered to be very polar phases that retain polar analytes capable of
hydrogen bond interactions. The increased sensor response reported for alkanes and
toluene contradicts this fact.

It has also been proposed that a chromatographic stationary phase can be
chemically modified to enhance its response to target analytes. Alder and Isaac [53]

investigated polyethylene glycol (PEG) coated QCM devices for use as toluene

17

diisocyanate sensors. Based on infrared spectroscopic results, a mechanism that involved
a slow, reversible interaction between the hydroxyl group of the polymer and the
isocyanate group was proposed. Chemical modification of the poly(ethylene glycol) film
was examined as a method to tune the selectivity to the target analyte while reducing the
response to water interferences The PEG was reacted with thionyl chloride, which
resulted in the chlorination of the hydroxyl endgroups. The modified film response to
toluene diisocyanate was found to be reversible and faster, while the response to humidity
was markedly decreased.

Ballantine et al. [54] correlated SAW sensor responses with solubility properties
and chemical properties using pattern recognition techniques. The interaction of 11
analytes including organic vapors, water, and toxic organophosphorous compound
analogs were studied on eleven films. Sensitivity to dimethyl methylphosphonate
(DMMP) was markedly increased on most films, and was proposed to be due to a dipolar
or polarizibility interaction. Rose-Pehrsson et al. [55] and Grate et al. [56] continued
working on the detection of hazardous vapors using pattern recognition techniques. The
responses of dimethyl methylphosphonate (DMMP) and N-N-dimethylacetamide (DMAC)
were distinguishable from lower boiling point interferences from 2-butanone and 1-
butanol. The interactions of these species was proposed to involve hydrogen bonding
between analytes and the film. Grate et al. [56] studied the responses of coated SAW
devices to trace organophosphorous and organosulfur vapors. A portable system for
automated sample analysis was described that included a thermal desorption polymeric
pre-concentration traps, an array of coated SAW devices, and a computer to perform

pattern recognition analysis.

1 8
1.6. Experimental Methods

1.6.1. Sensor Test Apparatus

There are two commonly used techniques to expose the acoustic sensor to the
analyte of interest Analyte can be introduced to the sensor cell as either a constant stream
or as a single pulse. A sensor that uses a continuous analyte stream relies on equilibrium
between the vapor phase and the coated film before measurements are made. Responses
in this mode are typically square shaped with the equilibrium occuning over several
minutes. Pulsed analyte delivery provides a narrow plug of analyte that interacts with a
film and causes spiked response profiles. The shape of the peak can be used to infer
interaction between the film and the analyte. Peaks that are slow retuming to the original
baseline (tailing) indicate slow diffusion of analyte from the film or strong interactions
between the analyte and the film. Sharp, symmetric peaks indicate minimal interaction,
and a peak that does not return to the original baseline can indicate irreversible bonding or
changes in the film properties. A schematic of the gas system and sensor cell used in this

work appears in Figure 1.2.

1.6.2. Fourier Transform Infrared (FTIR) / Attenuated Total Reflectance (ATR)

Spectroscopy

Fourier transform infrared (FTIR) spectroscopy is used to observe interactions
between analyte and polymer, infer modes of analyte/polymer interactions, and observe
changes in polymer morphology. Traditional FITR spectroscopy sampling techniques
involve transmission sampling. Sample preparation techniques for transmission infrared

studies of thin films include cast films, pressed pellets, and Nujol mulls. A more recent

l9

 

 

 

 

 

 

 

 

 

 

 

 

0 Injection Syringe
8 l
M .
S Heated GC Injector
Heated Transfer Line

N 2 M Test Cell

F

C

 

 

 

 

 

 

 

 

 

Figure 1.3. Sensor Gas Handling System.
MS = Molecular Sieve, MFC = Mass Flow Controller

20
sampling technique useful for the study of thin films is attenuated total reflectance (ATR)
sampling. Attenuated total reflectance sampling has been described by Fahrenfort [57]
and Harrick [58-59].
Total internal reflection occurs when light traveling through an optically dense
medium reaches a boundary with an optically less dense medium. If the angle of incidence
(6i) is greater then the critical angle (Go) all of the incident radiation will be internally

reflected. The critical angle is determined by

sin6c=n%l (1.13)

where nland n2 are the refractive indices of the optically dense and optically less dense
medium, respectively. The medium in which the beam propagates is called the internal
reflection element (IRE). At each reflection, an evanescent wave propagates a short
distance into the less dense optical medium, interacts with the coated film, and is
attenuated by absorption. Multiple interactions between the main beam and a probed
sample increase the observed signal-to-noise (S/N) ratio. The penetration depth (dp) of
the evanescent wave is a function of wavelength and refractive indices of the IRE and

sample and is calculated by

d = 7» (1.14)

p 21t(sin2(x - (112 Inf)“2

 

where tip is the penetration depth of the evanescent wave at wavelength 7)., sinzot
describes the angle of incidence between the infrared beam and the IRE, and n, and 112 are
the index of refraction for the IRE and the sample. The ZnSe IRE used in this work has
an index of refraction of 2.42. If a sample has an index of refraction of 1.5, the

penetration depth of the evanescent wave is approximately is 0.32 pm at a frequency of

 

21
1500 cm‘l. If the penetration depth of the evanescent wave is the close to the film
thickness, observation of interactions at the interface between the film and its surrounding
environment are possible.

FTIR/ATR has been used to examine the macroscopic and surface structures of
thin film polymers [60-62]. Information regarding chain structure, film morphology, and
bonding modes can be obtained from conventional ATR analysis [63-66]. Rates of
diffusion of liquid samples into polymer films have also been measured [67,68], and
diffusion of gaseous and liquid analytes in polymer matrices have been observed [69]. In
addition polymer coated internal reflectance elements have been used to concentrate a
solute in aqueous solution for analysis by FI'IR [70]. In all of the above references, the
researchers concentrate on the spectrum of the analyte and not the impact of the analyte

on the structure or absorbance of the polymer.

1.7. Focus of Research

The goal of this research is to observe the response of chromatographic stationary
phases as coatings on SAW sensors. A wide range of analytes chosen to represent various
functional groups and a wide range of boiling points will be used to compare the sensor
responses with known chromatographic behavior.

In order to fully understand the mechanism of SAW device response, a thorough
investigation into the fundamentals of film/analyte interactions is needed. This research
will focus on understanding the fundamental processes occuning between the film and
analyte. Characterization of chromatographic stationary phases with similar functionality
and/or similar backbone structures using SAW sensor measurements, in-situ Fl‘IR/ATR

experiments, and linear solvation energy relationships (LSER) will be performed.

22

The effect of changes in polymer Viscoelastic properties upon the sorption of
analyte will also be examined by a comparison of SAW sensor responses and in-situ
infrared spectroscopic experiments. An understanding of the effects of Viscoelastic
changes in the polymer film on the observed sensor response will allow examination of the
role and limits of film selectivity on mass sensor response.

When all experiments are compiled, compared, and contrasted, correlations
between SAW sensor response and the macroscopic and molecular properties of analytes
and analyte/film interaction will be drawn. This information will be used to better
understand the observed responses of polymer coated SAW sensors in real-world sensing

situations.

WEE-em 3

1.8.

99999!"

10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.

22.
23.

24.

25.

26.

27.

28.

23

References

Thompson, M.; Frank, M. D.; Heckl, W. M.; Marassi, F. M.; Vigmond, S. J.

Chemical Sensor Technology, Vol. 2 T. Seiyama ed., Elsevier, Amsterdam, 1989,

237-254.

Edmonds, T. E., ed. Chemical Sensors Chapman and Hall, New York, 1988.

Janata, J. Principles of Chemical Sensors Plenum Press, New York, 1989.

Nieuwenhuizen, M. S.; Venema, A. Sensors and Materials 1989, 5, 261.

Guilbault, G. 6.; Jordan, J. M. CRC Crit. Rev. Anal. Chem 1988, 19, l.

Rayleigh London Math. Soc. 1885, 7, 4.

Auld, B. A. Acoustic Fields and Waves in Solids Ch. 12, Wiley-Interscience, T
New York, 1973.

Wohltjen, H. Sensors and Actuators 1984, 5, 307.

Wohltjen, H.; Dessy, R. Anal. Chem. 1979, 51, 1458.

Wohltjen, H.; Dessy, R. Anal. Chem 1979, 51, 1465.

Wohltjen, H.; Dessy, R. Anal. Chem. 1979, 51 , 1471.

King, W. H., Jr. Anal. Chem 1964, 36, 1735.

Grate, J. W.; Snow, A.; Ballantine, D. 8., Jr.; Wohltjen, H.; Abraham, M. H.;
McGill, R. A.; Sasson, P. Anal. Chem 1988, 60, 869.

Ballantine, D. 8., Jr.; Wohltjen, H. IEE Ultrason. Symp. 1988, 559.

Martin, S. J.; Frye, G. C. IEEE Ultrason. Symp. 1991, 393.

Martin, S. J.; Ricco, A. J.; Frye, G. C. 1990 Solid State Sensor & Actuator
Workshop, ACS Spring Meeting, San Francisco.

Grate, J. W.; Klusty, M.; McGill, R. A.; Abraham, M. H.; Whiting, G.; Andonian-
Haftvan, J. Anal. Chem 1992, 64, 610.

Ballantine, D. S., Jr. Anal. Chem. 1992, 64, 3069.

Patrash, S. J .; Zellers, E. T. Anal. Chem. 1993, 65, 2055.

King, W. H., Jr. Anal. Chem 1964, 36, 1735.

Barton, A. F. M. Handbook of Solubility Parameters and Other Cohesion
Parameters CRC Press, 1991.

Grate, J. W.; Abraham, M. H. Sens. & Act. B. 1991, 3, 85.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chem. Soc.
Perkin. Trans. 2 1990, 1451.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chromatogr.
1990, 518, 329.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor, P. J. J. Chem.
Soc. Perkin. Trans. 2 1990, 521.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Duce, P. P.; Morris, J. J .; Taylor, P.
J J. Chem Soc. Perkin. Trans. 2 1989, 699.

Abraham, M. H.; Duce, P. P.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor, P.
J. Tetrahedron Lett. 1989, 29, 1587.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Monis, J. J.; Taylor, P. J .; Laurence,
C.; Berthelot, M. Tetrahedron Lett. 1989, 30, 2571.

29.

30.

31.

32.

33.
34.

35.
36.
37.
38.

39.
40.

41.
42.
43.
45.
46.
47.
48.
49.
50.
51.
52.

53.
54.

55.

24

Abraham, M. H.; Grellier, P. L.; McGill, R. A.; Doherty, R. M.; Kamlet, M. J.;
Hall, T. N.; Taft, R. W.; Carr, P. W.; Koros, W. J. Polymer 1987, 28, 1363.
Abraham, M. H.; Grellier, P. L.; McGill, R. A. J. Chem Soc. Perkin. Trans. 2
1988, 523.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Duce, P. P.; Morris, J. J.; Taylor, P.
J J. Chem Soc. Perkin. Trans. 2 1989, 699.

Kamlet, M. J.; Abboud, J-L, M., Abraham, M. H.; Taft, R. W. J. Org. Chem
1983, 48, 2877.

Patte, F.; Etcheto, M.; Laffort, P. Anal. Chem 1982, 54, 2247.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chromatogr.
1991, 587, 229.

Kovats, E. Helv. Chim. Acta 1958, 41, 1915.

Ettre, L. S. Anal. Chem. 1964, 36, 31A.

Schupp, O. E. ed. Compilation of Gas Chromatographic Data ASTM,
Philadelphia, 1967.

McReynolds, W. 0. Gas Chromatographic Retention Data Preston Publishing,
Evanston, IL., 1968.

Rohrschneider, L. J. Chromatogr. 1966, 22, 6.

Karger, B. L.; Snyder, L. R.; Horvath, C. An Introduction to Separation Science
Wiley and Sons, New York, 1973, pp. 211-246.

McReynolds, W. O. J. Chrom. Sci. 1970, 8, 685.

Guilbault, G. G.; Lopez-Roman, A. Environ. Lett. 1971, 2, 35.

Janghorbani, M.; Freund, H. Anal. Chem 1973, 45, 325.

Karasek, F. W.; Gibbins, K. R. J. Chromatogr. Sci. 1971, 9, 535.

Karasek, F. W.; Guy, P.; Hill, H. H.; Tiemay, J. M. J. Chromatogr. 1976, 12,
179.

Karasek, F. W.; Tiemay, J. M. J. Chromatogr. 1974, 89, 31.

Edmonds, T. E.; West, T. S. Anal. Chim. Acta 1980, 117, 147.

Tomita, Y.; Ho, M. H.; Guilbault, G. G. Anal. Chem 1979, 51, 9.

Ho, M. H.; Guilbault, G. G.; Reitz, B. Anal. Chem 1980, 52, 1489.

Am, D.; Amati, D.; Blom, N.; Ehrat, M.; Widmer, H. M. Sensors and Actuators B
1992, 8, 27.

Sanchez-Pedreno, J. A. 0.; Drew, P. K. P.; Alder, J. F. Anal. Chim. Acta 1988,
207, 285.

King, B. MSc Thesis University of Manchester Institute of Science and
Technology, 1984.

Alder, J. F.; Isaac, C. A. Anal. Chim Acta 1981, 129, 163.

Ballantine, D. 8., Jr.; Rose, S. L.; Grate, J. W.; Wohltjen, H. Anal. Chem 1986,
58, 3058.

Rose-Pehrsson, S. L.; Grate, J. W.; Ballantine, D. 5., Jr.; Jurs, P. C. Anal. Chem.
1988, 60, 2801.

56.

57.
58.
59.
60.
61.

62.
63.
64.
65.
66.
67.
68.

69.

70.

25

Grate, J. W.; Rose-Pehrsson, S. L.; Venezky, D. L.; Klusty, M.; Wohlthen, H.
Anal. Chem 1993, 65, 1868.

Fahrenfort, J. Spectrochim Acta 1961, 17, 698.

Harrick, N. J. Ann. N. Y. Acad. Sci. 1963, 101, 928.

Harrick, N. J. J. Opt. Soc. Amer. 1965, 55, 851.

Holland-Moritz, K.; Siesler, H. W. Appl. Spectosc. Rev. 1976, 11, 1.

Smith, A. L. Applied Infrared Spectroscopy John Wiley & Sons, New York,
1979.

Fieldson, G. T.; Barbari, T. A. Polymer 1993, 34, 1146.

Ito, M.; Pereira, J. R. C.; Porter, R. S. J. Polym. Sci, Polym. Lett. 1982, 20, 61.
Kuhn, K. J .; Hahn, B.; Percec, V.; Urban, M. W. Appl. Spectrosc. 1987, 41, 843.
Popli, R.; Dwivedi, A. M. J. Appl. Polym Sci. 1989, 37, 2469.

Evanson, K. W.; Urban, M. W. J. Appl. Polym Sci. 1991, 42, 2287.

Xu, J. R.; Balik, C. M. Appl. Spectrosc. 1988, 42, 1543.

Semwal, R. P.; Rao, N. B. S. N.; Hajela, B. P. Spectrochim Acta 1992, 48A,
1045.

Heinrich, P.; Wyzgol, R.; Schrader, B.; Hatzilazaru, A.; Lubbers, D. W. Appl.
Spectrosc. 1990, 44, 1641.

Meuse, C. W.; Tomellini, S. A. Anal. Lett. 1989, 22, 2065.

 

Chapter 2

The Effect of Elastic Properties on the Response of
Carbowax 20M Coated Surface Acoustic Wave (SAW) Devices

2.1. Introduction

The mass sensitivity, ruggedness, and size of piezoelectric based mass sensors
makes them attractive for use as chemical sensors. One type of device based on surface
acoustic waves (SAW) has mass detection limits on the order of 1 ng with signal-to-noise
ratios of greater than 20:1 [1]. The frequency response of these devices has been
determined to be a function on both mass loading and film elastic properties and is given

as follows [2-3]:

 

_ 2 ._ 2 4u'{7t'+u'
Af-(k1+k2)fohp worth; 04211)) (2.1)

where Af is the change in resonant frequency (response), k1 and k2 are substrate
constants, fo is the uncoated resonant frequency, h is the film thickness, p' is the film
density, X is the Lame constant, it. is the film shear modulus, and VR is the Rayleigh wave
velocity. The first term describes the mass loading effects on the device response and is a
function of added mass since the product hp' is the mass per unit area on the device. The
term (X + 11.)“); + 211') describes film elastic properties. If it is in the range of 0.5-1.0,
and the polymer is above its glass transition temperature, it is usually ignored since the
modulus terms are small compared to the square of the Rayleigh wave velocity. The
inclusion of material constants for quartz and the dropping of the second term reduces
equation (2.1) to a widely used working form given as:

Af = -1.26 x 106 fjhp (2.2)
26

27

Because of the difficulty in predicting and determining the effects of the second
term of equation (2.1), equation (2.2) has been widely used to determine the mass
response of polymer coated SAW devices. However, numerous workers noted that a
change in polymer Viscoelastic properties affects the propagation of acoustic waves in
piezoelectric mass sensors [2-11]. Work by Grate et al. [6] correlated SAW device
responses with analyte induced changes in the modulus of the selective polymer coating.
This resulted from correlating SAW device response data with data taken from
partitioning data as observed using gas-liquid chromatography, which was being studied
for use as a predictive tool to characterize the SAW device response properties of
polymers. They found that the response observed using GLC correlated well with that
observed on SAW devices, however the responses of the polymer coated SAW devices
were consistently higher than observed from the GLC data. Based on this increase, it was
suggested that the sorption of analyte vapor into the polymer coated on the SAW device
decreased the modulus of the film. By decreasing the modulus of the coating, the resonant
frequency of the SAW device decreases due to damping of the SAW into the softer film.
This decreased modulus resulted in a frequency shift that was larger than could be
explained simply by mass loading effects. These results suggested that the second term of
equation (2.1) cannot be ignored. In fact, they proposed that analyte induced swelling of
the polymer affected the propagation of the SAW to an extent that had not been
previously described.

In order to determine the effect of analyte sorption on changes of the Viscoelastic
properties of coatings used for SAW devices and the corresponding effects on SAW
device responses a film whose selectivity characteristics as both a coating on SAW devices
and in bulk techniques, is needed. A film that fulfills these criteria is Carbowax 20M. It is
a widely used gas chromatography stationary phase that separates analytes on the basis of

analyte molecular polarity and through hydrogen bond interactions. Because of its high

28

polarity and chromatographic selectivity towards polar compounds, Carbowax 20M has
also been studied as a selective coating for acoustic wave devices. Carbowax coated
devices have been reported to be selective for xylenes [12], chloroform [13], nitrobenzene
[14], styrene [15], toluene [16], mononitrotoluene [17], as well as acetone and chloroform
[18]. It is important to note that much of the previous work involved studying the
selectivity under careful controlled conditions using a limited set of target analytes.
Additionally, some of these selectivity claims [12,15-16] contradict selectivity predicted
from chromatographic properties because the compounds are non-polar and unable to
interact with Carbowax through hydrogen bonding.

The goal of this work is to study the interactions that occur between a Carbowax
20M coated SAW device and a series of 18 target analytes that were chosen so that a
variety of chemical functional groups and a range of analyte boiling points could be
examined. A pulsed injection system was used to introduce narrow gas phase analyte
pulses into a thermostatically controlled test cell. The response of the coated SAW
devices were compared to data from in-situ Fourier transform infrared spectroscopy
(FTIR) using attenuated total reflectance (ATR) sampling experiments which provided
insight into the effect of changes in polymer morphology on the response of the SAW
device. These comparisons will provide a better understanding of the effect of analyte
induced changes in film Viscoelastic properties on the response of polymer coated SAW

devices.

2.2. Experimental Methods

Reagents. Carbowax 20M (MW=20,000) was obtained from Aldrich Chemical
Company and used as received. The set of probe analytes used in this work is given in
Table 2.1 and includes five compounds used by McReynolds [19] to characterize

chromatographic stationary phases (benzene, butanol, 2-pentanone, nitropropane,

29

pyridine), a homologous series of hydrocarbons (n-hexane, n-heptane, n-octane, n-
nonane), and aromatic, alcohol, and aromatic heterocycles. Also included in Table 2.1 are
solvatochromic parameters for these analytes. These parameters are used to describe the
relevant molecular interactions that are possible for a specific analyte through hydrogen
bond donation (or), hydrogen bond acceptance (B), dispersion interactions (logL‘é),
electron donation/acceptance (R2), and charge stabilization (it? ). A more detailed
description of each variable and collections of parameters may be found in the extensive
work of Abraham et al. [20-34]. All probe analytes (>99% purity) were obtained from
Aldrich Chemical Company and used as received.

Surface Acoustic Wave (SAW) Device. SAW results were obtained using 200
MHz resonator devices manufactured by Femtometrics, Inc. (Costa Mesa, CA). The
devices are packaged in an array configuration that consists of six separate resonators
permanently mounted on one header that plugs into a motherboard containing the
individual resonator electronics. A separate reference oscillator is contained on the
motherboard. The 200 MHz resonators exhibit a resonant frequency shift of
approximately 904 Hz when perturbed by a surface mass change of 1 nanogram [35].

The SAW array and reference chips were plugged into a motherboard that
provides power, oscillating potential, and feedback circuitry to the sample and reference
devices. Data acquisition and processing was accomplished using an IBM PC compatible
computer and control software (Freq6., Femtometrics Inc.). The signals acquired were:
the individual resonator beat frequencies, which is a difference between the reference and
individual resonator frequencies. The data files for the array were then further processed
using an in-house developed peak processing routine under LabView (National

Instruments, Inc., Austin, TX).

 

 

30

 

 

2&5

cod

8.:
Ed
and
N3
8.:
cod
.2 .o
cod
and
$0
8.:
:3
3.:
8.:
we:
2.:
N:
92333:

M: .v cod
0E:::>< 82
mod cad
ww.m cod
mwd Rd
com wed
and mod
33“ 3 .3
Sum med
5 .m cod
cod mad
cod 2 d
on; and
5d cad
SN 3 .o
5N cod
5.6 $6
omd CW:

24 2
0:260:08— fiw:0.=m 000m
Some—305 “Eeni

cod

cod
cod
cod
hmd
cod
cod
cod
cod
cod
cod
mmd
cod
cod
cod
med
wed

505% 23..
28-:

#3
fimfi
02
#3
OS
w:
c:
_:
2:
we
E
SW
mm
5
em
00
no
cm

05%:

0530::
Beta
0:88-:
053508020050:
0:305:58
353.5;
053.3
0:038
0:053:09m
0:33;:
0535080205
0:23.03:
6:03:32
0:80:20?
0:3:0:
0:30;:
3550::
0:900:

@8309 22% BSD EON $33.80 go 865% :8 033 830032 0305

CF‘NMV‘WOI‘OO
v—rv—tv—tv—tv—rv—tv—rv—rv—r

NMVWOFOOG

v—t

 

AN 035.

31

Prior to coating, SAW resonators were cleaned by solution soaking and ozone
cleaning. The solution soaking involved rinsing the device with dry acetone, followed by
separate 30 minute soaks in dry toluene and electronic grade isopropanol. The device was
then allowed to dry in air. The dry device was then treated for 30 minutes in an ozone
cleaner (UV Ozone Photoreactor PR-100, UVP Inc., New Jersey), followed by a 30
minute isopropanol soak. For the final drying step, the device was kept in a dessicator.

Films for use in surface acoustic wave (SAW) experiments were spray cast from 1
mg/mL chloroform solutions using an air brush with dry nitrogen (99.95%, AGA Gas Co.)
as a canier gas. The films were deposited while the sensor device was oscillating to allow
observation of the frequency shift due to the mass of film deposited. Typical SAW device
coatings resulted in a 150 kHz frequency shift, which equals approximately 166 ng. The
film thickness was approximately 100 A

Sensor Test Apparatus. A schematic diagram of the experimental apparatus is
shown in Figure 2.1. The sensor test cell was machined from aluminum and has a volume
of ~8 mL. The lid and body were joined using stainless steel bolts and sealed with a thin
Teflon gasket. Electrical feedthroughs allowed the SAW device to be placed inside the
cell, with the feedthrough pins making electrical connection to the oscillator circuit.
Analytes were introduced into the system by syringe injection and reached the sensor
device as a narrow plug. Operating conditions were 40 mUmin of dry N2 flow (99.95%,
AGA Gas Co.) throughout the system, sample injector temperature of 250°C, and a test
cell temperature of 60°C. The temperature of the test cell was controlled by heating the
cell transfer line with nichrome wire and was monitored using a K-type thermocouple
positioned in the center of the cell. The injection volume used for each analyte was varied
depending on the magnitude of sensor response so that the maximum frequency difference
was below 300 kHz. At least six injections of each analyte were made to insure
reproducibility, and analyte injection order was varied to minimize the effect of analyte

order on the sensor response. Three different Carbowax 20M coated SAW devices were

32

 

 

Gas Handling
System

 

Teflon

Gasket \/

 

 

 

Injector

 

 

 

\ Aluminum

 

 

 

 

// / Blocks

 

 

 

 

 

 

Inlet

 

 

 

Figure 2.1. Schematic diagram of SAW experimental apparatus.

 

33

examined. Observed Carbowax 20M SAW device frequency responses typically had
relative standard deviations of less than 10%.

The frequency of the SAW devices was measured during exposure to analyte
vapor plugs. The observed frequency shift was then used to calculate the response of the
SAW device based on observed frequency shift (Hz) per umole of analyte injected, which
normalizes the responses. We plotted SAW response (Hz/umole) versus analyte boiling
point (°C), which allows observation of analyte condensation on SAW device responses.

Fourier Transform Infrared Spectrosc0py ( FT IR) Analysis. A Mattson Galaxy
5020 FTIR spectrometer equipped with a wideband mercury cadmium telluride (MCT)
detector was used to collect IR spectra of analytes, polymer films, and analytes sorbed in
films. The spectrometer is equipped with a modified attenuated total reflectance (ATR)
accessory (Specac, Inc.). The ATR accessory is a HPLC detector that has been modified
to perform analyte-filth studies. The attachment consists of a 500 uL flow thru cell in
contact with a ZnSe internal reflection element (IRE) (Specac Inc.) contained in a
temperature controlled water bath held at 60°C. The flow-thru fittings of the cell have
been changed to accommodate 1/4" stainless steel inlet and outlet fittings that are
connected to a gas handling system. The inlet of the ATR cell is connected to a sample
injector with a 1/4" stainless steel transfer line. The injector and transfer line are heated to
250°C. Analytes were injected as 1.0 11L pulses into a 20 mUmin flow of dry nitrogen
(99.95%, AGA Gas Co.) FTIR/ATR spectra were collected with 4 cm‘1 resolution. A
software package supplied by Mattson (FIRST Macros) was modified to collect time
dependent IR data. A spectrum of the coated ATR element is used as the background for
further data collection. The modified software collects eighty spectra (each spectrum is
the sum of 10 scans taken at one second intervals) following the injection of each analyte.
Using this data collection method, a decrease in peak absorbance (i.e., a “negative” peak)

suggests that the analyte affects the properties of the original coating by decreasing the

34

amount of film being observed. An increase in peak absorbance (a “positive” peak)
indicates that the analyte/film interaction involves increasing the amount of film being

observed.

2.3. Results

SAW Response Magnitudes. The variation in sensor responses as a function of
analyte boiling point for a Carbowax 20M coated SAW device at 60°C is shown in Figure
2.2. The average response of three uncoated devices with the analyte set (open circles,
dotted line) and similar data for Carbowax coated devices (closed circles, solid line) are
plotted on the same axes. The response for the uncoated device generally increases in a
smooth trend with increasing analyte boiling point, with responses to toluene (#11),
pyridine (#12), l-butanol (#13), nitropropane (#14), and tetrachloroethylene (#15)
forming a spike in the smooth trend. The response for the Carbowax coated SAW devices
(closed circles) exhibit deviations from the trend observed for the blank devices. Analytes
exhibiting increased response relative to the uncoated device are pyrrole (#17), and
nonane (#18). The responses to n-hexane (#3), n-heptane (#9), and n-octane (#16) were
very close to those observed on the blank devices. With the exception of n-hexane, n-
heptane, and n-octane, analyte responses were always greater on the Carbowax 20M

coated SAW devices that they were on the blank devices.

Another way to look at this data involves using response ratios (Table 2.2). We

use a response ratio that is the response of the Carbowax coated SAW divided by the

35

 

 

Resptrse (Hz/ undo)

 

 

 

 

B n. P 0 :‘c
Figure 2.2. Response of Carbowax 20M coated SAW (0) and blank SAW (0) devices
at 60°C.

36

 

 

ms :2 3: 05.89: 2
mm we 9.: saga 2
o: «a. a9. 23.8.: E
2. E Rom 8238520802 a
E 5.: $9. 05.885: 3
mm mm: Be. 6533.: 2
E w: : $9. 26:3 2
em 2 m SS 0522 _ _
Q cam wee acousaas 2
5 mm: SN 823;-.. a
we. we we. 82308253 w
as m: m: _ 828:: e
an a me. 35:88.: 0
3. .9. S. 05.6535 m
B me E. 0528 a.
S 3. mm 052.: m
3 z. m: 355...... N
am 3. m: 28.8.. :
35:38 5.3238 eeesfiE Eoeéfiv 9:39.. Benz semi
ovum 00:0am0m 0m:o%0m 22m 00:303. >35
as: 22 5.38.8

meanw— 0m:e%0m 00m>0Q >><m .N.N 030,—.

37
response of the blank SAW. With the exception of pyrrole, the average analyte response

ratio is approximately 4.5. Analyte responses for n—hexane, n—heptane, and n-octane were
all close to 1, which indicates that the Carbowax coating does not enhance the response of
the SAW device for these analytes. However, the response ratio for n-nonane is 7.5,
which is a dramatic enhancement over ratios observed for other n-alkanes. Additionally, a

dramatic enhancement is observed for pyrrole, with its response ratio being 23.

F TYR/ATR Studies. Time dependent spectra for gas-phase analytes interacting
over Carbowax 20M were collected using a background that was of the Carbowax coated
IRE so that time dependent features of the analyte interacting with the film can be
observed.

An example of an analyte interacting then desorbing from the film is shown in
Figure 2.3. This is the profile observed for a 1 11L pulse of n-octane vapor over a
Carbowax 20M film. The profile is expanded between 900 and 1800 cm’1 to show
spectral detail in this region. The lower spectra are references for Carbowax 20M (a), and
n-octane (b). At eleven seconds (c) and 22 seconds (d) after analyte injection n-octane is
not present in the film. Spectrum ((1) is a flat spectrum identical to the initial background
(Carbowax 20M film), which suggests that interaction of n-octane with this film has not
affected the background features.

Another type of profile is shown in Figure 2.4, which is for a 1 [11. pulse of
nitropropane over a Carbowax 20M film. The lower spectrum (a) is a reference spectrum
of a Carbowax film, while spectra (b) and (c) represent the film during and after
interaction with the nitropropane. Spectrum (b) was taken 23 seconds after sample

injection and shows positive features at 1553, 1468, 1445, and 1390 cm‘1 that are

38

 

l l l l

(d)

 

I 0.01 AU

(C) 10.01 AU

0.25 AU
(b) i
O 5 AU
(3)11M
l

 

 

 

 

1800 1600 1400 1200 1000
Frequency (cm‘ 1)

Figure 2.3. FTIR/ATR spectra for n-octane over Carbowax 20M at 60°C.
(a)=Carbowax 20M reference, (b)=n-octane reference,
(c)=11 seconds, (d)=22 seconds.

39

 

 

 

 

 

 

 

 

 

 

l l r I
T
l 002AU

(Q) L

l T

1 1

l

1 01AU

‘ i

z (b) L

I ,

l

l

l .

1 k05AU \JJK

{ .

(a) LAMA)

1 I l 1
3000 2000 1000

Frequency (cm‘ 1)

 

Figure 2.4. FTIR/ATR spectra for nitropropane over Carbowax 20M at 60°C.
(a)=Carbowax 20M reference, (b)=23 seconds, (c)=68 seconds.

40

attributed to nitropropane. The features at 2980 and 2861 cm'1 correspond to Carbowax
features seen in spectrum (a). Spectrum (0) was taken 68 seconds after injection, and
shows positive features at 1265, 1086, 1026, and 799 cm'1 that correspond to features
observed in the Carbowax spectrum (a). Additionally, the features that correspond to
nitropropane have disappeared, which suggests that nitropropane has desorbed from the
film, and the fact that this took 68 seconds suggests a stronger analyte/film interaction
than was observed for n-octane. The features are all shifted to a greater frequency than
observed in the reference spectrum, and the peak pattern is similar to that observed in the
Carbowax 20M reference spectrum. A similar overall profile was also observed for a 1 uL
pulse of pyridine over the Carbowax 20M film. ’

Figure 2.5 is for a l 11L injection of pyrrole over Carbowax 20M with the scale
expanded to show relevant features of both the analyte and film. The lower spectrum (a)
is a reference spectrum of Carbowax 20M, while spectrum (b) is a reference spectrum of
pyrrole. Spectrum (c) was taken 22 seconds after sample injection and numerous features
are present. The features at 1080, 1057, 1020, and 733 cm'1 are attributed to pyrrole.
Features attributed to Carbowax 20M are not present in this spectrum. Spectrum ((1) was
taken 91 seconds after injection, and features at 1265, 1091, 1026, and 799 cm“1 which are
attributed to the Carbowax 20M film are present.

Figure 2.6 is for a 1 1.1L injection of n-nonane over Carbowax 20M with the scale
expanded to show relevant features of the analyte and the film. The lower spectrum (a) is
a reference spectrum of Carbowax 20M, while spectrum (b) is a reference spectrum of n-
nonane. Spectrum (c) was taken 11 seconds after sample injection. In this spectrum, the

features at 1464 and 1381 correspond to nonane features, while the positive feature at

41

 

l l

0.02 AU
((0
(C) [0.08 AU

0.68 AU
(b)

[0.5 AU

(a)

 

 

 

1 1 1 1
4000 3000 2000 1000
Frequency (cm‘ 1)

Figure 2.5. FTIR/ATR spectra for pyrrole over Carbowax 20M at 60°C.
(a)=Carbowax 20M reference, (b)=pyrrole reference,
(c)=22 seconds, (d)=91 seconds.

42

 

 

 

 

 

 

 

 

 

 

 

 

1 1 1

(d) 1
0.01 AU
1 10.04 AU
1 (C)
1 fl ,
1
‘ 025AU ‘
1(b) 1 1 1
' A“
1 1
(a) [0.5 AU
1 1 1 1
1600 1200 800

Frequency (cm‘ 1)

Figure 2.6. FTIR/ATR spectra for n-nonane over Carbowax 20M at 60°C.
(a)=Carbowax 20M reference, (b)=n-nonane reference,
(c)=22 seconds, (d)=91 seconds.

43

1262, and the negative features at 1063, 1025, and 791 cm'1 corresponding to Carbowax
20M features. Spectrum ((1) was taken 23 seconds after n—nonane injection, and only
negative features at 1263, 1092, 1022, and 799 cm'1 are observed. These features all
correspond to features with similar frequencies observed in the Carbowax reference
spectrum (spectrum (a)). The lack of features due to n-nonane suggests that it has been
purged from the film 23 seconds after injection. A similar profile was observed for 1-

butanol, and it was desorbed from the film after 34 seconds.

A summary of analytes and the observed FTIR/ATR profiles is shown in Table 2.3.

Table 2.3. Summary of FTIR/ATR Spectral Profiles

 

FTIR/ATR
Analyte Number Analyte Boiling Point (°C) Spectral Profile
1 acetone 56 1
2 methanol 65 1
3 n-hexane 69 1
4 benzene 80 1
5 cyclohexane 81 1
6 isopropanol 82 1
7 thiophene 84 1
8 trichloroethylene 87 1
9 n-heptane 98 1
10 2-pentanone 10 1 1
1 1 toluene 1 1 1 1
12 pyridine 116 2
13 l-butanol 118 3
14 nitropropane 120 2
15 tetrachloroethylene 135 l
16 n-octane 126 1
17 pyrrole 1 3 1 2
18 n-nonane 151 3

1 = analyte features observed, no changes in spectrum after analyte purged.

2 = analyte features observed, positive features observed in spectrum after analyte purged.

3 = analyte features observed, negative features observed in spectrum after analyte
purged.

2.4. Discussion

Efiect of Coating on SAW Device Response. Results from characterization of
Carbowax 20M coated and blank (uncoated) SAW devices (Figure 2.2) show that the
observed response of blank devices is generally a function of analyte boiling point, with an
increased boiling point leading to an increased SAW device response. The trend for the
blank (uncoated) devices (open circles, dashed line) is a smooth increase except for
pyridine, l-butanol, nitropropane, and tetrachloroethylene. The increased response to
these analytes is due to an increased affinity between the quartz (a-SiOz) and these
analytes. The increasing SAW device response as analyte boiling point increased was first
noted by King [36], and can be attributed to condensation occurring in the test cell. Since
the analyte enters the sensor test cell as a gas it is not surprising that the higher boiling
point analytes partially condense on the relatively cool (60°C) sensor surface. Therefore,
one expects increased sensor response due to analyte condensation as the boiling point of
the analyte increases. This general trend is necessary as a reference trend for further
studies into the selectivity of films coated on the SAW device. Any significant deviations
from this observed response trend can be attributed to selective interactions between the
analyte and the film.

The macroscopic characteristics of Carbowax 20M films are dictated by the
molecular properties of the polymer. Carbowax is a high molecular weight poly(ethylene
glycol) film that consists of hydroxyl end groups and ether linkages between ethylene
groups. This very polar film interacts with analytes through hydrogen bond donation via
the hydroxyl groups and hydrogen bond acceptation through the ether oxygen atoms.
SAW devices coated with Carbowax show enhanced responses for thiophene, pyrrole, and
n-nonane. Data from Table 2.2 demonstrate that analyte responses on the Carbowax

coated SAW devices were generally greater than those observed on the blank devices.

45

The average response ratio was approximately 4.5 with an estimated deviation of i 1.
This average response factor suggests that the Carbowax 20M film concentrates most
analytes so that their presence in the film results in an average response increase 4.5 times
that observed for the blank, irregardless of the analyte chemical functionality. Analytes
with response ratios greater than the average enhancement value of 4.5 suggest that
selectivity phenemena are occurring between the analyte and the film.

The dramatic enhancement of the response to thiophene (C4H4S) and pyrrole
(C4H4NH) can be explained by strong analyte/film interactions. These are both aromatic
heterocyclic compounds that are capable of strong hydrogen bond interactions with the
Carbowax film. Thiophene can act as a hydrogen bond acceptor through interactions with
the hydroxyl endgroups of the Carbowax chains, and pyrrole can act as a hydrogen bond
donor through interactions between its amide proton and the ether oxygens of the
Carbowax backbone. Additionally, they both have small molar volumes which allows
stronger interactions because these molecules are more easily able to diffuse into the film
and interact.

The response of Carbowax 20M coated SAW devices to alkanes except for n-
nonane was very close to values observed on blank devices. This is due to the minimal
interactions that occur between Carbowax 20M and n-alkanes. As seen in Table 2.1,
alkanes have large dispersion interaction (longé) terms, which indicates that these
analytes are only capable of interaction through dispersion interactions. Carbowax 20M
films consist of highly polar regions that are separated by ethylene groups, and the
ethylene groups are the only functionality in the film which are capable of dispersion
interactions. The strength of dispersion interactions is dependent on the size of the non-
polar region that is capable of interaction, and in Carbowax 20M the ethylene regions are
not large enough to significantly interact with the non-polar n-alkanes. However, an

exception to the decreased SAW response for n-alkanes was observed for n-nonane,

46

which had a response factor of 7.5. The enhanced response cannot be due to an increased
affinity between the Carbowax 20M film because the interaction mechanisms between n-
nonane and Carbowax are quite small, as previously discussed. Therefore, the enhanced
response observed for n-nonane must be attributed to an event other than an strong
analyte/film interaction. The change in frequency of the SAW device can be both due to a
mass loading effect and to a change in the Viscoelastic properties of the film coated on the
SAW device [2-3]. Recent work by Grate et al. [6] suggests that analyte induced swelling
of a polymer coating on a SAW device can affect the observed frequency response of
SAW devices by decreasing the velocity of the acoustic wave propagation through the
polymer film. In fact, it was proposed that the response due to polymer swelling
dominates the response due to selective interactions. We believe that this is the effect that
is being observed between a Carbowax 20M coated SAW device and a pulse of n-nonane.
One of the reason that this film was studied as a coating on SAW devices was that
it has been well characterized for use as a stationary phase to separate gaseous species
based on their polarity in gas-liquid chromatography (GLC) techniques, and we hoped that
the GLC properties would correlate with observed SAW device selectivity. One of the
most common classification schemes used for GLC stationary phases is based on the
determination of McReynolds numbers for a specific stationary phase. The McReynolds
numbers are used to predict relative retention of a standardized series of analytes so that
the numerous stationary phases used for GLC can be selected based on the retention of
these test analytes. Previously determined numbers for Carbowax 20M are shown in
Table 2.4. A large McReynolds number is indicative of strong analyte retention. Values
from Table 2.3 predict that Carbowax should retain more polar species strongly (l-
butanol, nitropropane, and pyridine), with less retention of benzene and 2-pentanone. The
response of Carbowax 20M coated SAW devices do not demonstrate significant

enhancement to analytes that have large McReynolds numbers, which suggests that the

47

behavior of films as studied for use as chromatographic stationary phases does not

correlate well with observed SAW device responses.

Table 2.4. McReynolds Numbers for Carbowax 20M.

benzene l-butanol 2-pentanone nitropropane pyg'dine
332 536 368 572 510

Time Dependent F TIR/ATR Studies of Carbowax 20M / Analyte Interactions.
Time dependent FTIR/ATR spectra demonstrate that changes in spectral features that can
be attributed to the Carbowax 20M film occur through interaction with a short pulse of
analyte vapor. The background used for these studies consisted of the Carbowax 20M
coated IRE, so that changes that occur in the spectrum due to analyte interaction were
observed. However, changes that correspond to vibrational modes of the Carbowax 20M
film occurred for certain analytes. These changes could be due to analyte induced changes
in the film morphological properties or changes in the spectral properties of the film.
FTIR using attenuated total reflectance (ATR) sampling is used for studies of thin films,
and the sampling depth ((1,) is a function of the wavelength of the IR beam, as well as the
indices of refraction for both the film and IRE. If the film’s index of refraction changes
with sorption of analyte, the sampled depth of the film will change. This effect is seen in
Figure 2.6, which is n-nonane interacting with the Carbowax coated IRE. As n-nonane
interacts with the film (c), negative and positive features are observed. These features are
likely due to a change in the refractive index of the film from a value of ~ 1.5 towards
1.41, which is the refractive index of nonane. This results in a decreased sampling depth,
and a corresponding negative band because a smaller amount of film is sampled.
However, negative bands persist in the spectrum after the n-nonane has been purged from

the film, as evidenced by the lack of n-nonane features. Therefore the negative features

48

observed in the spectrum after analyte has been purged from the film are not due to
changes in the refractive index due to analyte sorbed in the film.

However, the changes in the spectrum can be used to infer changes in the
morphology of the polymer that is coated on the IRE. Since the polymer acts as a
background, the persistent changes in spectral features after analyte has been purged from
the film are due to a change in the film. Since we have observed negative bands persistent
in the spectra for n-nonane and l-butanol, it can be inferred that there has been an analyte
induced change in the Carbowax film. The negative bands can be attributed to an analyte
induced swelling of the Carbowax film. If the sampling depth is constant and the polymer
swells due to interaction with analyte, then less film will be sampled after the analyte has
induced swelling. Because less film is sampled and the polymer coated IRE is used as the
background, negative bands appear in the spectrum. Similarly, if the polymer is affected
by the analyte and changes its morphology so that it changes from a thin film to become
less continuous, less film will be sampled after analyte interaction, which will result in
negative spectral features. The reason that positive features appear in the spectrum are
not as well understood, however the changes in the spectrum can, at a first approximation,
be used to infer that some type of analyte induced change in the polymer has occurred.
The positive changes in the spectrum that are induced by analyte interaction requires

further study.

49

2.5. Conclusions

Carbowax 20M coated SAW devices have been characterized using SAW sensor
measurements with pulsed sample introduction. Time dependent FTIR/ATR data of
analyte/film interactions shows that n-nonane, which is not capable of strong interactions
with the Carbowax 20M film, induced morphological changes in the film. This
morphological change correlates well with the enhanced response of n-nonane on a
Carbowax 20M coated SAW device, which suggests that FTIR/ATR can be a useful
spectroscopic probe of analyte induced changes in the Viscoelastic properties of the
polymer coating on SAW devices. This data supports the idea that the response of
polymer coated SAW devices is affected by analyte induced swelling of the polymer
coating, however the swelling does not occur universally, which suggests that the
selectivity of the film remains an important factor in the selectivity of polymer coated

SAW devices.

2.6.

H
.

9‘5"

“99°.“

11.

13.
14.
15.
l6.
17.
18.

19.
20.

21.
22.
23.

24.
25.

26.

27.

28.

29.

50

References

Ballantine, D. 8., Jr.; Wohltjen, H. Anal. Chem. 1989, 61, 704A.

Wohltjen, H. Sensors and Actuators 1984, 5, 307.

Auld, B. A. Acoustic Fields in Waves and Solids Ch. 12, Wiley-Intersciece, New
York, 1973.

Martin, S. J.; Ricco, A. J.; Frye, G. C. 1990 Solid-State Sensor and Actuator
Workshop, ACS Spring Meeting, San Francisco.

Frye, G. C.; Martin, S. J. Appl. Spec. Rev. 1991, 26, 73.

Grate, J. W.; Klusty, M.; McGill, R. A.; Abraham, M. H.; Whiting, G.; Andonian-
Haftvan, J. Anal. Chem 1992, 64, 610.

Ballantine, D. 5., Jr.; Wohltjen, H. IEE Ultrason. Symp. 1988, 559.

Bartley, D. L.; Dominguez, D. D. Anal. Chem. 1990, 62, 1649.

Frye, G. C.; Martin, S. J.; Ricco, A. J. J. Sens. Mater. 1990, 1-6, 335.

Martin, S. J.; Ricco, A. J.; Frye, G. C. 1990 Solid State Sensor & Actuator
Workshop, ACS Spring Meeting, San Francisco.

Ballantine, D. S., Jr. Anal. Chem 1992, 64, 3069.

King, W. H., Jr. Anal. Chem. 1964, 36, 1735.

Persinger, H.E.; Shank, J. T. J. Chromatogr. Sci. 1973, II, 190.

Leoni, V.; Pucetti, G.; Grella, A. J. Chromatogr. 1975, 106, 119.

Edmonds, T. E.; West, T. S. Anal. Chim Acta 1980, 117, 147.

Lopez-Roman, A.; Guilbault, G. G. Anal. Lett. 1972, 15, 225.

Frechette, M. W.; Fasching, J. L.; Rosie, D. M. Anal. Chem. 1973,45, 176.
Sanchez-Pedreno, J. A. 0.; Drew, P. K. P.; Alder, J. F. Anal. Chim. Acta 1988,
207, 285.

McReynolds, W. O. J. Chrom. Sci. 1970, 8, 685.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chem Soc.
Perkin. Trans. 2 1990, 1451.

Abboud, J. L.; Kamlet, M. J.; Taft, R. W. J. Am. Chem. Soc. 1977, 99, 8325.
Abraham, M. H.; McGowan, J. C. Chromatographia 1987, 23, 243.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chromatogr.
1991, 587, 213.

Taft, R. W.; Abboud, J-L. M.; Kamlet, M. J. J. Am. Chem Soc. 1981, 103, 1080.
Abraham, M. H.; Grellier, P. L.; McGill, R. A. J. Chem Soc. Perkin Trans. 2
1987, 797.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chromatogr.
1990, 518, 329.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chromatogr.
1991, 587, 229.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor, P. J. J. Chem
Soc. Perkin. Trans. 2 1990, 521.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Duce, P. P.; Morris, J. J.; Taylor, P.
J J. Chem Soc. Perkin. Trans. 2 1989, 699.

30.

31.

32.

33.

34.

35.
36.

51

Abraham, M. H.; Duce, P. P.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor, P.
J. Tetrahedron Lett. 1988, 29, 1587.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor, P. J.; Laurence,
C.; Berthelot, M. Tetrahedron Lett. 1989, 30,2571.

Abraham, M. H.; Grellier, P. L.; McGill, R. A.; Doherty, R. M.; Kamlet, M. J.;
Hall, T. N.; Taft, R. W.; Carr, P. W.; Koros, W. J. Polymer 1987, 28, 1363.
Abraham, M. H.; Grellier, P. L.; McGill, R. A. J. Chem Soc. Perkin. Trans. 2
1988, 523.

Kamlet, M. J.; Abboud, J-L, M., Abraham, M. H.; Taft, R. W. J. Org. Chem.
1983, 48, 2877.

Bowers, W. D.; Chuan, R. L.; Duong, T. M. Rev. Sci. Instrum 1991, 62, 1624.
King, W. H., Jr. Anal. Chem 1964, 36, 1735.

Chapter 3

Studies of the Response of Tenax GC Coated
Surface Acoustic Wave (SAW) Devices

3.1. Introduction

Piezoelectric devices based on surface acoustic waves (SAWS) are useful for
chemical sensing applications because of their mass sensitivity, small size, low power
requirements, and durability. The first descriptions of SAW devices used as gas phase
chemical sensors were given by Wohltjen and Dessy [1-3]. It was found that when quartz
SAW devices were coated with selective coatings the device became a chemical sensor
whose selectivity was dictated by the selective properties of the coating. An ideal
selective coating would respond to a single target analyte in the presence of competing
species, however this level of selectivity has proven difficult to achieve. Numerous classes
of films have been studied as selective SAW device coatings including molecular
recognition centers [4—5], Langmuir-Blodgett films [6],and organic or inorganic polymers
[7-13].

It is known that the frequency response of SAW devices is a function of mass
loading and elastic properties of the coating [14-15] as described in equation 3.1.

_ 2 -_ 2 41111411
Af_(k,+k,)rohp k2f°h(V§KX+2u'j] (3.1)

 

where Af is the change in resonant frequency, kl and k2 are substrate constants, fo is the
uncoated resonant frequency, h is the film thickness, 6 is the film density, X is the Lame
constant, 11' is the film shear modulus, and VR is the Rayleigh wave velocity. The first

term describes the mass loading effects on the device response and is a function of added

52

53
mass since the product ho is the mass per unit area on the device. The term
(it + u')/(X +211) describes film elastic properties, and if within the range 0.5-1.0 and
the polymer is above its glass transition temperature it can be ignored since the modulus
terms are small compared to the square of the Rayleigh wave velocity.

While the effect of mass loading on the device response is well understood, the
effect of the film Viscoelastic properties are not well characterized or predictable. Grate et
al. [16] have proposed a model describing the response of polymer coated SAW devices
that is based on mass loading and polymer swelling. A relationship between vapor
(analyte) frequency shift and effects that occur on the coated SAW device given in

equation 3.2.

 

Afv =[AfszK]+[AfszKIAS/1W] (3.2)
p: pL a

where Af, represents the frequency shift on vapor sorption due to gravirnetric and
swelling effects. The first term describes the gravirnetric effect based on partitioning of
analyte of concentration C. into the film of thickness f,. The term ps is an approximation
for the mass of the film divided by the volume of the film containing the vapor. The
second term describes swelling effects, where CvK is the concentration of vapor in the film
and CVK/pL is the volume fraction increase of the film due to vapor sorption. ASAW was
defined to be the frequency change (in kHz) due to a 1°C change in the temperature per
kHz thickness of coating on the device, and or is the polymer thermal expansion
coefficient. It was proposed that SAW device responses are dominated by the second
term since it is believed that analyte induced swelling decreases the modulus of the
polymer film.

We have studied the response of SAW sensors coated with Tenax (poly(2,6-

diphenyl-p-phenylene oxide), a widely used chromatographic stationary phase and sorbent,

54

to a varied set of probe molecules introduced as narrow pulses. The morphological
properties of Tenax have been investigated by Sakodynskij et al. [17]. Tenax GC is a
porous medium with a pore volume of 0.667 cm3/g, and an average pore radius of 720 A.
It has been classified as a non-polar gas chromatographic stationary phase because the
ether oxygen of the phenylene oxide backbone is shielded by the phenyl and phenylene
groups, however its chromatographic retention behavior shows selectivity based on
analyte polarizibility and double bonds. It was also stated that retention of alcohols
relative to similar molecular weight alkanes was reduced. This retention property has led
to widespread use of Tenax GC as a concentration medium for aromatic and chlorinated
organic species [18-23]. If the macroscopic behavior of Tenax as a both a gas
chromatographic stationary phase and a concentrating medium are considered, it is
reasonable that a Tenax coated SAW device should show similar selectivity towards
aromatic and chlorinated analytes.

The goal of this work is to study the interactions that occur between a Tenax GC
coated SAW device and a series of 16 target analytes. The analytes were chosen so that
an examination of a variety of chemical functional groups from a range of boiling points
could be performed. Analytes were injected as narrow pulses in the vapor phase into a
thermostatically controlled test cell. Correlations between enhanced responses and
changes in polymer morphology using in-situ FTIR/ATR studies have been drawn, and we
believe that in-situ FTIR/ATR can be used as a spectroscopic tool to predict when analyte
induced polymer morphology changes occur. This data provide a predictive tool for
explaining the response of polymer coated SAW device that are attributed to analyte

induced changes in Viscoelastic properties.

55

3.2. Experimental

Reagents. Tenax GC (poly(2,6-diphenyl-p-phenylene oxide)) was obtained from
Alltech, Inc., and prior to use was solvent extracted in a Soxhlet apparatus following
techniques outlined by Picer and Picer [23]. This involved extraction for 24 hours with
dry acetonitrile, followed be a 12 hour extraction with diethyl ether and a 24 hour
methanol extraction. After drying overnight at 110°C, the polymer was placed in a sealed
bottle and stored in a dessicator.

The set of test (probe) analytes used in this work is given in Table 3.1. The
analyte list includes five compounds used by McReynolds [24] to characterize
chromatographic stationary phases (benzene, butanol, 2-pentanone, nitropropane,
pyridine), a homologous series of hydrocarbons (n-hexane, n-heptane, n-octane), aromatic
hydrocarbons, alcohols, and aromatic heterocycles. Also included in Table 3.1 are
solvatochromic parameters that have been determined for these analytes. These
parameters are used to describe the relevant molecular interactions that are possible for a
specific analyte through hydrogen bond donation (or), hydrogen bond acceptance ([3),
dispersion interactions (logL'é), electron donation/acceptance (R2), and charge
stabilization (nf ). A more detailed description of each variable and collections of
parameters may be found in the extensive work of Abraham et al. [25-39]. All probe
analytes (>99+% purity) were obtained from Aldrich Chemical Company or J. T. Baker
and used as received.

Surface Acoustic Wave (SAW) Device. SAW results were obtained using 200
MHz resonator devices manufactured by Femtometrics, Inc. (Costa Mesa, CA). The
devices are packaged in an array configuration consisting of six separate resonators
permanently mounted on a header which plugs into a motherboard that contains the

individual resonator electronics. A separate reference oscillator is contained on the

56

motherboard. The 200 MHz resonators exhibit a resonant frequency shift of
approximately 904 Hz when perturbed by a surface mass change of 1 nanogram [40].

The SAW array and reference chips were plugged into a motherboard that
provides power, oscillating potential, and feedback circuitry to the sample and reference
devices. Data acquisition and processing was accomplished using an IBM PC compatible
computer and control software (Freq6., Femtometrics, Inc.). The signals acquired were
the individual resonator beat frequencies, which is a difference between the reference and
individual resonator frequencies. The data files for the array were then further processed
using an in-house developed peak processing routine under LabView (National
Instruments, Inc., Austin, TX).

Prior to film coating, SAW resonators were cleaned by solution soaking and ozone
cleaning. The solution soaking involved rinsing the device with dry acetone, followed by
30 minute soaks in dry toluene and electronic grade isopropanol. The device was then
allowed to dry in air. The dry device was then treated for 30 minutes in an ozone cleaner
(UV Ozone Photoreactor PR-100, UVP Inc., New Jersey), followed by a 30 minute
isopropanol soak. For the final drying step, the device was kept in a dessicator.

Films for use in surface acoustic wave (SAW) experiments were spray cast from 1
mg/mL chloroform solutions using an air brush with dry nitrogen (99.95 %, AGA Gas Co.)
as a carrier gas. The films were deposited while the sensor device was oscillating to allow
observation of the frequency shift due to the mass of film deposited. Typical SAW device
coatings resulted in a 150 kHz frequency shift, which corresponds to a film thickness of
approximately 100 A.

 

 

cod
cod
Rd
$6
mod
3 d
mvd
cod
med
3 .o
and
cod
3 .o
cod
5.6
and

 

 

 

8d odd med

wad 3d mnm

mad Ed ad

ad 8d 8d

add med odd

de odd Ed

Ed Ed 0:

8d odd :d

M ed and odd

dmd med :3

ad ad 02

dd Rd dd

de Ed 93

8d 8d SN

3d wad 8d

Sd :d and
3
285 55:58 826825
530:me

5w:0:m 03m
0:092

cod
cod
cod
hmd
cod
cod
cod
cod
cod
cod
mmd
cod
cod
cod
med
3d

588m 23.
88-:

cm:
5:
GQ
w:
c:
:—
2:

8%:

0:88-:
053509520050:
0:30.505:
6:83- _
esdEd
0:039
0:::S:0:-m
05030::
0535080203
0:02:25
8:30.58:
0:008:20?
0:3:0:
0:30;:
65:50::
0:9000

d00_>0fl >9<m @8000 00 08:09 .6 862m :8 @003 8300.02 058m

 

._.m 030%

58

Sensor Test Apparatus. A schematic diagram of the experimental apparatus is
shown in Figure 3.1. The sensor test cell was machined from aluminum, and the lid and
body were joined using stainless steel bolts and sealed with a thin Teflon gasket. The total
volume of the test cell, including the transfer line between the injector and the cell, is ~8
mL. Electrical feedthroughs allowed the SAW device to be placed inside the cell, with the
feedthrough pins making electrical connection to the oscillator circuit. Analytes were
introduced into the system by syringe injection and reached the sensor device as a narrow
plug. Operating conditions were 40 mUmin of dry N2 (99.95%, AGA Gas Co.) flow
throughout the system, sample injector temperature of 250°C, and a test cell temperature
of 60°C. The temperature of the test cell was controlled by heating the cell transfer line
with nichrome wire and was monitored using a K-type thermocouple positioned in the
center of the cell. The injection volume used for each analyte was varied depending on the
magnitude of sensor response so that the maximum frequency difference was below 300
kHz. At least six injections of each analyte were made to insure reproducibility, and
analyte injection order was varied to minimize the effect of analyte order on the sensor
response. Three different Tenax GC coated SAW devices were examined. SAW device
frequency responses typically had relative standard deviations of less than 10%.

The frequency of the SAW devices was measured during exposure to analyte
vapor pulses, and the observed frequency shift was then used to calculate the response of
the SAW device based on observed frequency shift (Hz) per umole of analyte injected,
which normalizes the responses. We plotted SAW response (Hz/umole) versus analyte
boiling point (°C), to allow observation of analyte condensation on SAW device

responses.

59

 

 

 

t a ‘ 1
Gas Handling Injector
System
L J k J

 

 

 

 

 

 

 

 

Teflon \ Aluminum
Gaskét \/ // / Blocks

 

 

 

 

 

 

 

 

 

 

 

\\
Thermocouple
Inlet
v
r fl
Data
Acquisition

 

 

 

Figure 3.1. Schematic diagram of SAW experimental apparatus.

60

Fourier Transform Infrared Spectroscopy ( F TlR) Analysis. A Mattson Galaxy
5020 FTIR spectrometer equipped with a wideband mercury cadmium telluride (MCT)
detector was used to collect IR spectra of analytes, polymer films, and analytes sorbed in
films. The spectrometer is equipped with a modified attenuated total reflectance (ATR)
accessory (Specac, Inc.). The ATR accessory is a HPLC detector that has been modified
to perform analyte-film studies. The attachment consists of a 500 11L flow thru cell in
contact with a ZnSe internal reflection element (IRE) (Specac Inc.) contained in a
temperature controlled water bath held at 60°C. The flow-thru fittings of the cell have
been changed to accommodate 114" stainless steel inlet and outlet fittings that are
connected to a gas handling system. The inlet of the ATR cell is connected to a sample
injector with a 114" stainless steel transfer line. The injector and transfer line are heated to
250°C. Analytes were injected as 1.0 uL pulses into a 20 mUmin flow of dry nitrogen
(99.95%, AGA Gas Co.) FT IR/ATR spectra were collected with 4 cm'1 resolution.

A software package supplied by Mattson (FIRST Macros) was modified to collect
time dependent IR data. A spectrum of the coated ATR element is used as the
background for further data collection. The modified software collects eighty spectra
(each spectrum is the sum of 10 scans taken at one second intervals) following the
injection of each analyte. Using this data collection method, a decrease in peak
absorbance (i.e., a “negative” peak) suggests that the analyte affects the properties of the
original coating by decreasing the amount of film being observed. An increase in peak
absorbance (a “positive” peak) indicates that the analyte/film interaction involves

increasing the amount of film being observed.

61

3.3. Results and Discussion

Response of Tenax GC Coated SAWDevices. The variation in Tenax coated SAW
device responses (in Hz per timole of injected analyte) as a function of analyte boiling
point is shown in Figure 3.2. The average response of four uncoated devices with the
analyte set (open circles) and similar data for Tenax coated devices (closed circles) are
plotted on the same axes. The response for uncoated devices increases in a fairly smooth
trend with increased analyte boiling point, as first noted by King [41]. This trend is
attributed to analyte condensation occuning in the test cell. Since the analyte enters the
sensor test cell as a vapor, it is not surprising that the higher boiling point analytes partially
condense on the relatively cool (60°C) sensor surface. Therefore, one expects increased
sensor response due to analyte condensation as the boiling point of the probe analyte
increases. Any significant deviations from this observed smooth curve trend can be
attributed to an increased selectivity occurring between the probe analyte and the selective
film; we call positive deviations from this general smooth curve trend enhanced responses
and negative deviations from this curve suppressed responses. The response for the Tenax
coated devices (closed circles) closely follows the trend observed on the uncoated SAW
sensor, with increased responses being observed for benzene (#4), thiophene (#7),
trichloroethylene (#8), toluene (#11), pyridine (#12), and tetrachloroethylene (#15).
Responses to methanol (#2), isopropanol (#6), and l-butanol (#13) were not enhanced, in
fact the magnitude of the responses were similar to those observed on the blank devices,

with the response to l-butanol being lower than observed on the blank device.

62

 

 

 

 

 

 

 

 

33
g, 2111) r- .
1H.
11
E . 1 .
g 11. ‘1’
8 8 00
E KID '- /. 131 .1
e 0‘ .
P .0 ' \0
o / -—--~o
—-----o
0 _ 6N’8M\ .
l l l l l I l L
60 80 1(1) 120 140

Boilrng Pomt (°C)
Figure 3.2. Response of Tenax GC coated SAW (0) and blank SAW (0) devices at
60°C.

63

An additional way to look at the SAW response data involves using a response
ratio, which is the Tenax coated SAW response divided by the blank SAW response.
Response ratios are shown in Table 3.2. The response ratios can also be thought of as a
measure of the ability of the film to concentrate or block analyte on/from the SAW device,
with a large response ratio indicative of an effective concentration of analyte and a small
ratio signifying a blocking of analyte from the SAW device. The average ratio for the
Tenax GC coated SAW data is ~2 :0.5, with benzene, thiophene, trichloroethylene,
toluene, and tetrachloroethylene having significantly larger ratios than 2. Methanol,
isopropanol, and 1-butanol had ratios smaller than 2. The response ratio for pyridine was
not significantly above the average of 2.

The solvatochromic data in Table 3.1 indicates that the analytes that exhibited
enhanced SAW device responses (benzene, thiophene, trichloroethylene, toluene, pyridine,
and tetrachloroethylene) have large dispersion interaction (long6) and polarizibility (R2)
solvatochromic terms. Data for analytes exhibiting suppressed responses (methanol,
isopropanol, and 1-butanol) have large H-bond acid strengths (or) and small dispersion
interaction (long) and polarizibility (R2) terms.

The behavior of the use of Tenax GC as a gas chromatographic stationary phase
has been studied by Sakodynskii et al. [17]. They observed that Tenax GC separates polar
analytes based on its dipole moment and other analytes based on the presence of double
bonds. It was also observed that alcohol retention was reduced relative to alkanes of

similar boiling point Observed SAW device responses correlate well with these

64

 

 

e: :5. m: 838-: e:
S. E 5.: 828588283 2

3 $3 28 88:88:: E
med 8: d: : 6:38-: 2

:2 ME: 5:: 8:83 Q

3. :2 :5 88:9 _ _

ed :8 :dw 8888-: d:

:d :2 8d. 8:88-: d

:0 :2 2.: 88:58:28.5 m

N: a: 5. 888:: h
3 ed :2 3:888: e

S“ m: :2 88828: m

a}. d: in 88:8 8

mm s. Nd: 888-: m

I z. a. 6858: N

N.N .3: m0 0:900: fl

3:200: .88.: 33:38:: @8588 92:5. 8:82 2:8,:
ovum 02253: 028:3: >7<m 00:88”: 3<m
85: 00 88...:

2:20.00 :80: 88: e888: 8:6: 3% dd 28:

65
solvatochromic data and observed chromatographic behavior. Analytes with enhanced

responses have large dipole moments and double bonds, while those with suppressed
responses are alcohols with large H-bond acidity and low dispersion and polarizibilities.

Time Dependent F TIR/ATR Studies of Tenax GC / Analyte Interactions. Time
dependent spectra for gas-phase analytes interacting over Tenax GC were collected using
a background that was of the Tenax coated IRE so that time dependent features of the
analyte interacting with the film can be observed.

An example of an analyte interacting with the film, followed by desorption, is
shown in Figure 3.3. This is the profile for a 1 11L pulse of acetone vapor over a Tenax
film. The lower spectra are references for Tenax GC (a), and acetone (b). At 39 seconds
(c), analyte spectral features are observed; at 180 seconds ((1), negative bands at 1184,
1410, and 1419 cm‘1 are present, however these bands were also present at the time zero
spectrum, which means that no overall changes in the Tenax background. Similar profiles
were observed for methanol, isopropanol, and l-butanol.

Another type of FTIR/ATR profile is shown in Figure 3.4, which is for a 1 1.1L
pulse of benzene over a Tenax GC film. The lower spectra are references for Tenax GC
(a) and benzene (b), while spectrum (c) is from 30 seconds, and spectrum ((1) is from 192
seconds. At 30 seconds, benzene features are present, however at 192 seconds, these
features are greatly decreased. The negative bands observed at 1188 and 1414 cm’1 are
features attributed to the Tenax film. Similar profiles for n-hexane 3.5, toluene 3.6, n-
octane 3.7, and tetrachloroethylene 3.8 are shown in Figures 3.5, 3.6, 3.7, and 3.8,

respectively.

66

 

(d) 1’ * W

0.01 AU

(0) 10.03 AU MW

bad/0-06” Adi/1M

“0 40.107 AU WWW

7—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l l | I
4000 3 000 2000 1 000
Frequency (cm’ 1)

Figure 3.3. FTIR/ATR Spectra for acetone over Tenax GC at 60°C.

(aFTenax GC reference, (b)=acetone reference, (c)=39 seconds,
(d)=180 seconds.

 

——1
_

(‘1) 10.005 AU

1007 AU
(0

%

 

 

 

 

06AU
w)
107AU
(0
1 1 1 1
4000 3000 2000 1000

Frequency (cm’ 1)

Figure 3.4. FTIR/ATR Spectra for benzene over Tenax GC at 60°C.
(a)=Tenax GC reference, (b)=benzene reference, (c)=30 seconds,
(d)=l92 seconds.

68

 

(d)

 

 

 

 

 

 

 

b [0.2 AU
( ) 111
0.7 AU
0 1
| 1 1 I
000 3000 2000 1000

Frequency (cm' 1)

Figure 3.5. FTIR/ATR Spectra for n-heptane over Tenax GC at 60‘C.
(a)=Tenax GC reference, (b)=n-heptane reference, (c)=20 seconds,
(d)=290 seconds.

69

 

 

 

 

r 1 1
(d) 10.02 AU
101AU
(Q
03AU
w)
107AU
(0
1 1 1 1
4000 3000 2000 1000

Frequency (crrf 1)

Figure 3.6. FTIR/ATR Spectra for toluene over Tenax GC at 60°C.
(a)=Tenax GC reference, (b)=toluene reference, (c)=20 seconds,
(d)=90 seconds.

70

 

(®

0.05 AU

05AU
(Q

 

 

0.5 AU

(b) _1

 

107AU
(0

 

 

 

l l I l
4000 3000 2000 1 000
Frequency (cm‘ 1)

Figure 3.7. FTIR/ATR Spectra for tetrachloroethylene over Tenax GC at 60°C.
(a)=Tenax GC reference, (b)=tetrachloroethylene reference,
(c)=20 seconds, (d)=393 seconds.

 

 

 

T l
(d)
0.005 AU
10.05 AU
(C) ‘ W
0.5 AU
(b) 111
[0.7 AU
(6)
l l l l
4000 3000 2000 1000

 

 

 

Frequency (cm‘ 1)

Figure 3.8. FTIR/ATR Spectra for n-octane over Tenax GC at 60°C.
(a)=Tenax GC reference, (b)=n-octane reference, (c)=24 seconds,
(d)=662 seconds.

72

A slightly different profile is shown in Figure 3.9, which is for a 1 11L pulse of
thiophene over a Tenax film. The negative features at 1194, 1325, and 1414 cm'1 that are
present in spectrum ((1) are features from Tenax. The positive features at 2961, 2932, and
2872 cm-1 are also from Tenax.

The changes in FTIR bands in Figures 3.4-3.9 occur at a time that the analyte is
not present in the film because of the lack of analyte spectral features. Since these
experiments used a background spectrum of the Tenax coated IRE, persistent negative or
positive features present after the presence of analyte features indicate that these features
must be due to a change in the film coated IRE. Additionally, these analytes exhibited
enhanced SAW device responses (Figure 3.5). Grate et al. [16] have suggested that the
frequency response of polymer coated SAW devices is a convolution of effects due to
mass loading and changes in the Viscoelastic properties of the coated film. They believe
that analyte induced swelling is the dominant factor in SAW device responses, however it
must be noted that these experiments involved measuring SAW responses by passing a
continuous stream of analyte over the coated SAW for several minutes. Under these
conditions, it is likely that analytes induced greater changes in polymer Viscoelastic
properties simply because of increased exposure length to analyte. Our experiments used
pulsed analyte introduction, which is less likely to induce changes in polymer Viscoelastic
properties due to a brief analyte/film interaction time. However, we observed increased
SAW device responses (relative to blank devices) for relatively high boiling point analytes
(toluene and nitropropane). Additionally, we observed that these analytes caused changes

in polymer morphology using FTTR/ATR. Therefore, we believe that the enhanced

73

 

 

 

10.02 AU
(®
006AU
(Q
02AU
w) 1
107AU
(0
1 1 1 1
4 00 3000 2000 1000*

 

 

Frequency (cm‘ 1)

Figure 3.9. FTIR/ATR Spectra for thiophene over Tenax GC at 60°C.
(a)=Tenax GC reference, (b)=thiophene reference, (c)=10 seconds,
(d)=188 seconds.

74

response of toluene and nitropropane are due to analyte induced changes in polymer
morphology and/or Viscoelastic properties, and not due to an increased analyte/film

interaction strength.

3.4. Conclusions

The response of surface acoustic wave (SAW) devices coated with Tenax GC
(2,6-dimethyl-p-phenylene oxide) has been studied using pulsed sample introduction
techniques. Changes in polymer morphology were observed with in-situ FTIR/ATR
spectroscopy, and correlations with enhanced SAW device responses (relative to blank
device responses) indicate that these morphological changes can be used as an indication
of analyte induced changes in polymer Viscoelastic properties that affect polymer coated
SAW device responses. Additionally, the interactions between analytes and Tenax films
occurs in a semi-selective fashion, with enhanced responses for aromatic and chlorinated
species being observed, and suppressed responses for alcohols being observed. This agrees

with previously described properties of Tenax GC.

3.5.

NP‘P‘PPP?‘

10.

11.
12.
13.
14.

15.
16.

17.
18.
19.

20.
21.
22.

23.
24.
25.

26.
27.
28.

29.
30.

75

References

Wohltjen, H.; Dessy, R. Anal. Chem 1979, 51 , 1458.

Wohltjen, H.; Dessy, R. Anal. Chem 1979, 51 , 1465.

Wohltjen, H.; Dessy, R. Anal. Chem 1979, 51, 1471.

Dickert F. L.; Bauer, P. A. Adv. Mater. 1991, 3, 436.

Katrinsky, A. R.; Savage, G. P.; Pilarska, M. Talanta 1991, 38, 201.

Grate J. W.; Klusty, M. Anal. Chem 1991, 63, 1719.

Grate, J. W.; Snow, A.; Ballantine, D. S., Jr.; Wohltjen, H.; Abraham, M. H.;
McGill, R. A.; Sasson, P. Anal. Chem 1988, 60, 869.

Ballantine, D. 8., Jr.; Rose, S. L.; Grate, J. W.; Wohltjen, H. Anal. Chem. 1986,
58, 3058.

Rose-Pehrsson, S. L.; Grate, J. W.; Ballantine, D. 8., Jr.; Jurs, P. C. Anal. Chem
1988, 60, 2801.

Grate, J. W.; Rose-Pehrsson, S. L.; Venezky, D. L.; Klusty, M.; Wohltjen, H.
Anal. Chem. 1993, 65, 1868.

Snow, A.; Wohltjen, H. Anal. Chem 1984, 56, 1411.

Nieuwenhuizen, M. S.; Barendsz, A. W. Sensors and Actuators 1987, 11, 45.
Bowers, W. D.; Chuan, R. L. Rev. Sci. Instrum 1989, 60, 1297.

Auld, B.A. Acoustic Fields in Waves and Solids Wiley-Interscience, New York,
1973, Ch. 12.

Wohltjen, H. Sensors and Actuators 1984, 5, 307.

Grate, J. W.; Klusty, M.; McGill, R. A.; Abraham, M. H.; Whiting, G.; Andonian-
Haftvan, J. Anal. Chem 1992, 64, 610.

Sakodynski, K. I.; Panina, L. 1.; Klinskaya, N. S. Chromatographia 1974, 7, 329.
Leoni, V.; Pucetti, G.; Grella, A. J. Chromatogr. 1976, 106, 119.

Leoni, V.; Pucetti, G.; Colombo, R. J.; D’Ovidio, A. M. J. Chromatogr. 1976,
125, 399.

Vick, R. D.; Richard, J. J .; Svec, H. J.; Junk, G. A. Chemosphere 1977, 6, 303.
Billings, W. N.; Bidleman, T. F. Atmosph. Environ. 1983, 17, 383.

Grate, J. W.; Rose-Pehrsson, S. L.; Venezky, D. L.; Klusty, M.; Wohltjen, H.
Anal. Chem 1993, 65, 1868.

Picer, N.; Picer, M. J. Chromatogr. 1980, I93, 357.

McReynolds, W. O. J. Chrom. Sci. 1970, 8, 685.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chem. Soc.
Perkin. Trans. 2 1990, 1451.

Abboud, J. L.; Kamlet, M. J.; Taft, R. W. J. Am Chem Soc. 1977, 99, 8325.
Abraham, M. H.; McGowan, J. C. Chromatographia 1987, 23, 243.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chromatogr.
1991, 587, 213.

Taft, R. W.; Abboud, J-L. M.; Kamlet, M. J. J. Am Chem Soc. 1981, 103, 1080.
Abraham, M. H.; Grellier, P. L.; McGill, R. A. J. Chem Soc. Perkin Trans. 2
1987, 797.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.
41.

76

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chromatogr.
1990, 518, 329.

Abraham, M. H.; Whiting, G. S.; Doherty, R. M.; Shuely, W. J. J. Chromatogr.
1991, 587, 229.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor, P. J. J. Chem
Soc. Perkin. Trans. 2 1990, 521.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Duce, P. P.; Morris, J. J.; Taylor, P.
J J. Chem. Soc. Perkin. Trans. 2 1989, 699.

Abraham, M. H.; Duce, P. P.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor, P.
J. Tetrahedron Lett. 1988, 29, 1587.

Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Morris, J. J .; Taylor, P. J.; Laurence,
C.; Berthelot, M. Tetrahedron Lett. 1989, 30,2571.

Abraham, M. H.; Grellier, P. L.; McGill, R. A.; Doherty, R. M.; Kamlet, M. J.;
Hall, T. N.; Taft, R. W.; Carr, P. W.; Koros, W. J. Polymer 1987, 28, 1363.
Abraham, M. H.; Grellier, P. L.; McGill, R. A. J. Chem Soc. Perkin. Trans. 2
1988, 523.

Kamlet, M. J.; Abboud, J-L, M., Abraham, M. H.; Taft, R. W. J. Org. Chem.
1983, 48, 2877.

Bowers, W. D.; Chuan, R. L.; Duong, T. M. Rev. Sci. Instrum 1991, 62, 1624.
King, W. H., Jr. Anal. Chem 1964, 36, 1735.

Chapter 4
Future Work

4.1. Instrumental Development

4.1.1. Comparison of Sample Introduction Methods

Grate et al. [1] have proposed that the response of a polymer coated SAW sensor
is dominated by analyte induced swelling of the polymer matrix. They suggest that the
sensor response due to partitioning is approximately four times less than the response due
to the swelling of the polymer coating. This was determined through the comparison of
partition coefficients determined by gas-liquid chromatography (GLC) and SAW device
frequency shifts. The GLC partition coefficients were determined from retention data of a
pulse of sample injected into a chromatographic column that contained the film of interest
coated on a solid support. The SAW partition coefficients were determined from steady
state frequency shifts of the coated device after several minutes of interaction with the
film. The chromatographic sampling technique allows the analyte to interact numerous
times with the stationary phase, and it is assumed passing a continuous stream of analyte
over a coated SAW sensor mimics the conditions in a chromatographic column.

There has not been a comparison between of SAW device partition data based on
pulsed sample introduction and continuous sample introduction techniques. We will be
able to introduce samples in a pulsed mode rapidly and reproducibly using the
chromatographic injection system detailed previously [2]. Additionally, we will bypass the

chromatographic injector apparatus and install an analyte bubbler that is capable of

77

78

generating a continuous stream of analyte vapor. By using an electrically actuated valve,
we will be able to sample the vapor stream and determine the gas-phase analyte
concentration by injection of the sample into a gas chromatograph. By measuring
partition coefficients determined by the response of a coated SAW device using both
pulsed and continuous sample introduction techniques, we will be able to determine if the
partition data derived from pulsed data collection agrees with that from continuous sample

introduction.

4.2. Analyte Induced Viscoelastic Effects on SAW Responses

4.2.1. Background

As described by Wohltjen [3], SAW frequency responses are a function of mass
loading and changes in the Viscoelastic properties of the polymer. The effect of mass
loading on SAW sensor response is well understood, but the effect of changes in polymer
Viscoelastic properties remains unclear. Studies of the effect of changes in polymer
Viscoelastic properties on the response of polymer coated piezoelectric devices have been
described [4— 9], but these studies have not addressed the role of film/analyte selective
interactions in the sensor signal.

One method to enhance the selectivity of coated sensors involves incorporation of
a molecular recognition center in the coating matrix [10- 12]. If the dominant effect in
sensor response is the change in Viscoelastic properties due to analyte sorption, selectivity
due to the recognition center will not be observed. This effect will only be observed if the
polymer used to immobilize the recognition center is susceptible to changes in Viscoelastic
effects. We intend to perform a series of experiments in which the concentration of

molecular recognition sites in a polymer is carefully controlled. By varying the density of

79

recognition centers, the overall effect of both analyte induced Viscoelastic changes and
recognition centers can be observed. The response of a SAW device coated with a film
containing a high density of recognition sites should be dominated by the recognition
centers, while the film with a low concentration of recognition centers should exhibit
responses that are dominated by changes in the film Viscoelastic properties.

Cyclodextrins are particularly attractive for use as host molecules in guest/host
interaction chemistry [i]. The most widely studied cyclodexrrin is B-cyclodextrin, which

has a conical shape that is created by linking 7 glucose monomers. The resulting cup

structure (Figure 4.1) has a hydrophobic interior and a hydrophilic exterior.

4
‘

V

 

 

Figure 4.1. B-cyclodextrin Structure and Dimensions.

The properties of B—cyclodextrin coated SAW devices have been examined by
Dickert and Bauer [10]. They noted that pure cyclodextrins were not suited to sensing
applications because of irreversible bonding of water to the hydrophilic exterior of the
cup. They studied silated and methylated cyclodextrins embedded in siloxane polymers

and found that these deactivated cyclodextrins were not sensitive to water vapor. In

80

addition, they found that perchloroethylene interacted with the cyclodextrin by forming a
guest/host inclusion complex. Cyclodextrins immobilized in non-selective matrices and
coated on piezoelectric mass sensors have been found to respond selectively to benzene
[13] and hydrocarbons [14].

The gas chromatographic separation properties of cyclodextrins and modified
cyclodextrins have also been detailed [15]. It has been found that it is necessary to
alkylate the outer hydroxyl functional groups on the cyclodextrins in order to make them
soluble in a non-selective siloxane matrix. B-cyclodextrins immobilized in siloxane
matrices have been widely used for the gas phase separations of enantiomers [15- 19].

A synthetic method that allows control of recognition center density and dispersion
within a polymer matrix has been described by Xu et al. [20]. This work details a method
for preparing polymeric B-cyclodextrins by linking cyclodextrin units together with
poly(ethylene glycol) segments. The synthetic route was able to produce polymeric [3-
cyclodextrins with molecular weights varying from 3000 to 6000. These polymers were
found to form inclusion complexes with pyrene with selectivity factors that were
significantly larger than observed with monomeric cyclodextrins. The polymerization

chemistry followed a reaction scheme detailed below:

82

solution with 30 wt. % reactants with a 3 hour reaction. Neutralization and
recrystalization resulted in polymers with molecular weights ranging from 3700 to 6000.

Purity and structure of polymers were confirmed with proton NMR.

4.2.2. Proposed Research

The response characteristics of a series of various molecular weight poly(ethylene
glycol) films coated on SAW sensors will be studied using methods described previously.
SAW device response, polymer morphology changes, and LSER determinations will
provide information on the response characteristics of these films. These experiments will
provide a data set that will allow correlations between sensor responses and changes in the
relative number of hydrogen bond donating groups (-OH) and hydrogen bond accepting
groups (-O-) to be drawn. It is expected that the lower molecular weight poly(ethylene
glycol) films will have greater selectivity towards hydrogen bond accepting species
because of the increase in relative number of hydroxyl groups. LSER determinations for
these films should show that these films have a large b term, which corresponds to the
hydrogen bond acidity of the film. As the molecular weight increases, we expect to
observe a decrease in the hydrogen bond acceptor selectivity corresponding to a decrease
in the density of hydroxyl groups. This should also result in a decrease in the magnitude
of the b term from LSER calculations.

These experiments will also serve as reference experiments for work with various
molecular weight polymeric cyclodextrins. Polymeric cyclodextrins will be synthesized
and characterized using previously described methods [20]. By changing the relative
density of recognition centers in a poly(ethylene glycol) backbone in a controlled fashion,
the effect of molecular recognition phenomenon on the sensor response will be

determined.

83

Probe analytes used in these studies will be expanded to more include aromatic
compounds, chlorinated hydrocarbons, alcohols, ketones, n-alkanes, and esters so that a
complete characterization by LSER determinations can be performed. The use of in-situ
FTIR/ATR will provide observation of changes in polymer morphology, and these changes
will be correlated with SAW sensor peak profiles. We believe that observed changes in
polymer morphology observed using in-situ FTIR/ATR can be correlated with changes in
the Viscoelastic properties of the film.

The experiments involving poly(ethylene glycol) and polymeric cyclodextrins
coated on SAW devices will allow a further understanding of the overall SAW sensor
response. The conditions at which changes in polymer Viscoelastic properties dominate
sensor response will be deduced. This will allow an enhanced understanding of the
conditions which allow the selectivity of a molecular recognition center to dominate the

properties of polymer coated SAW devices.

4.3.

I"

@SQMPP’

10.
ll.
12.

13.

14.
15.
16.
17.
18.
19.
20.

84

References

Grate, J. W.; Klusty, M.; McGill, R. A.; Abraham, M. H.; Whiting, G.; Andonian-
Haftvan, J. Anal. Chem 1992, 64, 610.

Townsend, E. B. IV; Rasimas, J. P.; Hoffmann, D. P.; Ledford, J. S. Anal. Chem.
In Preparation

Wohltjen, H. Sensors and Actuators 1984, 5, 307.

Ballantine, D. 8., Jr.; Wohltjen, H. IEE Ultrason. Sympos. 1988, 559.

Bartley, D. L.; Dominquez, D. D. Anal. Chem 1990, 62, 1649.

Martin, S. J.; Frye, G. C. Appl. Phys. Lett. 1990, 57, 1867.

Martin, S. J.; Frye, G. C. Proc. 1991 IEEE Ultrason. Symp. 1991, 393.

Martin, S. J.; Ricco, A. J.; Frye, G. C. 1990 Solid State Sensor and Actuator
Workshop, ACS Spring Meeting, San Francisco, CA.

Ballantine, D. 8., Jr. Anal. Chem. 1992, 64, 3069.

Dickert, F. L.; Bauer, P.A. Adv. Mater. 1991, 3, 436.

Yan, Y.; Bein, T. J. Phys. Chem 1992, 96, 9387.

Lucklum, R.; Henning, B.; Hauptrnann, P.; Schierbaum, K. D.; Vainhinger, S.;
Gopel, W. Sensors and Actuators A 1991, 25-2 7, 705.

Lai, C. S. I.; Moody, G. J.; Thomas, J. D. R.; Mulligan, D. C.; Stoddart, J. F.;
Zarzycki, R. J. Chem Soc. Perkin Trans. 2 1988, 319.

Tong, L.; Hou, Z.; Inoue, Y.; Tai, A. J. Chem Soc. Perkin Trans. 2 1992, 1253.
Schurig, V.; Nowotny, H-P. Angew. Chem Int. Ed. Engl. 1990, 29, 939.
Armstrong, D. W.; Li, W.; Chang, C. D.; Pitha, J. Anal. Chem 1990, 62, 914.
Berthod, A.; Li, W.; Armstrong, D. W. Anal. Chem 1992, 64, 873.

Feibush, B.; Woolley, C. L.; Mani, V. Anal. Chem 1993, 65, 1130.

Armstrong, D. W.; Tang, Y.; Ward, T.; Nichols, M. Anal. Chem 1993, 65, 1114.
Xu, W.; Demas, J. N.; DeGraff, B. A.; Whaley, M. J. Phys. Chem 1993, 97,
6546.

”11.111111111111111 11111111111115