Confocal microscopes are useful research tools for spatial characterization of heterogeneous coating systems. In a confocal microscope, emitted, reflected, or scattered light from the sample is detected through a spatial pinhole, which blocks most of the out-of-focus lights to enhance depth resolution and image quality. The depth profiling feature essentially enables optical sectioning of samples without destructive sample preparation, like cross-section microtoming. The article described three illustrative applications of confocal Raman microscopy (CRM) and confocal laser scanning microscopy (CLSM) to investigate some common coatings phenomena.

Quantification of Component Distribution Using Raman Intensity Ratio
In Example 1, lateral and depth mapping of a styrenated additive was performed by comparing the intensities of Raman transitions associated with phenyl ring of the additive with that of the carbonyl groups of the acrylic polymer. In pigmented coatings, titanium dioxide (TiO2) is a strong absorber at low wavenumbers, and film opacity causes depth attenuation of spectral intensities. By focusing on the spectral range for organic components, the authors monitored the intensity ratios of styrene to acrylic signature bands at various depths to minimize the influence of signal attenuation. As a result, spatial analysis of the additive present at only 2 wt % in the pigmented acrylic paint formulation was determined (semi)quantitatively. The additive concentration, reflected by a constant intensity ratio throughout the film, confirmed its uniform distribution in non-transparent, pigmented coatings.

Visualization and Quantification Using CLSM

Figure 1—Fluorescence images of acrylic polymer film (blue) stained by grape juice (pink).

Compared with CRM, CLSM using reflection or fluorescence contrast can provide real-time 3D imaging with greatly improved speed and spatial resolution. An example is shown in Figure 1 (Figure 9 in the original article), which depicts layer-by-layer optical sectioning of latex film stained by grape juice. The intrinsically fluorescent color compounds in the grape juice provided distinct microscopic contrast and spatial differentiation in the color-coded spectral images. CLSM therefore enabled direct visualization and quantification of stain penetration in the polymer matrix.

(Semi)quantitative Analysis Through Novel Data Processing Methods

Figure 2—CRM data on effect of staining time.

When standard peak-fitting or intensity ratio calculation is not possible, special data processing techniques are required to extend the use of CRM for quantitative and semi-quantitative analyses. In Example 2, the authors explored the second derivative analysis of Raman spectra to resolve low concentration of a surfactant leached out on the film surface of a semi-gloss white paint. In Example 3, a novel data analysis method was developed to take advantage of the fluorescence emission of grape juice. Fluorescence is generally an unwanted limitation encountered often in Raman spectroscopy. The area ratio of the fluorescence envelope to the C-H region was exploited to characterize adsorption, penetration, and removal of grape juice stains. Figure 2 (Figure 16a in the original article) illustrates a typical example of CRM extracted data. The intensity ratio at zero depth was related to the surface concentration of grape juice. The minimum in the depth profile plot was taken as the end point of grape juice penetration. Using this approach, effects of different binder polymers, staining time, and washing protocols were investigated. The CRM semi-quantitative analysis yielded excellent agreement with color measurements from the empirical washability test, while providing deeper insight into the staining and stain removal processes.

In summary, the article demonstrated that confocal microscopes can be employed to skillfully map spatial locations of various chemical species even in pigmented, multi-component coating systems.

Wenjun Wu, Arkema, Inc., Arkema Coating Resins.������������:��wenjun.wu@arkema.com. Dana Garcia, Arkema, Inc. Jeffrey Schneider, Arkema, Inc., Arkema Coating Resins.��

Electrochemical Impedance Spectroscopy (EIS) is a powerful technique that can be used to study the corrosion protection provided by coatings and the mechanism(s) when they fail. This information can assist scientists in assessing how well and how long a coating can protect the substrate, and how to improve its performance. Although this technique typically is used in the laboratory, instrumentation is available to assess performance in the field (e.g., railcars, bridges, etc.).

Corrosion is an electrochemical process of metal degradation, including painted metals, in which electrons move from one part of metal (anode) to another (cathode). With the movement of electrons, the process can be described as an electrical circuit, and the components within this system (i.e., coating, metal, electrolyte) act like and can be modeled as elements such as resistors and capacitors within a circuit, along with their associated parameters. In direct current (DC) systems, R (resistance) = V (voltage)/I (current). For alternating current (AC) systems, impedance, Z, is a circuit’s tendency to resist (or impede) alternating electrical current. The related relationship is Z = V/I. EIS can be used to measure these parameters. For example, EIS can be used to measure the impedance over a wide range of frequencies of a coating system exposed in a corrosive environment such as saltwater.

Figures 1 and 2 show an exposure cell and instrumentation used for EIS measurements. The cell is constructed in the same manner as used in electrochemical measurements, such as studying corrosion. A sample electrode (working electrode; i.e., coated metal), a reference electrode (e.g., saturated calomel), and a counter electrode (inert material; e.g., platinum) are immersed in an electrolyte solution (e.g., 5% NaCl, common for corrosion testing).

EIS instrumentation includes a waveform generator to produce sine waves and a potentiostat to control the potential. It controls the DC potential as well as the AC voltage. The instrumentation measures AC voltage, the current and the phase relationship between them over a range of frequencies. This data is used to calculate the impedance of the system. The instruments are connected to an electrochemical cell which contains the coated metal. A computer runs the experiment, which includes collecting and analyzing the data. EIS experiments can expose and measure the parameters at a specific point in time, but also measure the parameters as exposure time continues, thus allowing assessment of coating degradation and/or corrosion events over an extended exposure period. The sample can contain any metal/coating /electrolyte desired, preferably ones that represent real world conditions.

For example, studying a marine environment, a cell may consist of a steel alloy with a marine coating which is immersed in salt water. To study corrosion that may occur in a vehicle radiator, one may choose a copper alloy, a protective coating, and an antifreeze solution. Performing EIS studies over exposure time will provide information on any coating degradation and/or substrate corrosion, if any. This information will give valuable clues on the mechanisms occurring over time, which can then be addressed to improve performance.

The data collected includes the frequency of the AC waveform and either the magnitude and phase of the impedance at each frequency, or the real and imaginary components of the impedance. Modern equipment and software perform this task and assist in analyzing the results. Common output displays are Bode (Impedance and Phase Angle vs. Frequency) and Nyquist (Imaginary vs. Real Impedance) plots, which will be described in more detail in the next column on EIS. Once data has been collected, it is analyzed. The software can create an equivalent circuit model. This approach makes the analogy between the electrochemical cell (i.e., corrosion or lack thereof) and electrical circuit components such as resistors and capacitors. In short, this clearly indicates how good (or bad) the coating is performing. In greater detail, it provides information on how the coating may be failing.

In the October 2021 issue of CoatingsTech, Part 2 of this summary will focus on what EIS can tell us about coating performance and failure mechanisms.

CoatingsTech August 2021 | vol. 18, no. 8

The coatings technologies of waterborne, high solids, powder, and radiation curable coatings generally require high-molecular-weight latex polymers or strategically designed low-molecular-weight polymers, oligomers, and reactive additives. The design of these resin materials requires molecular weight distribution (MWD) information in the very high and very low-molecular-weight ranges which are accessible by size-exclusion chromatography (SEC). In addition, there is a need for compositional distribution information as a function of molecular weight, particularly for oligomers; absolute MWD information, particularly for water-soluble polymers; and chain-branching information for high-molecular-weight polymers.

This overview discusses the SEC separation mechanism, molecular weight calibration methods including the use of hydrodynamic volume, data treatment methods, and polymer chain-branching determination. The use of molecular size sensitive detectors (viscometer, light scattering) and compositional sensitive detectors (UV-visible, IR) are discussed in the context of illustrative qualitative and quantitative examples. The practice of high-resolution SEC analysis of oligomers is discussed and illustrated with problem-solving examples.

INTRODUCTION

During the past several decades, new coatings technologies, such as high solids, powder, waterborne, and radiation curable coatings have been developed to meet the challenges of: (a) governmental regulations in the areas of ecology [volatile organic compounds (VOC) emissions]; (b) long-term increasing costs of energy, and petroleum-based solvents; (c) more active public consumerism; and (d) the continual need for cost-effective, high-performance coatings in a highly competitive and global business environment.

These new coatings technologies require the use of water as the major solvent with water-soluble or high-molecular-weight latex polymers or the use of strategically designed low-molecular-weight polymers, oligomers, and reactive additives that when further reacted produce high-molecular-weight and crosslinked polymers. Knowledge of the molecular weight and molecular-weight distribution (MWD) of the polymer components in a coatings system is essential for the optimization of polymer design for specific end-use properties.

Since its introduction, many decades ago, gel-permeation chromatography (GPC)¹, or size-exclusion chromatography (SEC), has become an important and practical tool for the
determination of the MWD of polymers. Numerous studies have been published on the use SEC in plastics, elastomers,
and coatings systems including several monographs ^2-15. With the advent of high-efficiency columns, the resolution in the lower-molecular-weight region (molecular weights in the range of 200 to 10,000) has been greatly improved and the speed of analysis increased. These features make high-performance SEC (HPSEC) an indispensable characterization tool for the analysis of oligomers and polymers used in environmentally acceptable coatings systems.

SEC Separation Mechanism

Size-exclusion chromatography separates the polymer molecules by their molecular size or “hydrodynamic volume” in solution. The separation occurs as the polymer molecules elute through one or more columns packed with a porous support. Smaller molecules are retained in the pores to a greater degree than the larger molecules. As a result, the largest size molecule (or the molecule having the greatest hydrodynamic volume) elutes from the column first followed by the smaller molecules.

The volume of liquid at which a solute elutes from a column or the volume of liquid corresponding to the retention of a solute on a column is known as the retention volume (V _R) and is related to the physical parameters of the column, such as interstitial (void) volume (V₀), and internal pore volume. The dependence of molecular size in solution upon retention volume is schematically illustrated in Figure 1. The void volume V₀ corresponds to the total exclusion of solute molecules from the pores. The excluded solute molecules are significantly larger than the largest available size pores. Between V₀ and V_R the solute molecules are selectively separated based on their molecular size in solution.

Beyond the total column volume V_T, separation will not be achieved by a liquid exclusion chromatography mechanism. If molecules appear to separate beyond V_T they are being retained on the column support by an affinity mechanism. The fundamental aspects of the SEC separation mechanism have been treated theoretically by Casassa et al.^16-20, Giddings²¹, and Yau et al.^{22, 23} These treatments are based on an equilibrium distribution of species between the mobile phase in the interstitial volume and the species in the pore volume of the column support.

During the film formation process, an applied liquid film transforms into an adherent solid film through physical and/or chemical processes. The film formation process involves changes in rheology, evolution of mechanical properties, and the resultant ultimate film structure and morphology. In this study, the film formation process of coatings was monitored using a novel and innovative technology denoted as Adaptive Speckle Imaging Interferometry (ASII). Film formation data obtained from ASII was compared with information obtained from more traditional methods, such as set-to-touch time, tack-free time and dry-hard time (from a mechanical recorder), rate of volatiles evaporation from thermogravimetric analysis, and degree of chemical cure from differential scanning calorimetry, where applicable. Combining information from several techniques provides a more comprehensive picture of the film formation process for a wide variety of waterborne and solvent-based coatings: architectural interior and exterior coatings, industrial maintenance, marine, and gel coat coatings. Excellent repeatability and reproducibility of ASII drying curves were demonstrated. ASII drying curves can differentiate film thickness, substrate type, volatile organic compound level, pigment volume concentration, resin type, and coating manufacturer for a specific type of coating.

CONTINUE READING IN THE OF COATINGSTECH.

Introduction

Over the last 40-plus years, newer coatings technologies, such as high solids, powder, waterborne, and radiation-curable coatings have had significant market share growth. These coatings technologies have been developed to meet the challenges of: (a) governmental regulations in the areas of ecology (volatile organic compounds (VOC) emission); (b) long-term increasing costs of energy and petroleum based solvents; (c) more active public consumerism; and (d) the continual need for cost-effective high-performance coatings in a highly competitive and global business environment. These new coatings technologies require the use of water as the major solvent with water-soluble or high molecular weight latex polymers or the use of strategically designed low-molecular weight polymers, oligomers, and reactive additives which, when further reacted, produce high molecular weight and crosslinked polymers. This has led to a need for improved methods of materials characterization in diverse areas, which include molecular weight distribution analysis, particle size distribution (PSD) measurement and assessment and characterization, rheology of coatings, film formation and cure process characterization, morphological surface and bulk characterization, and spectroscopic analysis, as well as a need for improved methods for modeling and predicting materials properties and processes.

Concurrent with the major technological changes in the coatings industry was the significant increase in the rate of change in instrumentation technology. This change was driven by significant advances in electronics and computer and sensor technologies to produce computer-aided, more user-friendly, reliable, and cost-effective instrumentation. The advances in instrumentation and computer technology are filling the need for improved polymer and coatings characterization methods in the context of the newer coatings technologies.

The newer coatings technologies have driven the need for improved methods of PSD assessment and characterization for each of the coatings technologies in terms of size ranges and component particulates as follows:

Waterborne coatings (0.01–50 mm)

• Latex
• Pigments (size and shape)
• Emulsions and dispersions

Powder coatings (0.1–100 mm)

• Resin-pigment composite particle:
  (a) dry powder;
  (b) wet concentrated dispersions
• High solids
• Pigment grind particle size analysis (0.1–100 μm)

The particle size assessment and characterization needs resulting from the challenges presented by the newer coatings technologies and their component materials leads to the following measurement requirements:

• Wide dynamic particle size range
• Improved resolution
• Measurements made in the concentrated dispersion regime
• Assessment of dispersion stability
• Description of structural and textural morphology

The changes in the field of PSD characterization and measurement, aided by�� the advances in electronics computer
and sensor technologies, can be characterized by at least
four activities:

• End user innovation
• Revitalization of older instruments
• Evolution of research grade instrumentation into
  low-cost, user-friendly instrumentation
• Attempts to meet extremely difficult technical challenges

The advances in PSD measurement methods will be discussed in the context of the four above-mentioned activities in terms of measurement principles and the strengths and weaknesses of these methods for characterizing PSDs.

End User Innovation—Commercial Development

The first activity in the changes in PSD characterization and measurement is the commercial development of instrumental methods originally developed in a few academic and industrial laboratories by individuals with high levels of skill and expertise. Innovations in instrumentation and methods usually are driven by the customers who are the technological and scientific leaders and prototype developers, while the instrument vendors are the technological followers who have the engineering skills to do cost-effective commercial development. Examples of such methods are hydrodynamic chromatography (HDC), capillary hydrodynamic fractionation (CHDF), and field flow fractionation methods involving the use of sedimentation, flow, and thermal fields (e.g., SdFFF, FIFFF, and ThFFF).

Hydrodynamic Chromatography (HDC)

HDC, first reported in 1976, was invented in an industrial laboratory by Hamish Small¹ to fulfill an analysis need of The Dow Chemical Company. It took 10 years to be transferred from a laboratory method requiring a high degree of skill into a commercial instrument requiring a moderate degree of skill. The commercialization was carried out by Micromeritics Corporation. However, the commercialization failed in the marketplace because of technological limitations inherent in the method and because of competitive market pressures.

HDC instrumentation is comprised of a liquid chromatograph with an accurate and precise pumping system. Detection was accomplished with a sensitive fixed wavelength ultraviolet (UV) detector. Fractionation and separation of particles occurred in a column packed with uniform, non-porous beads. The separation of particles by size took place in the interstices between the beads and was primarily a function of the bead size and the ionic strength of the medium. The separation mechanism is shown in Figure 1. The interstices between the beads can be treated as capillaries of varying sizes. Larger particles ride higher up on the parabolic flow profile and are eluted first while some smaller particles hug the walls of the capillary experiencing slower flow streamlines and exit later. Other contributing factors to the separation mechanism include electrical double layer effects and Van der Waals attraction. The fatal technical flaw in the methodology was the unpredicted occurrence and amount of particle deposition, which took place in the packed columns. The overall features, benefits, and limitations of this methodology are shown in Figure 2.

Capillary Hydrodynamic Fractionation (CHDF)

The “Tubular Pinch Effect” mechanism operative in CHDF was first discovered by Segre and Silberberg² in 1962 and applied to particle size analysis in 1979 by Regnier and Ball³. However, it did not develop into a viable commercial instrument until innovative experimental and theoretical work was carried out by Silebi and Dos Ramos.^4,5 The commercial embodiment of CHDF had become available from MATEC Applied Sciences. CHDF instrumentation is quite analogous to HDC in that a liquid chromatograph with precise, accurate, and reproducible flow is required. In this instrument, the separation does take place in capillaries of defined dimension using split-flow injection. The mechanism of this separation method is essentially the same as that postulated for HDC. However, now the column is an empty capillary, so there is much less propensity for the particles to deposit in the column. The factors influencing the particle size fractionation are shown in Figure 3, and the associated features, benefits, and limitations of CHDF are shown in Figure 4.

Field Flow Fractionation (FFF)

Field Flow Fractionation methods were invented by J. Calvin Giddings and co-workers and reported in 1967.^6-7 The technology transfer of FFF methods from an academic laboratory method to viable commercial instrument has taken a minimum of 20 years to occur. The DuPont Instrument Company commercialized a sedimentation FFF instrument (SdFFF) in 1986 based upon the pioneering work of J. Calvin Giddings and the subsequent research of Kirkland and Yau.⁸ Unfortunately, the DuPont Instrument Company’s SdFFF instrument was a commercial failure and unavailable by 1991. However, a small entrepreneurial company known as Fractionation, Inc. began making SdFFF instruments commercially available in 1988, followed by FIFFF (flow FFF) instruments in 1991 and, more recently, ThFFF (thermal FFF) instruments.

Field flow fractionation is a form of one-phase chromatography. The instrument is a liquid chromatographic system in which the separation takes place in a flat, narrow channel (typically 100 to a few hundred micrometers in width). A parabolic flow profile is created in this narrow channel by the mobile phase. An external field is applied across the face of the channel such that the field extends over the channel’s thin dimensions and is perpendicular to the flow field. Particles are driven toward the accumulation wall of the channel and form a diffuse cloud, which has an exponential concentration distribution as a result of Brownian motion acting against the field. For particles less than 1–2 µm, smaller particles are displaced farther from the accumulation wall. When the flow field is turned on, the smaller particles, which ride higher on the parabolic flow profiles elute first. The fields that have been used include SdFFF, ThFFF consisting of a temperature gradient across the channel walls, cross flow (FIFFF), electric (ElFFF), and magnetic (MFFF). Commercial instruments are available to perform SdFFF, ThFFF, and FIFFF.

A schematic of the instrumentation separation mechanism and some channel configurations are shown in Figure 5. When the particles are greater than 1 µm, effects of Brownian motion become negligible, and the larger particles ride higher on the parabolic flow streamlines because the velocity-dependent lift forces increase relative to the driving forces produced by the perpendicular external field. In the ~1–100 µm size range, the larger particles elute first, analogous to CHDF. This mechanism is termed the steric mode of FFF (StFFF). In the commercially available instrumentation, field programming options are available for extending the size separation range, reducing the analysis time and optimizing the resolution per unit of time. Examples of separation by SdFFF, FIFFF, ThFFF, and StFFF are shown in Figure 6. The features, benefits, and limitations of FFF chromatography methods are shown in Figure 7. For a review of FFF applied to particle characterization, the reader is referred to the many papers and review articles of Giddings and co-workers.⁹

The commercialization of the instrumental techniques of HDC, CHDF, and FFF have attempted to address the market pull by many industries for PSD analysis methods, primarily in the 0.01–1.0 µm size range. The coatings industry material characterization needs, as a result of the growth of waterborne coatings technology for improved particle size distribution analysis of latex, pigments, and emulsions, is a component of the overall market pull for these techniques.

Revitalization of Older Instrumental Methods

Gravitational and Centrifugal Sedimentation

The second activity in the renaissance of PSD characterization and measurement is characterized by the revitalization of older instrumental methods such as gravitational and centrifugal sedimentation methods. Redesign, modernization with advanced electronics, and user-friendly computer-aided analysis have extended the instrument product life cycle. A good example is disc centrifuge photosedimentometry (DCP).^10,11 The basic technology has been around since the late 1950s and was embodied into a commercial instrument in the 1960s by the Joyce Loebel Company. During the 1970s, the Joyce Loebl Disc Centrifuge Photosedimentometer was the only commercially available instrument for obtaining PSD information by the DCP technique. This technique was particularly well suited for the particle size analysis requirements of the coatings industry for the analysis of latexes, pigments, and emulsions.^12-14 However, it had deficiencies that made it unsuitable as a plant quality control instrument. Over a period of about 10 years, industrial scientists at The Glidden Company developed an improved instrument¹¹, method of use^15,16 and user-friendly data analysis system. The Glidden Company’s DCP technology was licensed to the Brookhaven Instruments Corporation in 1986, which made additional engineering improvements and successfully commercialized the enhanced and revitalized DCP technology. The current instrument can operate in both the line start and homogeneous start modes.

In addition, Brookhaven Instruments Corporation extended the technology by developing an X-ray detection system that facilitated the analysis of heavy small inorganic particles. The X-ray disc centrifuge photosedimentometer (X-DCP) became commercially available in 1991 and was based upon a prototype developed by Terry Allen at DuPont. The X-DCP can operate both in the gravitational and sedimentation modes.

The range of variations in sedimentation instrumentation mode, detection, and experimental method is shown in Figure 8. Also shown in Figure 9 are Stokes’ laws for gravitational and centrifugal sedimentation, which govern the fractionation of particles by size in a gravitational or centrifugal force field. The line start mode^11,12 can produce very high resolution separations. An example of such a separation is shown in Figure 10 for a mixture of nine Duke polystyrene latex standards covering a size range of 107–993 nm in approximately 100-nm increments. Baseline resolution was achieved in seven out of the nine standards. Another major advance in the method of operation is protecting the aqueous meniscus from evaporative cooling and disruption of the density gradient in the fluid by sealing the surface with a small amount of dodecane (usually 1 ml of n-dodecane, injected a few minutes after the density gradient has formed, in about 15–20 ml of an aqueous spin fluid). The sealing of the fluid surface inhibits evaporative cooling and thereby maintains a stable density gradient for several hours and extends the analysis size range. The homogeneous start mode is faster, covers a wider dynamic size separation range, but has less resolution than the line start mode for multi-modal separations. The features, benefits, and limitations of sedimentation methods are shown in Figure 11.

An instrument designed for particle size analysis by sedimentation in a gravitational field was patented by Oliver, Hicken, and Orr and commercially embodied, in 1969, by Micromeritics into an instrument known as the Sedigraph. This instrument produces a cumulative distribution of particle sizes and has been used heavily for the analysis of pigments such as titanium dioxide. Starting in 1988, the Sedigrap’s design and function were revitalized by adding an automatic sample introduction system for unattended operation as well as improving the data analysis system.

Evolution of A Research Instrument Into A Routine User-Friendly Instrument

Dynamic Light Scattering (DLS, PCS, QELS)

The third activity is the evolution of a research instrument into a low-cost instrument that requires a minimum degree of skill to use. An excellent example of this process is the transformation of research-grade photon correlation spectrometers into low-cost, easy to use, limited-function instruments for routine analysis applications.

Dynamic light scattering (DLS) was first known as quasielastic light scattering (QELS) derived from the fact that when photons are scattered by mobile particles, the process is quasielastic. This gave rise to the acronym DLS, since QELS gave information on the dynamics of the scatterer. Since measurements are made with a digital correlator, the acronym PCS (photon correlation spectroscopy) is widely used. The first QELS measurements were made in 1964 by Cummins et al.¹⁷ The first commercial instruments were available circa 1976 and only suitable for use by experts. The early measurements were concerned with obtaining translational diffusion coefficients of macromolecules and particles. The technique was used to gain information about particle size by relating the diffusion coefficient to particle size through the Stokes Einstein equation for spheres.¹⁸

By the early 1970s, the fundamentals had been established. From the mid-1970s onward, improvements occurred in the technology for digital autocorrelators as well as in automation and integration of advances in microprocessors and laser technology. From the early 1980s onward, the PCS instrumentation was increasingly used for particle size analysis. By 1985, low-cost routine 90° instruments were commercially available for routine analyses. The PCS instrumentation was made more automatic and useful for the analysis of particle size and estimating PSD with the commercialization of an automatic sample dilution system invented by Nicoli and Elings.¹⁹

The DLS method utilizes the fluctuations in light scattering intensity observed over a range of time intervals. The autocorrelation function of light scattering intensity at time t��is compared to time zero, C(τ) = ( l(t)-I(0)), for a range of Δt intervals. The autocorrelation function can be described by an exponential decay function of Δt. This process is schematically shown in Figure 12. For a monodisperse sphere, the correlation function is represented by an exponential decay function as shown in Figure 12 where D_T��is the translational diffusion coefficient, q is the scattering wave vector. The particle size d is related to D_T��through the Stokes-Einstein equation as shown in Figure 12. It has been shown that particle sizes obtained on monodisperse spheres are accurate, precise, and highly reproducible. In addition, the PCS instrument is fast and has high throughput.

A schematic of a PCS instrument with auto-dilution is shown in Figure 13. This type of instrument also has been shown to have potential for online measurement of particle size for an emulsion polymerization reactor.²⁰

For polydisperse distributions, an average size and estimate of dispersity can be obtained from the method of cumulants. Inversion of correlograms to obtain PSDs for multimodal distributions has had limited success. If particle size modes are in a ratio of 3:1 or greater, then it should be possible to extract the peak modes from the data. This is a low-resolution method compared to techniques such as CHDF, FFF, or sedimentation in a centrifugal force field.

The features, benefits, and limitations of DLS are summarized in Figure 14.

Attempts to Meet Extremely Difficult Technical Challenges

The fourth activity characterizing the changes in PSD characterization and measurement methods involve attempts to meet extremely difficult technological challenges as a result of advances in electronics and computing and sensor technologies. These technological challenges will be discussed below and can be cataloged as follows:

• Measurements in concentrated dispersions
• Wide dynamic particle size range measurement capability in a single instrument
• On-line and at line analyses
• Structural and textural morphology characterization of particle shape

Concentrated Dispersion Measurements

Fiber Optic Quasi-Elastic Light Scattering {FOQELS)

The need to measure particles in concentrated dispersions has led to the development of fiber optics QELS, given the acronym FOQELS by Brookhaven Instruments, for making measurements in concentrated dispersions up to 40% by weight. This commercial development is based on the work of Dhadwal et al.^21,22 Visible light from a laser diode is focused into the sample by a monomode fiber, and the scattered light is collected by a second monomode fiber at an angle of 153°. The fluctuations in scattered light are analyzed by the photon correlation technique. As long as the dispersion has observable fluidity, translational diffusion coefficients can be measured and transformed into particle size information. The fiber optics technology enables remote measurement of processes. For small particles less than 100 nm, accurate measurements of size in concentrated dispersion can be obtained. As the particle size increases, deviations from absolute size are observed as the concentration increases due to multiple scattering effects. The precision, reproducibility, and size range of measurement associated with PCS measurements applies to the FOQELS measurements. FOQELS applications reported include particle size growth in emulsion polymerization, nucleation processes in metallic oxide manufacture and monitoring protein crystal growth.²³

Laser Light Diffraction

Over the last 30 years, there has been intense activity to revitalize Fraunhofer light diffraction technology to measure concentrated dispersions over a wide dynamic particle size range (0.1–800 mm) by combining Fraunhofer diffraction technology with light scattering detectors to generate a hybrid instrument. There are at least 10 instrument vendors involved in this marketplace. The market pull for this type of instrument has been the need to provide particle size distribution information above and below 1.0 mm with a single instrument. Fraunhofer diffraction physics dates back to 1840. The advent of laser technology coupled with advances in computer technology has made this a viable commercial method of measuring particle size since 1972. The measurement is fast and can be used with dry powders and often the instrument of choice for powder coatings PSD measurements. Combining Fraunhofer diffraction with light scattering detection via Mie theory has enabled measurements to be made down to 0.06 mm, as reported by some vendors. The technique is a moderate to low-resolution technique for extracting multimodal distribution information. Below 2 mm, one has to carefully check the validity of the specific instrument vendor’s software to extract multimodal distributions.

Electroacoustic Efforts

There have been at least three vendors (Matec Applied Sciences, Sympatec, and Malvern Instruments) that have produced instruments which take advantage of electroacoustic techniques in concentrated dispersion to measure both the particle size and the electrophoretic mobility distributions. These techniques are well suited for dense, inorganic materials such as titanium dioxide.²⁴ Exploratory experimentation has been underway to evaluate the technique for organic (low-density) dispersions. The technique is limited with respect to bimodal or multimodal distribution analysis in the same way as is the DLS technique.

Other Methods

Dielectric spectroscopy has some promise for particle size characterization in concentrated colloids, as reported by Sauer et al.²⁵ The authors showed that unique Cole-Cole plots could be obtained as a function of concentration and ionic strength for monodisperse latex particles.

Rheological flow curve analysis using the Cross equation for concentrated dispersions shows correlations of one of the rheological parameters to particle size and distribution.

Both dielectric spectroscopy and rheological flow curve analysis merit further study as possible methods for characterizing the PSD of concentrated dispersions.

Wide Dynamic Particle Size Range Measurement Capability

There has been a lot of activity in developing this capability manifested by the development of hybrid instruments to cover a broad particle size range, some of which has been previously mentioned.

• Fraunhofer Diffraction/Mie Scattering (0.06–8000 µm)
• High resolution gravitational/centrifugal sedimentation
  (0.05–100 µm)
• Combination of FFF modes (e.g., StFFF with FIFFF or SdFFF)
• Single-particle optical sensing/DLS (0.1–300 µm)

The combination of single-particle optical sensing (SPOS) with DLS offers the possibility of analyzing a broad particle size range conservatively from 0.1–300 µm. The single particle optical sensing method is a photozone sensing method in which particles are counted by their ability to obscure light as they move through an orifice. This technique is automated through an auto-diluter. The sensing size range of SPOS is from about 1–300 µm. The SPOS technique is a high-resolution technique with respect to extracting multimodal distributions and is extremely sensitive to low levels of large particles in the presence of small particles (contaminant analysis). A schematic of the auto-dilution SPOS apparatus is shown in Figure 15 with a separation of a hexamodal mixture of monodisperse particles shown in Figure 16 to demonstrate the resolution of the method. When coupled with DLS, this hybrid instrument can characterize the full PSD of particulate systems in which the main contributor to the PSD are small particles.

DLS provides characterization of the small particles while SPOS characterizes the small level of larger particles, barely discernible in the DLS size distribution. An example of this type of analysis is shown in Figure 17.

On-Line, At-line Analysis

The main techniques available for on-line and at-line analysis of particle size distributions have been discussed to some degree. These techniques are:

• Fiber optic QELS (FOQELS) for concentrated dispersions
• Automatic dilution PCS
• Automatic dilution turbidity analysis
• Laser sensor-backscatter measurements

Angular scattering and absorbance measurements as a function of wavelength with photo diode array instruments and automatic dilution coupled with sophisticated mathematical analysis of PSDs is revitalizing the classical turbidity measurement and has been the subject of research by Prof. Luis Garcia-Rubio at the University of South Florida at Tampa.^26-28

Lasentech has developed a laser backscatter instrument to measure effective particle size (chord length across a particle) in concentrated dispersion from several microns to larger sizes and has found practical application for monitoring particle size changes in many processes.

Structural and Textural Morphology Characterization of Particle Shape

The need for structural and textural morphology information has led to cost/performance improvements in automated image analyzers that now provide shape information using signature wave form characterization and fractal dimension characterization of shape and texture.²⁹

Conclusions

As we look forward toward the future, we can expect further advances in PSD characterization and measurement. It is anticipated that additional hybrid type instruments will be produced to cover a wide dynamic particle size range with improved resolution. The quest for PSD analysis methods for concentrated dispersions with good resolution over a wide dynamic range will continue and will be extremely difficult to achieve. However, advances will continue to be made in providing structural and textural information as a result of advances in computer technology. Another major technical challenge which will probably be met is the development of in-line or in-situ particle size monitoring technology for latex reactors.

In addition, we expect the instruments to be more user-friendly, easier to use, more automated and smarter with respect to programmed intelligence.

Acknowledgement

This article is based partly on an American Chemical Society (ACS) workshop on “Modern Methods of Particle Size Distribution: Assessment and Characterization,” which originated as part of the ACS Division of Polymeric Materials: Science and Engineering technical programming. The author acknowledges the contributions of many of the workshop speakers over the last 30 years for their technical contributions to this evolving workshop and for numerous stimulating discussions on the subject.

Specifically, the author wishes to acknowledge the following individuals: Dr. J. Gabriel Dos Ramos, Dr. David Fairhurst, Mr. Peter Faraday, Dr. J. Calvin Giddings, Mr. Kerry Hasapidas, Mr. Richard Karun, Dr. Brian Kaye, Dr. David Nicoli, Dr. Remi Trottier, Dr. Bruce Weiner, Dr. Kim (Ratanathanawongs) Williams, and Dr. Stewart Wood.

Reprinted from Progress in Organic Coatings, 32 /1-4, Theodore Provder, Challenges in particle size distribution measurement past, present and for the 21st century, Pages 143-153, Copyright (1997), with permission from Elsevier.

References

1. Small, J., Colloid Interface Sci. 1976, 57, 337.

2. Segre, G. and Silberberg, A. J. Fluid Mech. 1962, 14, 136.

3. Regnier, E. and Ball, D., Pittsburgh Conference on Spectroscopy and Analytical Chemistry, 1977, paper No. 447.

4. Dos Ramos, J. G. and Silebi, C. A., J. Colloid Interface Sci., 1989, 133, 302.

5. Silebi, C. A. and Dos Ramos, J. G., J. Colloid Interface Sci., 1989, 130, 14.

6. Thompson, G. H., Meyers, M. N., and Giddings, J. C., Sep. Sci. 1967, 2, 797.

7. Giddings, J. C., J. Chem. Phys., 1968, 49, 81.

8. Kirkland, J. J., Yau, W. W., Doerner, W. A., and Grant, J. W. Anal. Chem., 1980, 52 1944.

9. Giddings, J. C., Ratanathanawongs, S. K., and Moon, M. H., KONA Powder Particle, 9, 1991, 200.

10. Koehler, M E., Zander, R.A., Gill, T. T., Provder, T., and Niemann T. F., ACS Symposium Series No. 332, T. Provder, Ed., 1987, 180.

11. Koehler, M. E., Provder, T., and Zander,R.A. US Patent 4,311,039, 1982.

12. Provder, T. and Holsworth, R. M., ACS Div. of Org. Coatings and Plastics Chemistry Preprints, 1976, 36, 150.

13. M. J. Devon, T. Provder, and A. Rudin, “Measurement of Particle Size Distribution with A Disc Centrifuge: Data Analysis Considerations”, ACS Symposium Series, 472, 134 (1991).

14. M. J. Devon, E. Meyer, T. Provder, A. Rudin, and B. B. Weiner, “Detector Slit Width Error in Measurement of Latex Particle Size Distribution with a Disc Centrifuge”, ACS Symposium Series, 472 154 (1991).

15 Holsworth, R. M., Provder, T., and Stansbrey, J. J., ACS Symposium Series No. 332, T. Provder, Ed., 1987, p. 191.

16. Holsworth, R. M. and Provder, T., US Patent 4,478,073, 1985.

17. Cummins, H. Z., Knable, N., and Yeh, Y., Phys. Rev. Let., 1964, 12, 150.

18. Foord, R., et al., Nature, 1970, 227, 242.

19. Nicoli, D. F. and Elings, V. B., US Patent 4,794,806, Jan. 3, 1989, “Automatic Dilution Systems.”

20. Nicoli, D .F., Korti, T., Gossen, P., Nu, J. S., and MacGregor J. F., “Particle Size Distribution II. Assessment and Characterization”, ACS Symposium Series 472, T. Provder, Ed., 1991, 86.

21. Dhadwal, H., Ansari, R., and Meyer, W., Rev. Sci-lnstrum. 1991, 62 (12), 2963.

22. Dhadwal , H., et al., Proc. SPIE, 1993, 16 ,1884.

23. Application Notes, Brookhaven Instruments, FOQELS.

24. O’Brien R. W., Rowlands, W. N., and Hunter, R. J., Proceedings of the NIST Workshop on “Electroacoustics for Characterization of Particulates in Suspensions”, 1993.

25. Sauer, B. B., Stock, R. S., Lim, K. H. and Ray, W. H.,
J. Applied Polymer. Sci., 1990, 39, 2419.

26. Sacato, P., Lanza, F., Suarez, H., and Garcia-Rubio, L. H. Proceedings of the ACS Division of Polymeric Materials: Science and Engineering, 1996, 75, 30.

27. Bacon, C. and Garcia-Rubio, L. H. Proceedings of the ACS Division of Polymeric Materials: Science and Engineering 1996, 75, 32.

28. A. Brandolin, L. H. Garcia-Rubio, T. Provder, M. E. Koehler and C. Kuo., “Latex Particle Size Distribution from Turbidimetry Using Inversion Techniques: Experimental Validation”, ACS Symposium Series, 472, 20 (1991.

29. Kaye, B., American Laboratory, April 1986, 55.

CoatingsTech | Vol. 17, No. 10 | October 2020

Waterborne coatings are multi-component, heterogeneous systems containing a variety of additives, in addition to water, polymer binders, and pigment particles. The distribution of various ingredients in the coating films directly affects a coating’s appearance and performance; for example, optical properties, substrate adhesion, chemical resistance, and mechanical durability. Over the years, a plethora of analytical tools have been utilized in coatings characterization and defect identification. Microscopy and vibrational spectroscopy are complementary techniques and the combination of the two can provide the most complete information regarding the distributions of different coatings ingredients, as well as the fate and the impact of liquids or particulates that come into contact with the dried coating films.

This article reports new applications of two confocal microscopy techniques in coatings research: confocal Raman microscopy (CRM) and confocal laser scanning microscopy (CLSM). In a confocal microscope, most of the out-of-focus structures are suppressed at image formation. The emitted/reflected/scattered light from the sample is detected through a pinhole at the back focal plane of the microscope, enabling depth resolution and a strongly reduced background signal. As a result, confocal microscopes, compared to conventional wide-field microscopes, significantly enhance targeted analysis and image quality. The depth profiling feature of the confocal scanning microscope essentially allows optical sectioning of samples without destructive sample preparation like cross-section microtoming.

CRM combines the chemical sensitivity and specificity of vibrational spectroscopy with a high resolution confocal microscope.¹ Chemical depth profiling by CRM allows for tracking of individual compounds or functional groups as a function of depth.^1-9 Since Raman intensities scale linearly with concentrations, quantification of chemical species is possible by calculating the band area or the intensity ratio of the characteristic Raman transition relative to an internal reference band that retains its spatial homogeneity and temporal consistency.^2-9�� Combining depth profiling and lateral mapping of characteristic Raman bands of different materials, CRM has proven to be a useful research tool for spatial characterization of heterogeneous systems such as polymer blends^1-2 as well as multi-component coatings based on different polymer technologies.^3-9 For example, CRM has been successfully applied to investigate the degree of curing in UV-curable coatings,^3,6 melamine enrichment in polyester/melamine coil coatings and its effect on mechanical properties,⁸ the drying process and movement of the drying front in alkyd coatings,⁵ the photooxidative degradation mechanism and its rate under accelerated QUV weathering conditions,^6,9,10 etc.

CRM has proven to be a useful research tool for spatial characterization of heterogeneous systems such as polymer blends as well as multi-component coatings based on different polymer technologies.

The use of CRM to investigate nontransparent, especially highly pigmented coatings, has been limited because Raman transitions are relatively weak and often overpowered by the background fluorescence of many materials in the heterogeneous coating systems. Moreover, spectrum acquisition for lateral mapping and depth profiling to gain full spatial compositional distribution is time consuming. Compared to CRM, CLSM can perform real-time 3D imaging using reflection or fluorescence contrast with greatly improved speed and spatial resolution.¹¹ Utilizing the reflection intensities, CLSM was able to reveal the distribution and orientation of platelet-shape effect pigment in a metallic automotive basecoat^11,12�� and to monitor blister initiation and growth in an in-situ study of the corrosion process.¹³

By adding a fluorescent dye molecule, color-coded fluorescence images generated by CLSM allowed the observation of water uptake over time and cavity formation during the hydration-degradation-erosion process of antifouling marine coatings.¹⁴�� Similarly, CLSM has been employed to quantify the penetration depth and distribution of UV-curing fluorescent ink pigments in coated and uncoated paper.¹⁵�� Because of the noninvasive nature of spatial-spectral analysis, fluorescence imaging is a rapidly growing field of the biological sciences and has found increasingly broad applications in preclinical and clinical studies.^16-19 These studies also rely on the spectral signatures of either intrinsically fluorescent molecules (e.g., fluorescent proteins in tumor cells) or fluorescent markers as probes. Extensive research and successful applications have enabled real-time observation of diffusive movement of individual fluorescent molecules²⁰ and improved the accuracy of disease detection and intrasurgical metastatic diagnosis.²¹�� With its high sensitivity and specificity, multispectral fluorescence imaging is a powerful, noninvasive method that can be employed to monitor treatment responses²² and even to guide neurosurgeries.²³ Much can be learned and leveraged from other scientific fields to promote the development and application of CLSM techniques for coatings research.

In this article, CRM and CLSM are employed to analyze some common coatings phenomena such as component distribution, surfactant leaching, stain penetration, and removal. The applicability and benefits of these two confocal techniques in coatings research will be demonstrated by three examples. The spatial distribution of a styrenated additive (SA) and surface enrichment of a polymerization surfactant in pigmented acrylic paints are characterized by CRM. Utilizing the intrinsically fluorescent color compounds present in grape juice, CLSM was explored to directly visualize and quantify stain penetration. The fluorescence emission from grape juice, a normally undesired interference for Raman spectroscopy, was utilized to semi-quantitatively characterize stain penetration and stain removal. This article also describes the data analysis techniques developed in these studies to effectively compensate for the weak Raman signals and therefore expand the use of CRM for quantitative and semi-quantitative analyses.

Experimental

Materials

Latex and Paint Samples

The acrylic emulsion polymers AP1 and AP2 were prepared by semi-continuous emulsion polymerization of butyl acrylate, methyl methacrylate, methacrylic acid, and N-(2-methacryloyloxyethyl) ethylene urea. A C14–C16 alphaolefin sulfonate was used as the polymerization surfactant at 0.72 parts based on total monomers (BOTM) to emulsify the monomer mixture and to stabilize the resulting latex particles. The glass transition temperatures (T_g) of both acrylic polymers were 0°C as measured by differential scanning calorimetry (DSC). The SA was an ammonium-neutralized solution of poly(styrene/a-methylstyrene/acrylic acid) terpolymer. The SA has a number-averaged molecular weight of 6500 and an acid number of 205. Its T_g��was determined by DSC to be 110°C.

The acrylic polymers were formulated into low-VOC coatings using the same formulation [35% volume solids (VS) and 31% pigment volume concentration (PVC)] described in a previous publication.² The corresponding paint samples were utilized in the three case studies and denoted as Paint I-SA (based on blend of AP1 and SA), Paint I (based on latex AP1), and Paint II (based on latex AP2).

Fluorescence of Concord Grape Juice

The staining liquid, Welch’s Concord grape juice, was purchased from grocery stores. Fluorescence absorption and emission spectra of the diluted grape juice in water were collected on an FLS980 Spectrometer (Edinburgh Instruments) in the Laboratory for Imaging and Kinetic Spectroscopy at North Carolina State University. Absorption was scanned from 200 to 800 nm. The emission spectrum was generated with an excitation wavelength at 532 nm.

Confocal Raman Microscopy

The paint films for CRM measurements were prepared on Leneta white panels using a 7-mil drawdown bar and dried at 25°C and 50% relative humidity (RH) for seven days. Confocal Raman investigation of paint films was conducted on an Almega™ Raman system (ThermoFisher Scientific). The Almega Raman utilized the 532 nm laser (Nd:YVO4 frequency doubled) with a power of ~25 mW. Spectra were collected from 64 exposures at 1 second each using a 50X long working distance (LWD) objective. Multiple measurements were made for both lateral and depth scans to ensure reproducibility. Depth profiling started from the surface of the film with two-micron step change. To ensure experimental consistency, the zero-depth position was determined under the objective of the Raman microscope. The sample stage was raised until the topographical features of the sample were in optical focus with reflected light and then the C–H stretching Raman transition was maximized using the 532 nm laser source. This position was defined as the origin (film surface) and the depth was determined from the stage Z direction micrometer.

Confocal Laser Scanning Microscope Techniques

The spatial imaging capability of CLSM was explored for rapid confirmation as to whether grape juice’s fluorescence can be utilized to detect penetration of grape juice into polymer and paint films. The CLSM experiments were performed using the Nikon A1 multispectral fluorescence microscope in the imaging center on the St. Paul campus of the University of Minnesota. Nikon A1si is a confocal system equipped with a point-scan scan head and mounted on a Nikon Ti2000E inverted, fluorescence microscope with DIC optics. A 560 nm laser at 40 mW was used in the fluorescence imaging and the NIS Elements imaging software was used to control data acquisition and analysis.

Stain Resistance and Washability

Sample preparation and testing conditions were described in a previous paper.²⁴ The paint films were prepared on the Leneta white or black scrub test panels using a 7-mil DOW bar and allowed to dry at 25°C and 50% RH for a minimum of three days. Common household stains including mustard, ketchup, hot coffee, grape juice, wine, and blue fountain ink, were tested. For liquid stains like grape juice, a strip of single-ply paper towel was used to hold the liquid stains in place. The residence time of the stain (staining time) in this study varied from 20 min to 24 h to span the range of MPI cleansability and ASTM washability tests. The stained panels were rinsed under running tap water, and excess stains were gently removed by a soft brush before a sponge wash for 100 cycles with water or other alkaline cleaning solutions (e.g, Formula 409®, Leneta standardized nonabrasive scrub medium). The Garner Straight Line Washability and Wear Abrasion Machine was used for the washability test.

Stain resistance can be visually assessed by comparing the water rinse section before and after staining and water rinse. The degree of staining was determined using the ΔE values measured by a BYK-Gardner colorimeter of unstained vs stained and washed portions of the paint film. For each sample, two paint drawdowns were prepared on the same Leneta chart to ensure that drying, staining, and washing conditions were the same, and duplicated color measurements and CRM experiments were recorded.

Results and Discussion

CRM is an effective method for quantifying the distribution of individual constituents in heterogeneous systems based on the unique Raman spectra of different materials.^1-9 The first two examples discussed in this article illustrate that CRM can be employed to map relative spatial locations of various chemical species even in pigmented, multi-component coating systems. The methods described in this article can be applied to characterize component interactions and distributions in various polymer blends as well as other additives present in paint formulations. Problems such as material stratification, segregation, or migration across interfaces due to formulation instability or material incompatibility can thus be identified and addressed.

Example 1: Spatial Distribution of Styrenated Additive in Acrylic Paint Films

The effects of an alkali soluble resin on the morphology and mechanical properties of neat latex films have been reported previously.² CRM was employed to quantify the distribution of the SA within the polymer films. Since the acrylic polymer contained no aromatic moieties, quantitative analysis was done by comparing the intensities of Raman transitions associated with phenyl rings (I_ph) on the additive to that of carbonyl groups (I_C=O) of the acrylic polymer. The Raman intensity ratios confirm a uniform spatial distribution of the SA throughout the latex film.² The previous CRM investigation was performed on the clear latex films containing 10% and 20% SA. The SA concentration was 2.0% in the paint formulation, equivalent to 3.7% based on paint solids. The same quantification approach (I_ph/I_C=O) was applied in this study to investigate SA distribution in the pigmented paint films.

Figure 1 shows the Raman spectra with depth increments of 10 microns (top red trace) up to 50 microns (bottom blue trace) into the film of Paint I-SA. The presence of rutile titanium dioxide (TiO₂) is clearly seen from the Raman spectra. The Ti-O stretching and bending transitions are in the spectral range of 650 to 100 cm^-1.²⁵ All spectral intensities decrease progressively due to depth attenuation as CRM probes deeper into the nontransparent paint film. Contrasted with the latex films, the depth attenuation is dramatic for the TiO₂-containing paint samples owing to film opacity.

Fortunately, the strong Raman transitions of TiO₂ at low wavenumbers do not impede the analysis of other spectral features belonging to the acrylic polymer and the SA. Figure 2 displays the Raman spectra of interest from 4000 to 800 cm^-1 for the Paint I-SA film. The Raman transitions indicating the presence of the SA component are marked in Figure 2: aromatic C–H bending and ring breathing at 998 cm^-1, aromatic ring motion at 1598 cm^-1, and aromatic C–H stretching at 3060 cm^-1. The intensity ratios of styrene to acrylic signature bands at various depths were employed for quantification of SA concentration, so that the influence of signal attenuation was minimized.

Figure 3 shows the Raman intensities of TiO₂ and the styrene transition at 998 cm^-1 as a function of depth in Paint I-SA film. The intensity of the styrene peak at 998 cm^-1 is weaker than TiO₂; thus, a multiplication factor of 20 is used to display the styrene signals on the same graph with TiO₂. Figure 3 demonstrates that the depth signal attenuation is similar for TiO₂ and the styrene component, indicating a uniform SA distribution across the film thickness.

The distribution of the SA in the paint film was quantified by comparing the styrene peaks at either 998��cm^-1 or 1598 cm^-1 to the C=O peak at 1727 cm^-1 representing the acrylic polymer. Figure 4 displays the Raman intensity ratios of both styrene transitions to the carbonyl stretching band. The intensity of the 998��cm^-1 transition is greater than that of the 1598 cm^-1 transition, its styrene to acrylic peak intensity ratio I_ph/I_C=O is therefore higher, consistent with the spectral features shown in Figure 2. Nevertheless, the two sets of data in Figure 4 show little variation as a function of film depth and both lead to the same conclusion, i.e., the SA concentration, proportional to the intensity ratio, remains unchanged across the thickness of the paint film. Figure 4 represents the averaged intensity ratios from three run profiles at different sample locations. The small error bars displayed in the graph demonstrate reasonably constant lateral concentrations of SA as well as good reproducibility of the Raman depth profiling. The results further support the conclusion that the SA is uniformly (i.e., a constant concentration profile) distributed in the pigmented paint films. The CRM experiments yielded an excellent agreement with the previous finding regarding uniform distribution of the styrene additive in the neat latex films.²

Example 2: Surfactant Enrichment on Paint Film Surface

Quantification by standard peak-fitting or intensity ratio requires a “clean” Raman spectrum and strong material-specific Raman bands, as exemplified by the CRM analysis of SA distribution. Difficulties often arise in CRM analysis because of sample fluorescence, baseline shifts, and/or overlapping peaks. A smoothed second-derivative processing approach has been shown to provide an alternative quantification procedure.²⁶ The second derivative sharpens those subtle features that are difficult to quantify in the original spectra and therefore overcomes some of the above-mentioned problems. The example below demonstrates the advantage of this data analysis strategy.

The Raman spectra of the Paint I film from the air-film interface (red trace) down to 50 microns (blue trace) are displayed in Figure 5. The Paint I film exhibits compositional inhomogeneity in the first eight microns of the surface, but the spectral differences are small and can be better discerned in the second derivatives of the Raman spectra. Figure 6a shows the second derivatives of Raman spectra in the C–H regions (3000~2800 cm^-1) for the surface of Paint I film. The transitions between 2920 and 2925 cm^-1 correspond to the C–H stretching of methylene linkage –(CH₂)_n– in long chain hydrocarbon structures. The red trace represents surface (0–4 mm) and the green trace represents depth (4–8 mm). Higher absolute y-axis values in Figure 6 indicate higher Raman intensities in this spectral region (the negative sign is a result of the derivative calculation).

The differences in Figure 6a thus indicate that a compound rich in aliphatic hydrocarbon units resides in the top four microns of the Paint I film, presumably the C14–C16 alpha-olefin sulfonate surfactant used in the polymerization of the acrylic latex (AP1). Figure 6b displays the second derivatives of Raman spectra for the surface of the Paint I film after water rinse. The second derivatives of Raman intensities are now much more similar and the difference across film depth is no longer detected. This suggests that a simple rinse of the top surface with water is all that is required to eliminate most of the surface compositional inhomogeneity found in the top eight microns of the Paint I film. The small spectral changes in band shapes are likely a consequence of the rinse process as the fine surface topography may have been altered.

Surfactant leaching is an issue frequently encountered with exterior waterborne paints, especially in high humidity environments. Aside from the unsightly “snail trail” on the paint surface, a nonuniform surfactant distribution caused by surfactant migration can lead to other performance problems like blister formation and adhesion failures. Surface enrichment of surfactants has been reported for many latex films involving different nonionic and anionic surfactants used in synthesis of emulsion polymers.^27-34 In paint formulations, the concentration of polymerization surfactant is significantly reduced. The C14–C16 alpha-olefin sulfonate surfactant was used at 0.72% based on total monomer in the synthesis of AP1. Its concentration in the liquid paint sample was 0.15 wt% (concentration in water ~0.32 wt%), which is still 2~5 times higher than its critical micelle concentration (cmc) of 0.07 wt%. This example demonstrates that the sensitivity of CRM allows for detection of migration and surface enrichment of low concentration species like polymerization surfactant on the paint film surface.

Example 3: Characterization of Stain Resistance, Stain Penetration, and Stain Removal

Stain resistance, and easy and clean stain removal are desirable properties of interior wall paints and a key performance category that the Consumer Reports® tests and ranks for interior paint products.³⁵ Stain resistance is the ability of a coating surface to withstand discoloration caused by contact with liquids and solid particulates. A washability test assesses the relative ease and completeness of removal of a specific soilant from a coating surface by scrubbing. The qualitative generalization was made in a previous study that good washability usually corresponds well to high resistance to the water-based household stains.²⁴ Two confocal scanning microscopy techniques, CRM and CLSM, are explored in this study to characterize washability or stain removal in relation to stain resistance and stain penetration of polymer binders as well as the effect of staining time and washing protocols.

Selection of Representative Staining Material and Paint Samples

Welch’s Concord grape juice was used to probe stain adsorption and penetration. It was chosen as the surrogate staining agent in this study because 1) it is a common food stain and some of the color-contributing chemicals in grape juice are also responsible for the color, flavor, and aroma of red wine.³⁶ Consequently, any finding from grape juice is expected to correlate with red wine, another liquid stain often specified in the paint washability test; and 2) the colors of other food stains such as mustard, coffee, and tea are also attributable to phenolic compounds. The spectroscopy and microscopy methods developed for grape juice are likely to be applicable to those food and beverage stains that have their own specific absorption and emission spectra.

The intense color of Concord grape juice has been attributed principally to the presence of anthocyanins, which are water-soluble fluorescent pigments.³⁷ Figure 7 displays the emission spectrum with excitation at 532 nm for Welch’s Concord grape juice diluted in deionized water. The emission spectrum of the juice was recorded from 545 to 800 nm since fluorescence is expected to be lower energy emissions at longer wavelengths than the excitation wavelength. The 532 nm excitation wavelength was chosen to match the laser wavelength of the Almega Raman system. This wavelength selection will be discussed in the section on “Characterization of Stain Adsorption, Penetration, and Removal Using CRM” later in this article.

For a given stain, stain resistance and washability are affected by the composition, surface characteristics, and porosity of the paint film. Depending on the binder choice and the paint formulation strategy, stain resistance and washability of the resulting coating can vary greatly. Residual color from each stain is typically measured as color change, DE.�� Small DE values are thus desirable, denoting slight or no staining of the paint surface. Figure 8 plots the total DE values of Paint I and Paint II surfaces after the 5-stain washability test. Paint I and Paint II are based on the two acrylic latexes, AP1 and AP2, in the same paint formulation. Figure 8 illustrates vividly the influence of binder choice, highlighting that AP2 drastically improves removal of the five stains investigated. Therefore, Paint I and Paint II with strikingly different stain resistance and washability performance, are ideal samples that can be used to provide contrast and validation for the observations by confocal scanning microscopy techniques.

Visualization and Quantification of Stain Penetration Using CLSM

Since the presence of fluorescent substances in the grape juice stain has been confirmed (Figure 7), the fluorescence imaging capability of CLSM can be employed for visualization and quantification of juice penetration into coating films. For simplicity of the feasibility check, grape juice was applied to the clear polymer films via a strip of single-ply paper towel and let stand for two hours. A quick screening of the stained AP2 film produced promising results: the color-coded spectral images of the grape juice and acrylic polymers are pink and blue, respectively (Figure 9). This distinct microscopic contrast is encouraging and can be utilized to provide spatial differentiation for the grape juice stain in the polymer matrix. Visualization and quantification of grape juice penetration depth is made possible.

The benefit of fluorescence imaging and depth profiling is fully illustrated in Figure 9 which depicts layer-by-layer optical sectioning of the stained latex film. The pink color indicating the presence of grape juice was found at the film surface and the first two microns under the surface. Figure 10 shows the CLSM images of the stained polymer film after the sponge wash using Formula 409, a common all-purpose household cleaner. The pink color of grape juice was no longer visible even on the surface image. Figure 10 suggests relatively shallow penetration of grape juice into the well-coalesced film of the low T_g polymer. After cleaning with Formula 409, the grape juice stain was completely removed from the polymer film surface.

Using this CLSM technique, red wine and coffee were found to also produce colored spectral images distinguishable from the acrylic polymer matrix, confirming the utility of this powerful tool for visualizing a wide range of staining materials. This is the first reported application of CLSM for direct visualization of spatial distribution and penetration of soilants or stains in coating systems. The meaningful extension of this technique to the pigmented paint samples met a few challenges. Our preliminary investigation indicated that higher laser powers were required because the intensity of fluorescence was reduced in the opaque paint films, but the staining materials decayed rapidly under these conditions. Quantitative measurement and visual display of grape juice penetration into latex films and clean stain removal afterwards demonstrate the usefulness and advantage of fluorescence imaging, which hopefully will encourage more research efforts to advance the method development for more complex coating systems.

Characterization of Stain Adsorption, Penetration and Removal Using CRM

In this study, the initial objective was to determine whether a grape juice-specific Raman transition can be found and utilized. Figure 11 shows the Raman shifted spectrum of partially dried grape juice. Not surprisingly, strong fluorescence dominated the spectrum and no specific Raman transitions were detected on the fluorescence envelope. This interpretation is corroborated by the emission spectrum in Figure 7.

The principle of Raman spectroscopy is based on the inelastic scattering of monochromatic light when the frequency of photons changes upon interaction with a sample. The frequency of reemitted photons from the sample is shifted in comparison with the original monochromatic frequency, which is known as the Raman Effect. This energy shift is characteristic for the type and coordination of the molecules involved in the scattering process and, therefore, provides information about vibrational and rotational energies of molecular bonds.

Figure 12 compares the unshifted Raman spectrum with the emission spectrum from grape juice using 532 nm excitation. The Raman excitation laser is also at 532 nm, which allows the unshifted Raman spectrum to be recalculated from the Raman shifted spectrum (Figure 11) and the excitation wavelength at 532 nm. The two spectra in Figure 12 are similar in overall shape. The observed shift in the maximum is possibly a consequence of recording the fluorescence in a partially dried state for the Raman experiment and in water for the emission spectrum. One way to minimize the fluorescence is to use a longer wavelength laser; the drawbacks, however, would be lower Raman intensity. It will be difficult with a lower energy laser to resolve the low surface concentration effect of color bodies in the grape juice.

The fluorescence envelope was also observed for Paint I and Paint II samples stained with grape juice. Figure 13 displays exemplary Raman spectra taken on the film surface of Paint I. A “hump” between 4146~3071 cm^-1 was present in the spectra of the stained and water-rinsed paint films, consistent with the Raman spectrum of the dried grape juice in the same spectral range. This region of the fluorescence envelope is denoted as “high frequency curvature.”

With no juice-specific Raman peaks, a novel data analysis methodology was developed to exploit the fluorescence envelope in the “high frequency curvature” region. Semi-quantitative analysis of grape juice concentration and penetration depth was performed by normalizing the area under the curvature 4146~3071 cm^-1 to the area of C–H stretching transition between 3075~2798 cm^-1 to account for experimental variables. This intensity ratio signifies the concentration of grape juice. The results from depth profiling of the unstained and the stained samples of the two paints are presented in Figure 14. The intensity ratios of both unstained samples, as expected, are relatively constant. After two hours of contact with grape juice followed by a gentle water rinse, the samples displayed an intensity attenuation for the top surface layers. The minimum in the depth profile plot was taken as the end point of grape juice penetration.

The difference in the surface concentration of grape juice between the two paint films was supported by the extent of discoloration. Figure 15 provides photos of two stained and water-rinsed paint panels along with the measured color change or DE due to grape juice staining. It is clear from the visual comparison of the two photos and the measured DE values that Paint II has greater stain resistance and better washability than Paint I. A highly stain-resistant coating surface is more difficult for stains to adhere to and allows easier removal of any residual stain during washing cycles. For Paint II, a simple water rinse removed most of the stains, suggesting that the adhesion and adsorption of grape juice on the paint surface is weak and the stain penetration is minimal.

Figure 16 illustrates the effects of grape juice staining time on Paint I and Paint II films. In both cases, the intensity ratio of the high frequency curvature area to the C–H stretching area increases with increasing stain residence time. Figure 16a shows that longer staining time leads to a higher surface concentration of the color compounds left by grape juice. A large upward shift is seen when the staining time changes from 20 min to two hours. Beyond two hours, the shifts become smaller with further increase of staining time. Compared to Paint I, a smaller change in surface concentration and penetration distance was observed for Paint II. This indicates that the impact of longer staining time was reduced in a more stain-resistant coating system.

Besides higher surface concentration, the grape juice stain was also detected at increased depth in Paint I with longer staining time. For Paint I (Figure 16a), the penetration depth was 6~8 microns after staining for 20 min. When grape juice was left on the paint films for two hours or longer, its penetration depth increased to 10~12 microns. The maximum stain penetration depth in Paint II was ~6 microns after the longest staining time of 24 h. Lower surface concentration and reduced stain penetration result in an easy-to-clean paint surface. These conclusions from semi-quantitative analyses of Raman spectra are in excellent agreement with the color measurements on the same panels.

Color change as a function of grape juice staining time is shown in Figure 17 for Paint I and Paint II. The DE or degree of discoloration corresponds to the concentration of staining compound adsorbed by and entrapped in the coating. Similar to the magnitude of curve shifting in Figure 16, the DE value increases linearly with the increase of the grape juice staining time from 20 min to four hours. Further increases of staining time up to 24 h did not cause substantially heavier discoloration, possibly because the system was approaching adsorption equilibrium after four hours. In the range of staining time studied, Paint II consistently exhibited less color change from the grape juice stain. The slope of the DE change was not as steep as that of Paint I in the linear region of staining time between 20 min and four hours. The results shown in Figures 8, 16, and 17 are self-consistent and all demonstrate that Paint II has greater stain resistance and better washability.

Figure 18 shows the depth profiling results for the stained and cleaned Paint I films. The experimental variables compared in Figure 18 are staining time (2 vs 24 h) and cleaning procedure (water rinse vs water rinse followed by sponge wash). To avoid the likely spectral interference of other cleaning chemicals, only the results from the water rinse and the sponge scrubbing using water are reported here to illustrate the effect of washing protocol. For both sets of data, the intensity ratio indicative of stain concentration is lower after the sponge wash using water. The mechanical forces involved in the sponge washing procedure more effectively removed the grape juice stain than just a water rinse. The minimum between 10~12 microns in the intensity ratios, however, was not changed by the more forceful sponge washing. Figure 18 also demonstrates that once stain penetration occurred, the staining depth was basically unchanged by mechanical sponge scrubbing. The results indicate that film erosion was not the mechanism for stain removal from the Paint I surface, and that the thickness of dried grape juice did not introduce significant measurement error. Therefore, the data in Figure 18 verified that depth determination using CRM is reasonably accurate and reproducible.

Latex paints all have varying degrees of porosity, depending on the formulation (e.g., choices of binder, pigments, and PVC). The penetration of a liquid in porous systems is usually viewed as a spontaneous process driven by capillary forces.³⁸ Penetration of a liquid stain can be estimated using the classic Lucas-Washburn-Rideal (LWR) equation³⁹:��

where γ and η are the surface tension and the viscosity of the stain, respectively; θ is the contact angle; and R is the porosity of the coating film. In addition, the LWR equation also describes the dependence of penetration distance on time: .

No attempt was made to model grape juice penetration with regards to its time dependence because of the data range. Raman depth profiling was done in 2-μm step-scan increments (i.e., a noncontinuous variable) and the deepest stain penetration distance was 10~12 microns (i.e., a relatively small range from 4 to 12 microns). Additionally, to hold the liquid stains in place, the grape juice was applied to a layer of absorbent paper towels covering the paint surface. This particular method of stain application complicates or alters the interaction of��the stain with the coating surface. Consequently, the experimental data are not expected to strictly follow the theoretical predictions of a stain’s wetting, adhesion, and penetration behaviors. However, the semiquantitative determination of stain adhesion/adsorption, penetration, and removal using the CRM data analysis procedure was validated by the visual observations and washability results. Not surprisingly, staining time and washing procedures both influence stain adsorption/desorption and ultimately stain removal or washability. The results of the two acrylic paints clearly demonstrate that good washability can be achieved by optimizing stain resistance of paint films and minimizing stain adsorption and penetration. A stain-resistant coating like Paint II is thus highly preferred because it can minimize or prevent wetting, adhesion, and penetration of staining materials.

Conclusions

This article reports on surface and depth profile characterization of waterborne coatings using confocal Raman microscopy (CRM) and confocal laser scanning microscopy (CLSM). Using Raman intensity ratios, a uniform spatial distribution of a styrenated additive (SA) throughout the acrylic paint film was confirmed. The second derivative analysis of CRM data revealed that surfactant migration was detectable in the pigmented paint film as well, even though the concentration of polymerization surfactant is significantly reduced in the paint formulation. Utilizing the intrinsically fluorescent color compounds present in the Welch’s grape juice, CLSM was explored for direct visualization and quantification of stain penetration and removal. A novel data analysis approach developed in this work allowed the use of CRM for semiquantitative characterization of adsorption, penetration, and removal of grape juice stains. Fluorescence has been an unwanted limitation of Raman spectroscopy, but this shortcoming was effectively utilized in this work. The results correlated well with the data obtained by the conventional stain removal or washability test and at the same time provided valuable insight into staining and stain removal processes.

Both CRM and CLSM offer chemical specificity and high-resolution spatial analysis and are thus powerful tools for determining chemical composition and component distribution of multi-component coating systems. The examples described in this article effectively demonstrate the usefulness and advantage of these confocal scanning techniques and the need for continued method development to expand their application in coatings research.

Acknowledgments

The confocal fluorescence images were generated by Dr. Gang Pu and Dr. Jilin Zhang in Professor Steve Severtson’s group at the University of Minnesota. Fluorescence absorption and emission spectra were collected by Dr. Evgeny Danilov in the Laboratory for Imaging and Kinetic Spectroscopy at North Carolina State University. Their assistance is greatly appreciated.

References

1. “Confocal Raman Microscopy,” in Springer Series in Optical Sciences 158, Chapt. 11, Diening, T., Hollricher, O., and Toporski, J. (Eds.), Heidelberg: Springer-Verlag Berlin, 2010.
2. Wu, W., Miller, C.M., and Severtson, S.J., “Alkali-soluble resins (ASR) and acrylic blends: influence of ASR distribution on latex film and paint properties,” J. Coat. Technol. Res., 13, (4) 655-665 (2016).
3. Schrof, W., Beck, E., Königer, R., Reich, W., and Schwalm, R., “Depth profiling of UV cured coatings containing photostabilizers by confocal Raman microscopy,” Prog. Org. Coat., 35, 197-204 (1999).
4. Adamsons, K., “Chemical surface characterization and depth profiling of automotive coating systems,” Prog. Polym. Sci., 25, 1363-1409 (2000).
5. Marton, B., van de Ven, L.G.J., Otto, C., Uzunbajakava, N., Hempenius, M.A., and Vanso, G.J., “A depth-resolved look at the network development in alkyd coatings by confocal Raman microspectroscopy,” Polymer, 46, 11330-11339 (2005).
6. Posset, U., Gigant, K., Schottner,G., Baia, L., and Popp, J., “Structure-property correlations in hybrid Sol-Gel coatings as revealed by Raman spectroscopy,” Opt. Mater., 26, 173-179 (2004).
7. Zhang, W.R., Lowe, C., and Smith, R., “Depth profiling of clear coil coating by confocal Raman microscopy,” Prog. Org. Coat., 66, 141-148 (2009).
8. Zhang, W.R., Zhu, T.T., Smith, T., and Lowe, C., “An investigation on the melamine self-condensation in polyester/melamine organic coating,” Prog. Org. Coat., 69, 376-383 (2010).
9. Zhang, W.R., Zhu, T.T., Smith, T., and Lowe, C., “A non-destructive study on the degradation of polymer coating I: Step-scan photoacoustic FTIR and confocal Raman microscopy depth profiling,” Polym. Test., 31, 855-863 (2012)
10. Dupuie, J.L., Weber, W.H., Scholl, D.J., and Gerlock, J.L., “Clearcoat analysis in isolated and multi-layer paint systems by confocal Raman microscopy,” Polym. Degrad. Stab., 57, 339-348 (1997).
11. Schrof, W., Beck, E., Etzrodt, G., Hintze-Brüning, H., Meisenburg, U., Schwalm, R., and Warming, J., “Depth-resolved characterization of UV cured coatings by confocal Raman and two-photon microscopy,” Prog. Org. Coat., 43, 1-9 (2001).
12. Ketler, W.H. and Richter, G., “Investigation on topology of platelet-like effect-pigments in automotive surface-coating,” Prog. Org. Coat., 31, 297-306 (1997).
13. Schneider, O., Hevbare, G.O., Scully, J.R., and Kelly, R.G., “Confocal Laser Scanning Microscopy as a Tool for In Situ Monotoring of Corrosion Underneath Organic Coatings,” Electrochem. Solid-State Lett., 4 (12) B35-B38 (2001).
14. Faÿ, F., Linossier, I., Peron, J.J., Langlois, V., and Vallée-Rehel, K., “Antifouling activity of marine paints: study of erosion,” Prog. Org. Coat., 60, 194-296 (2007).
15. Li, Y. and He, B., “Characterization of ink pigment penetration and distribution related to surface topography of paper using confocal laser scanning microscopy,” BioResources, 6 (3) 2690-2702 (2011).
16. Bearman, G. and Levenson, R., “Biological Imaging Spectroscopy.” In: Vo-Dinh, T. (Ed.), Biomedical Photonics Handbook, 1st ed., CRC Press: Boca Raton, FL, 8-1–8-5, 2003.
17. Hiraoka, Y., Shimi, T., and Haraguchi, T., “Multispectral imaging fluorescence microscopy for living cells,” Cell Struct. Funct., 27, 367–374 (2002).
18. Levenson, R.M. and Mansfield, J.R., “Multispectral imaging in biology and medicine: slices of life,”�� Cytometry A, 69 (8) 748–758 (2006).
19. Mansfield, J.R.,�� Hoyt, C., and Levenson, R.M., “Visualization of microscopy-based spectral imaging data from multi-label tissue sections,” Curr. Protoc. Mol. Biol., 84, 14.19.1–14.19.15 (2008).
20. Nie, S., Chiu, D.T., and Zare, R.N., “Robing individual molecules with confocal fluorescence microscopy,” Science, New Series, 266 (5187) 1018-1021 (1994).
21. Zhou, L. and El-Deiry, W.S., “Multispectral fluorescence imaging,” J. Nucl. Med., 50 (10) 1563-1566 (2009).
22. Virostko, J., Xie, J., Hallahan, D.E., Arteaga, C.L., Gore, J.C., and Manning, H.C., “A molecular imaging paradigm to rapidly profile response to angiogenesis-directed therapy in small animals,” Mol. Imaging Biol., 11, 204–212 (2009).
23. Manning, H.C., Shay, S.D., and Mericle, R.A., “Multispectral molecular imaging of capillary endothelium to facilitate preoperative endovascular brain mapping,” J. Neurosurg., 110, 975–980 (2009).
24. Wu, W., Kaufman, M., Schneider, J., and Grieb, R., “New acrylic polymers for high performance interior wall paints,” JCT CoatingsTech, 29-35 (August 2016).
25. Lukačevic, I., Gupta, S.K., Jha, P.K., and Kirin, D., “Lattice dynamics and Raman spectrum of rutile TiO₂: The role of soft phonon modes in pressure induced phase transition,” Mater. Chem. Phys., 137, 282-289 (2012).
26. Nichols, M.E., Seubert, C.M., Weber, W.H., and Gerlock, J.L., “A simple Raman technique to measure the degree of cure in UV curable coatings,” Prog. Org. Coat., 43, 226-232 (2001).
27. Aramendia, E., Mallégol, J., Jeynes, C., Barandiaran, M.J., Keddie, J.L., and�� Asuza, J.M., “Distribution of surfactants near acrylic latex film surfaces: A comparison of conventional and reactive surfactants (surfmers),” Langmuir, 19 (8) 3212-3221 (2003).
28. Du Chesne, A., Gerharz, B., and Liser, G., “The segregation of surfactant upon film formation of latex dispersions: an investigation by energy filtering transmission electron microscopy,” Polym. Int., 43 (2) 187-196 (1997).
29. Zhao, C.L., Holl, Y., Pith, T., and Lambla, M., “FTIR-ATR spectroscopic determination of the distribution of surfactants in latex films,” Colloid Polym. Sci., 265 (9) 823-829 (1987).
30. Gundabala, V.R., Zimmerman, W.B., and Routh, A.F., “A model for surfactant distribution in latex coatings,” Langmuir, 20 (20) 8721-8727 (2004).
31. Xu, G., Dong, J., Zhang, J., Severtson, S.J., Houtman, C.J., and Gwin, L.E., “Characterizing the distribution of nonylphenol ethoxylate surfactants in water-based pressure-sensitive adhesive films using atomic-force and confocal Raman microscopy,” J. Phys. Chem. B., 112 (38) 11907-11914 (2008).
32. Zhang, J. and Severtson, S.J., “Characterizing the distribution of sodium alkyl sulfate surfactant homologues in water-based, acrylic pressure-sensitive adhesive films,” J. Phys. Chem. B., 115 (25) 8138-8144 (2011).
33. Zhang, J., Zhao, Y., Dubay, M.R., and Severtson, S.J., “Surface enrichment by conventional and polymerizable sulfated nonylphenol ethoxylate emulsifiers in water-based pressure-sensitive adhesive,” Ind. Eng. Chem. Res., 52 (25) 8616-8621 (2013).
34. Mallégol, J., Gorce, J.-P., Dupont, O., Jeynes, C., McDonal, P.J., and Keddie, J.L., “Origins and effects of a surfactant excess near the surface of waterborne acrylic pressure sensitive adhesives,” Langmuir, 18 (11) 4478-4487 (2002).
35. Consumer Reports, January 2017 Online issue.
36. Agati, G., Matteini, P., Oliveira, J., de Freitas, V., and Mateus, N., “Fluorescence approach for measuring anthocyanins and derived pigments in red wine,” J. Agric. Food Chem., 61 (42) 10156–10162 (2013).
37. Sastry, V.L.N., “Nature of pigments in Concord grapes and their behavior during heat processing and storage,” Retrospective Theses and Dissertations, Paper 12809, Iowa State University.
38. Hamraoui, A. and Nylander, T., “Analytical approach for the Lucas–Washburn Equation,” J. Colloid. Interface Sci., 250, 415-421 (2002).
39. Washburn, E.W., “The dynamics of capillary flow,” Phys. Rev., 17 (3) 273-283 (1921).

Industry Profile: Dr. Wenjun Wu

Dr. Wenjun Wu is currently a Research Fellow at Arkema Coating Resins and global Waterborne Technology and Innovation Leader. She has developed numerous technology platforms and successfully commercialized over 20 new binders and rheology modifiers for applications in architectural coatings, personal care, and aircraft anti-icing fluids. A certified Six Sigma Black Belt, she has also held positions such as research leader and intellectual capital manager in The Dow Chemical Company.

Wu received B.S. and M.S. degrees from Peking University, and a Ph.D. from the University of Southern California. She was a postdoctoral researcher at Henkel Corporation from 1995 to 1996.

Wu is the recipient of many industry honors. In 2019 alone, she has been recognized for her work with two prestigious awards. In March, she and co-author Dr. Christopher Miller of Arkema received the European Coatings Show Best Paper Award for their paper, “Developing a Deeper Understanding of the Effect of Latex Design Parameters on Final Coating Film Properties.” In February, she was recognized with the Siltech Best Paper for Innovation Award at the 46th Waterborne Symposium in New Orleans. Prior to that, she has been the recipient of an ACS Best Paper Award, a Roon Foundation Award, and Society of Cosmetic Chemist Award. In addition, Wu has received 39 internal awards including two-time Arkema CEO Performance Awards, two best paper awards, and two technology center awards from The Dow Chemical Company. She is a recognized expert on emulsion polymerization, colloid and surface science, latex polymer characterization, as well as design of experiment (DOE) and statistical analysis.

When asked about what motivates her work, Wu responds, “I find a great deal of excitement in problem solving and continuous learning. I’m so blessed to have worked with and learned from many great scientists during my career. Industrial research and development has been a fun and satisfying journey because of ample opportunities to provide technical solutions to enhance and enrich people’s lives.” A personal goal is helping to develop young talents. She wishes to devote more time to coaching and mentoring researchers who are interested in the coatings field.

As evidenced in this article, Confocal Microscopy techniques have been an area of focus. According to Wu, “Non-destructive Confocal Microscopy techniques have been employed by other researchers to study curing and weathering of clear coating finishes. We have been able to extend Confocal Techniques for characterization of component distribution in non-transparent, pigmented coatings. By sharing our research results, we hope to encourage continued method development efforts to further expand the application of Confocal Techniques for more complex coating systems.”

This paper received the Siltech Best Paper for Innovation Award at the 46th Waterborne Symposium, February 24–March 1, 2019, in New Orleans, LA.

CoatingsTech | Vol. 16, No. 4 | April 2019

This article addresses the value of collecting mechanical and thermal property data for better understanding the development of coating properties. As a case study, the rheological and thermal characterization data of two solventborne systems (a polyamide-cured epoxy and a moisture-cured RTV silicone) and a 100% solids system (polyetheramine-cured epoxy) are discussed along with the experimental considerations when conducting these types of studies. For the two solventborne systems, dynamic mechanical analysis (DMA) and differential scanning calorimetry (DSC) were used to characterize the liquid-to-solid transformation of these functional coatings. In particular, for the polyamide-cured epoxy, the effects of solvent addition and cure temperature (ambient vs heat-cured) on the resulting properties were investigated. For the 100% solids polyetheramine-cured epoxies, an in situ DMA immersion test for measuring the solvent resistance of coatings is presented. For these materials, lower crosslink density formulations exhibited more dramatic decreases in modulus with solvent uptake.

Introduction

Polymer coatings are applied to a substrate to impart or improve a particular property or functionality. Among others, the properties that can be enhanced through the use of coatings are aesthetic, mechanical, protective, anticorrosive,¹ conductive,² and biomedical³ in nature. Polymer-based dispersions or solutions use water or organic solvent(s) as the vehicle for applying the coating to the substrate, where the use of solvent often helps improve the film formation process.

For these coatings to be successful, transformation of the initially liquid material to a solid, film-formed state of given properties is essential. Although extensive analytical tests or techniques are available for characterizing the liquid coating and the final solid film, less characterization effort is devoted to measuring the properties during the coating drying or curing process. The reason for this may be that film formation involves a change of physical state, loss of material, and property changes that vary by many orders of magnitude so that finding an experimental apparatus and geometry can be challenging. Despite the difficulties, characterization of the transformation from the liquid to solid state is critical to understanding the resulting coating properties. As a result, these data are necessary to properly specify a material for a given application, as well as for designing next-generation coatings that address the limitations of current materials.

Various techniques are available for measuring the properties of polymer coatings. Among others, these include thermal methods^4,5 such as differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), and thermomechanical analysis (TMA), mechanical methods^6,7 such as DMA and nanoindentation,⁸ as well as impedance spectroscopy.^9,10 Because these techniques are complementary, mechanical methods are often coupled with thermal methods for characterizing coatings.^11-15

One of the challenges associated with performing a more complete characterization of the mechanical properties of coatings is the ability to obtain free-standing pieces. In the absence of a free-standing polymer film, experimental conditions that ensure that the results are not dominated by the typically stiffer and more massive substrate must be employed. Despite the challenge of a substrate, researchers have been able to characterize the mechanical properties of the polymer coating without removing the substrate.^8,16-18 Similar considerations apply when trying to use TMA to characterize coatings on substrates. To collect more complete data, efforts were made in this work to produce free-standing coating films

As a case study, two solventborne systems were chosen: a polyamide-cured epoxy and a moisture-cured room temperature vulcanizing (RTV) silicone. For the epoxy system, a diglycidyl ether of bisphenol A (DGEBA)-type resin was mixed with a high imidazoline content polyamide, and xylene as a solvent. A one-part, moisture-cured RTV silicone dispersion that can be used as a conformal coating was used for the silicone system. In addition, a 100% solids system comprised of a DGEBA-type epoxy resin cured with a blend of polyetheramines was examined.

These two solvent-based systems were chosen as representative examples of polymer coatings; commercial formulations may be more complicated. The use of these three systems in this work is for illustrative purposes only.

Experimental

Three systems were chosen to demonstrate the utility of using rheological and thermal techniques to characterize coatings. In particular, DMA and DSC were employed to characterize the rheological and thermal properties, respectively.

Briefly, DMA uses a sinusoidal deformation to measure the rheological properties of a material, including the storage modulus (E′ or G′), loss modulus (E″ or G″), and tan(d).¹⁹ Tan(d), which is equal to E″/E′ or G″/G′, is a measure of the degree of viscoelasticity of a material, or, more simply, the balance between liquid-like and solid-like and properties. Rheological properties are commonly determined as a function of temperature and frequency; both of these variables are illustrated in this work. For the rheological testing of coatings, free-standing films are preferred, since the results are not influenced by the mechanical properties of the substrate and the characterization can be more definitive. It is recognized that, for some coatings systems, obtaining a free-standing film may not be possible. As will be discussed below, atypical substrates were used in this work to afford free-standing films. While using the actual substrates associated with a given application has merit, the approach used herein enables a greater level of understanding of the coating system.

Specifically, an RSA G2 rheometer (TA Instruments) was used to measure the tensile dynamic mechanical properties for both coating systems as a function of cure time. This rheometer was also used to measure the mechanical properties of polyetheramine-cured epoxies during immersion in water as a demonstration of an in situ method for examining the solvent resistance of a coating.

DSC measures the heat flow into or out of a sample relative to a reference standard and enables one to measure properties such as the glass transition temperature ( T_g), heat of reaction, and degree of cure for thermosetting systems. DSC must be used with care when applied to water- or solventborne coatings as evaporation of residual solvent can often dominate the resulting thermal signal. For semi-crystalline thermoplastic polymers, DSC is useful for examining the degree of crystallinity, as well as the breadth of the crystal melting and crystallization profiles. Two heating scans were used in this study; the first scan measures the properties of the material in its current state, while the second scan determines the properties after nearly complete cure. For this work, a Q200 DSC (TA Instruments) was used. Given the relatively small sample size used for DSC (20 mg), representative sampling is important.

Initial experiments using TGA to examine solvent loss as a function of time for the polyamide/epoxy/xylene formulations were largely unsuccessful as the expected monotonic trend in loss of solvent was not observed. The relatively thick films used made it difficult to obtain representative coating specimens for testing. Since a gradient in solvent concentration through the thickness of the coating will exist until all of the solvent is lost, taking samples for TGA testing only from the bulk or the surface will result in data that are not characteristic of the coating.

Polyamide/Epoxy Formulations

A stoichiometric mixture of a DGEBA-type epoxy resin and a polyamide curative was prepared and served as a 100% solids control. In addition, a stoichiometric formulation of the same epoxy resin and polyamide curative was prepared with 20% by weight xylene (based on the total resin and curatives weights) to illustrate the impact of solvent on mechanical properties. For these epoxy-based formulations, about 1/32-in. thick specimens were cast into silicone rubber molds for subsequent rheological and thermal testing. Relatively thick coatings were used to facilitate sample handling and lengthen the time required for solvent removal. Rheological and thermal properties were measured after 2, 6, and 21 days of cure at ambient temperature and humidity. Because the results of initial DSC testing performed on the samples with xylene were dominated by solvent evaporation, subsequent DSC data were not collected on this solventborne system.

Heat-cured specimens of the 100% solids and xylene-modified formulations were also prepared to compare the properties of the coatings cured at ambient temperature with those ultimately attainable using an elevated temperature cure. Heat-curing consisted of one hour at 120°C in a forced air oven after 21 days of cure at ambient temperature.

The tensile dynamic mechanical properties of the ambient- and heat-cured polyamide/epoxy coatings (both with and without xylene) were measured as a function of temperature over the –50 to 120°C range using a 5°C/min ramp. This relatively high ramp rate was chosen to facilitate capturing the true mechanical properties of the coating and minimizing additional sample cure within the rheometer. Data were using a 6.28 rad/s deformation frequency.

Moisture-Cured RTV Silicone Formulation

An as-received RTV silicone dispersion sample was cast 20 mils wet on a silicone rubber substrate (to facilitate subsequent removal of the coating for mechanical testing). Although it is recognized that some of the solvent in the dispersion may diffuse into the silicone rubber substrate during cure thereby altering the cure rate, this shortcoming was deemed not to be particularly relevant for the purpose of this work. Immediately after casting, the coatings were placed in an environmental chamber held at ambient temperature and 30% relative humidity (RH). Free-standing films were carefully peeled from the substrate and the tensile dynamic mechanical properties of the coatings were measured after 1 hr, 24 hr, and 7 days of cure at ambient temperature and 30% RH. For improved statistics, two specimens were measured for each test and each time point.

For each time point, rheological properties were obtained in tension as a function of frequency over the 1 to 100 rad/s range. To minimize any further evaporation of solvent, testing was conducted at ambient temperature and did not use the RSA G2’s environmental chamber (which uses a flow of dry N₂ to control temperature).

Polyetheramine/Epoxy Blends

Because polyetheramine (PEA) curatives are available in a wide range of functionality and molecular weights, they offer the opportunity to make networks of defined structure. A series of PEA-based formulations was prepared with a DGEBA-type epoxy to demonstrate the use of an in situ rheological technique to examine the solvent resistance of coatings, which is especially important in the potential use of coatings in protective or secondary container-type applications.

A relatively low molecular weight diamine polyetheramine (PEA Low) and a relatively high molecular weight diamine polyetheramine (PEA High) were selected as curatives. Specifically, 60/40, 70/30, and 80/20 blends (by weight) of PEA Low/PEA High curatives were examined. For each blend, a stoichiometric mixture was prepared with a DGEBA-type epoxy resin, cast into ~10 mil-thick films, and heat-cured for 3 hr at 150°C in a forced air oven. The choice of these blends was made to produce materials that illustrate the impact of solvent uptake on a convenient experimental time scale.

In addition to measuring the tensile dynamic mechanical properties of these three formulations as a function of temperature (similar to that performed for the polyamide/epoxy materials), the rheological properties of these three materials were measured in tension as a function of immersion time (3 hr) in water at ambient temperature using the immersion fixture of the RSA G2. A schematic of the immersion apparatus is shown in Figure 1. Basically, this apparatus consists of a thin-film fixture housed inside a stainless steel cylinder enabling one to submerse a polymer in a fluid of interest (water for these experiments) and collect data as a function of time and/or temperature. This apparatus avoids the difficulties and limitations associated with the more traditional approach of collecting solvent resistance data by avoiding having to remove, handle, and load samples taken from a benchtop vial or vessel containing the polymer and fluid of interest. While water was chosen for these experiments, other fluids of interest (e.g., solvents, oils, etc.) can be used with this apparatus. The safety aspects of immersion experiments performed with a nonaqueous fluid should be considered prior to performing these experiments, especially for nonambient test conditions.

For these PEA/epoxy immersion experiments, 10 minutes of data were collected at ambient temperature prior to the addition of water at ambient temperature to serve as a baseline for the immersion data. Data were collected every 30 sec using a 6.28 rad/s deformation rate.

Results

Polyamide/Epoxy Formulations

Rheological Data

The tensile dynamic mechanical properties of the polyamide/epoxy/20% xylene sample after two days of cure at ambient temperature and humidity are shown in Figure 2. The data show a glassy region up to about 0°C, a T_g region from 10 to 50°C, and a rubbery plateau above T_g. The midpoint T_g, as evidenced by the peak in the tan(d) data, is 36±2°C. The onset T_g, as evidenced by the peak in the tensile loss modulus (E”) data is in the 5–20°C range. The increase in the tensile storage modulus (E’) with increasing temperature in the rubbery region indicates that additional cure and/or loss of xylene is occurring while the sample is in the rheometer. This increase in E’ is not surprising, considering that the sample may not be fully cured and that solvent is likely still present in the sample.

The data in Figure 2 are qualitatively similar to those for the other polyamide/epoxy samples (either with or without xylene). As a result, and for reasons of brevity, subsequent figures for the polyamide/epoxy samples only focus on the E’ data.

An overlay of the E’ data as a function of temperature and cure time for the xylene-based samples cured at ambient conditions, together with the data for the heat-cured sample, is shown in Figure 3. The midpoint T_g, minimum in E’, molecular weight between crosslinks (Mc), and E’ at 120°C data for these formulations are summarized in Table 1. The following equation was used to calculate Mc¹⁹:

Mc = [3ρRT]/E’

where ρ is the density, R is the universal gas constant, and T is the absolute temperature. For fully cured thermosets, the E’ data at T_g + 40°C are often used to calculate Mc. This was not done in this work since E’ data at T_g + 40°C would have resulted in additional cure for the partially cured materials. Instead, the minimum in E’ was used for the Mc calculations; this should result in Mc values that are more reflective of the as-cured material properties.

The Mc calculations assume a density of 1.0 g/cm³ and should be viewed only as estimates, as the materials are continuing to cure in the rheometer. The E’ data at 120°C provide a measure of the network properties after a defined amount of thermal exposure in the rheometer; for most of these samples, this additional exposure results in additional cure and/or loss of solvent, both of which will increase E’. Also calculated in Table 1 is a cure ratio defined by E’ at 120°C/minimum in E’. This ratio provides a relative measure of the amount of additional build in modulus (due to cure or solvent loss) that each sample experiences, with a larger ratio indicating that substantial property changes occurred in the rheometer.

The data in Table 1 show that for the xylene-based formulations cured at ambient temperature and humidity, the T_g and minimum in E’ increase with increasing cure time; this trend is expected as the degree of cure increases with increasing cure time. The decrease in the cure ratio with increasing cure time for the ambient-cured formulations provides additional evidence that the degree of cure increases with time. The similarity of the 6- and 21-day data, coupled with the fact that these data are lower than those observed for the heat-cured xylene-based sample, suggests there is a limit to the properties attainable at ambient temperature. The 50–55°C T_g for the 6- and 21-day ambient-cured samples suggests that these formulations are vitrified (or glassy) at ambient temperature, making it difficult to readily advance the degree of cure.

An overlay of the E’ data as a function of temperature and cure time for the polyamide/epoxy samples (without xylene) cured at ambient, together with the data for the heat-cured sample, is shown in Figure 4. The midpoint T_g, minimum in E’, Mc, E’ at 120°C, and cure ratio data for these formulations are also summarized in Table 1. The trends observed for the xylene-based formulations are also seen for these 100% solids materials. Relative to Figure 4, the difference in data between the heat-cured and ambient-cured samples is considerably larger than the comparable difference for the xylene-based materials.

To more effectively compare the effect of solvent addition and cure time, an overlay of the E’ data for all eight of the polyamide/epoxy samples is shown in Figure 5. In addition to the differences in midpoint T_g and the magnitudes of E’ in the rubbery region, the solvent-based systems have broader E’ vs temperature profiles compared to the corresponding 100% solids formulations. Other, more specific, observations from Figure 5 suggest that:

After two days of cure, the xylene-based formulation displays a 6°C lower midpoint T_g, higher Mc, and lower cure ratio compared to the 100% solids sample. The difference in midpoint T_g is relatively small compared to the ~15°C difference in onset T_g (as evidenced by the decrease in E’ as the samples enter the T_g region). The presence of solvent appears to have depressed T_g, but afforded a higher degree of cure. At ~120°C the E’ data for the two samples are comparable suggesting that they reach similar cure states.

After six days of cure at ambient conditions, the xylene-based epoxy displays a 4°C higher midpoint T_g and higher crosslink density (lower Mc) than the 100% solids sample. Similar to the trend observed after two days of cure, the onset T_g for the formulation with xylene is lower than the one without xylene, though the two six-day samples are more comparable in onset T_g than the two two-day samples. At ~120°C, the E’ data for the two samples are comparable and higher than the E’ data for the formulations cured for two days. The higher E’ data at 120°C for the six-day samples compared to the two-day data could be the result of the additional cure afforded by the additional four days at ambient temperature resulting in more crosslinked (i.e., higher rubbery E’) materials.

After 21 days of cure, the xylene-modified and neat polyamide/epoxy formulations display midpoint T_gs that are nearly the same (52–53°C). These T_gs are relatively high compared to the cure temperature (ambient temperature) and result in systems that are vitrifying. Vitrification causes the reaction rates to decrease substantially so that long times are required to advance the degree of cure. The presence of xylene affords a higher crosslink density and lower onset _��T_g. The E’ data at 120°C are comparable and higher than the corresponding data for the two- and six-day samples.

The two 120°C heat-cured show a dramatic difference in midpoint T_g (~33°C), with the xylene-based formulation exhibiting the lower T_g (64°C). As was seen with the ambient-cured formulations prepared with xylene, the difference in the onset and midpoint T_gs for the heat-cured material made with xylene is greater than that observed for the neat material. Both heat-cured samples display cure ratios close to 1, suggesting that there is little additional cure occurring in the rheometer.

Thermal Data

The 1st and 2nd heating scan DSC data for the polyamide/epoxy formulation (prepared without xylene) immediately after mixing are shown in Figures 6 and 7, respectively. A summary of the key thermal data from the 1st and 2nd heating scans for this and the other 100% solids samples is provided in Table 2. As with the data in Figure 6, and the subsequent 1st scan DSC data, differences in how the baselines are drawn impact the numerical values of the resulting tabulated data, especially for the cure exotherm. As a result, care must be taken not to over-interpret relatively small differences in thermal data.

The data in Figure 6 show a T_g in the –20°C range (attributable to the two reactants) and an exotherm associated with cure over the 50 to 200°C range. Since the data were obtained immediately after mixing, the ~231 J/g exotherm reflects the total heat associated with cure. A single T_g of ~99°C and no residual cure exotherm are observed in Figure 7; since this is second scan data, this T_g is closer to the maximum T_g attainable for this system and is in good agreement with the T_g observed from the DMA experiments for the 100% solids heat-cured formulation.

The thermal data for the polyamide/epoxy sample cured two days at ambient temperature and humidity are shown in Figures 8 and 9. The 1st scan data indicate an inflection T_g of ~38°C and a residual cure exotherm of ~45 J/g. Based on the cure exotherm data in Figures 6 and 8, the degree of cure for this sample is ~81%. An apparent relaxation endotherm is also witnessed at the end of the glass transition region. Similar to the sample analyzed immediately after mixing (shown in Figure 7), the 2nd scan data in Figure 9 indicate an inflection T_g of 99°C. The fact that the data in Figures 7 and 9 are in agreement indicates that both materials result in comparable networks (at least with regard to T_g) and that the maximum attainable T_g does not appear to be dependent on the cure path, i.e., a ramp cure in the DSC vs two days at ambient temperature followed by a ramp cure in the DSC.

The DSC data after six days of cure at ambient temperature and humidity for the 100% solids sample are shown in Figures 10 and 11. Similar to the data in Figure 8, the 1st scan data exhibit a T_g, stress relaxation endotherm, and a residual cure exotherm. Relative to the two-day data, the T_g has increased, while the cure exotherm is comparable. The 2nd scan data indicate an inflection T_g of 99°C, the same as the data immediately after mixing and after two days of ambient cure.

The 1st and 2nd scan DSC data for the formulation cured 21 days at ambient temperature and humidity are shown in Figures 12 and 13, respectively. Relative to the data in Figures 6, 8, and 10, the 1st scan T_g continues to increase, while the residual cure exotherm is about the same as two- and six-day data. The increases in the 1st scan T_g data with increasing cure time are in agreement with the DMA data in Table 1. The lack of change in the cure exotherm, coupled with the increase in the minimum E’ (summarized in Table 1), suggests that DSC does not appear to be as sensitive to the latter stages of cure compared to DMA.

The DSC data for the 120°C heat-cured polyamide/epoxy formulation are shown in Figures 14 and 15. The 1st scan data indicate a small T_g around 48°C and a primary T_g of about 91°C. No residual cure exotherm is observed, which is consistent with the sample being fully cured prior to DSC analysis. The 2nd scan T_g of 98°C is consistent with the prior 2nd scan DSC results. The endotherm at 155°C is likely an experimental artifact.

Inflection T_gs of 47 and 65°C are observed for the 120°C heat-cured polyamide/epoxy/20% xylene formulation; these data are shown in Figure 16. A residual cure exotherm is not evident. The impact of any residual xylene on the DSC data is not obvious. The 2nd scan DSC data for this material are shown in Figure 17 and indicate an inflection T_g of 67°C.

The 2nd scan T_gs for the two heat cured samples (98°C for the 100% solids formulation, 67°C for the xylene-based formulation) are in very good agreement with the corresponding DMA data (97°C, 64°C). The relatively low ultimate T_g for the heat-cured sample prepared with xylene is either due to the plasticization of the crosslinked network by residual xylene or the presence of the xylene has somehow hindered or altered network formation.

Moisture-cured RTV Silicone Formulation

Figure 18 displays the tensile dynamic mechanical properties as a function of frequency for the moisture-cured RTV silicone after one hour of cure at ambient temperature and 30% RH. The rheological data are frequency-dependent, which is typical for a viscoelastic material. In particular, E’ increases and tan(d) decreases with increasing frequency. The decrease in tan(d) indicates more solid-like (rather than liquid-like) behavior with increasing frequency. These rheological data are typical of those collected over the seven-day period.

Figure 19 displays the E’ data as a function of frequency after 1 hr, 24 hr, and 7 days for both replicate specimens. The increase in E’ with increasing cure time is indicative of additional cure. As might be expected, the largest increase in E’ occurs early in the cure. In addition, the frequency dependence of E’ decreases with increasing cure time indicating that a more elastic material is created with further cure. Figure 20 presents the corresponding tan(d) data over the seven-day cure period. The decrease in tan(d) with increasing cure time corroborates the E’ data shown in Figure 20. These kinds of data provide an objective measure of the time required to achieve different degrees of cure, as well as insight into how the material properties change with time. Similar studies could be conducted at other environmental conditions where it would be expected that experiments conducted at higher RHs would result in a faster cure of the moisture-cured RTV material.

Polyetheramine/Epoxy Blends

Figure 21 is an overlay of the E’ and tan(d) data for the three cured PEA Low/PEA High/epoxy formulations. The increase in T_g and crosslink density (as evidenced by the E’ data in the rubbery region) with increasing PEA Low level is evident. This is not surprising considering that PEA Low has a smaller molecular weight (and lower equivalent weight) and yields a tighter network compared to PEA High-based formulations.

The E’ data for the 60%/40%, 70%/30%, and 80%/20% PEA Low/PEA High blends cured with epoxy resin as a function of immersion time are shown in Figure 22. The magnitude of E’ prior to the introduction of the water follows the expected trend, with the 80%/20% blend exhibiting the highest modulus at ambient temperature. The immersion data indicate a decrease in E’ with increasing immersion time that is due to the uptake of water into the crosslinked networks. This decrease in E’ is more pronounced as the level of PEA High in the formulation increases. One explanation for this is that by increasing the average molecular weight between crosslinks of the network, it is easier for water to diffuse into the polymer network, causing the modulus to decrease with increasing immersion time because of plasticization.

While these experiments were performed with water at ambient temperature, organic solvents, as well as temperatures other than ambient could be used to understand the solvent resistance of coatings. Immersion experiments performed at a higher temperature or with thinner films would likely yield comparable results on a shorter time scale. These immersion experiments are particularly useful as they collect data in real time and provide kinetic-type information and alleviate the issues associated with traditional solvent resistance experiments, namely the need to remove and handle materials from the immersion medium and load them into a rheometer without excessive deformation. Of course, if the time required for a coating to experience the impact of a solvent is particularly long or requires exposure conditions that cannot be accommodated by the immersion fixture and rheometer, the more traditional approach of immersing and removing specimens for testing needs to be employed.

Summary

The rheological and thermal properties of a polyamide-cured epoxy coating (prepared with and without xylene) and a moisture-cured RTV silicone coating were characterized during cure using DMA and DSC. For both coatings, substrates that enabled the analysis of free films were used. DSC measurements were limited to materials that had relatively little solvent, as the endotherm associated with the evolution of relatively large amounts of solvent dominated the thermal curves, making it impossible to determine _��T_gs and residual cure exotherms.

Both the polyamide/epoxy (100% solids) and polyamide/epoxy/20% xylene formulations exhibit increases in T_g with increasing cure time. For comparable cure times at ambient temperature and humidity, the presence of solvent results in lower onset _��T_gs and lower Mcs. Dynamic mechanical and DSC data on 120°C heat-cured samples of the xylene-modified and 100% solids formulations indicate that the presence of xylene results in a significantly lower T_g data (~97 vs 64°C); part of this may be due to the difficulty of xylene to diffuse out of the relatively thick coating that was used.

For the silicone-based coating formulation, the DMA data show a steady increase in E’ over the course of a seven-day cure at ambient temperature and 30% RH, coupled with an increase in the elastic character of the material with increasing cure time.

Two polyetheramine curatives, one with a relatively low molecular weight (PEA Low) and one with a relatively high molecular weight (PEA High), were used to prepare three epoxy-based coatings of varying crosslink density. For each formulation, decreases in E’ as a result of immersion in water were measured in situ as a function of time using an immersion fixture. Of the three networks, the one with the largest Mc (60% PEA Low/40% PEA High/Epoxy) exhibits the largest decrease in E’; this is likely due to the relative ease with which water can penetrate and plasticize the matrix. This approach to evaluating the solvent resistance of coatings avoids the handling issues associated with traditional immersion studies that rely on periodically removing and testing a sample that has been immersed.

By measuring rheological and thermal properties as a function of cure time, it is possible to understand the rate with which properties develop and the transformation from the liquid to the solid state. It is particularly useful when these types of studies can be compared against the performance of a control formulation. The analytical approaches shown here can be extended to other coatings systems and can be complemented by techniques that suit the specific needs of the coatings technologist.

Acknowledgment

I wish to thank Intertek Allentown for supporting this endeavor. I would also like to acknowledge the support of Nausheen Habib, Scott Voth, Dr. Ling Yang, and Dr. John Zielinski for collecting the rheological and thermal data and for useful discussions.

References

1. Ates, M., J. Adhes. Sci. Technol., 30 (14), 1510-1536 (2016).
2. Kuhn, H.H., Child, H.D., and Kimbrell, W.C., Synth. Met., 71 (1-3), 2139-2142 (1995).
3. Smith, J.R. and Lamprou, D.A., Trans. IMF, 92 (1), 9-19 (2014).
4. Belder, E.G., Rutten, H.J.J., and Perera, D.Y., Prog. Org. Coat., 42 (3-4), 142-149 (2001).
5. Bley, O., Siepmann, J., and Bodmeier R., J. Pharm. Sci., 98, 651–664 (2009).
6. Koleske, J.V., “Mechanical Properties of Solid Coatings,” Encyclopedia of Analytical Chemistry, Meyers, R.A. (Ed.), Chichester: John Wiley and Sons,
   1-15, 2000.
7. Schlesing, W., Buhk, M., and Osterhold, M., Prog. Org. Coat., 49 (3), 197-208 (2004).
8. Malzbender, J., den Toonder, J.M.J., Balkenende, A.R., and de With, G., Mater. Sci. Eng., R, 36 (2-3), 47-103 (2002).
9. Mansfeld, F., Han, L.T., Lee, C.C., and Zhang, G., Acta, 43 (19-20), 2933-2945 (1998).
10. Mansfeld, F., J. Appl. Electrochem., 25 (3), 187-202 (1995).
11. Skrovanek, D.J. and Schoff, C.K., Prog. Org. Coat., 16 (2), 135-163 (1988).
12. Hegedus, C.R., Pepe, F.R., Dickenson, J.B., and Walker, F.H., J. Coat. Technol., 74 (927), 31-39 (2002).
13. Ramis, X, Cadenato, A., Morancho, J.M., and Salla, J. M., Polymer, 44 (7), 2067-2079 (2003).
14. Mafi, R., Mirabedni, S.M., Attar, M.M., and Moradian, S., Prog. Org. Coat., 54 (3), 164-169 (2005).
15. Mafi, R., Mirabedni, S.M., Naderi, R., Attar, M.M., Corros. Sci., 50 (12), 3280-3286 (2008).
16. Gupta, M. K., Thermochim. Acta, 166, 157-167 (1990).
17. Carlier, V., Sclavons, M., and Legras, R., Polymer, 42, 5327-5335 (2001).
18. Spinks, G. M., Liu, Z., Brown, H., Swain, M., See, H., and Evans, E., Prog. Org. Coat., 49 (2), 95-102 (2004).
19. Menard, K.P., Dynamic Mechanical Analysis: A Practical Introduction, Second Ed., Boca Raton: CRC Press. 2008.

CoatingsTech | Vol. 14, No. 7 | July 2017