Solving Adhesive Quality Issues: A Proposed Roadmap To The Answer
Michael Bradshaw, Chemist, Chemsultants International, Mentor, OH.
When dealing with complex problems it is best to take a step-by-step approach. This methodology can aid in the investigation of quality issues encountered with pressure sensitive adhesive (PSA) products. In the production of pressure sensitive products a number of materials are typically brought together: adhesive components are blended or compounded together, release liners are produced and finished prior to being silicone coated, face stocks may be extruded or top coated and primed, and finally all of these components are married in a lamination process. Finally, this construction of materials is finished or slit to size prior to packaging and shipping to a tape or label converter. This is quite an involved process for a simple product such as a label or tape to be born.
Troubleshooting adhesive failures can be overwhelming with the many variables that are directly
related to or can be contributing factors to a failure. We propose a "roadmap" to aid the investigator in narrowing the window of potential variables to the most likely few. At that point, it is up to the investigator to perform further analysis to exclude the non-contributing possibilities.
The first step in any problem-solving scenario is to verify that there is indeed a problem. This may seem odd at first, but one has to be careful in knowing exactly why a product is deemed "bad". Even with the advent of physical testing instruments, such as probe tack testers, adhesion/release testers, etc., it is not uncommon to see the "finger-stick" tack test or "hand - peel" adhesion test performed by end end-users. With such simplistic “tests” the detection range is so vast the tests have little merit (i.e. a sample with tack and a sample with no tack). We require actual numerical data to identify a significant difference. One person’s observation of a "problem" may be another person’s observation of “within specification" performance.
Common physical testing methods available to the adhesive investigator include, but are not limited to: peel adhesion, probe tack, loop tack, shear, and S.A.F.T. When possible, all tests should be setup to duplicate the actual use of the end product in the field in order to gather meaningful information. If actual use conditions can not be duplicated the use of a standard substrate is suggested. Conditions such as the proper peel angle, suitable substrate, and suitable aging condition(s) must be adhered to whenever possible. For example, no meaningful information is gathered via by aging a label under UV lamps (used to simulate outdoor sunlight) when the product is intended for indoor use. The same principle applies to peel angle. As an example, an end-user would be more likely to peel the release liner off a pest glue board at 180º as opposed to 90°, so testing should be performed at 180º. To arrive at potential root causes of failure, it is best to test a known "good" sample along side a "suspect" sample. This provides the investigator the additional help of benchmarked “good sample results and "suspect" sample results to have a basis for comparing the results. Once a statistically significant difference in physical properties is identified, the process of narrowing the variable window begins.
Once the common benchmarking tests have been completed the verification stage is complete. Samples were created and tested in a controlled laboratory setting thus eliminating application conditions as a source of deviation. The next step is to determine what may be causing the differences or potential performance problem. The most common and subsequently easiest easiest-to-overlook issues may be how much adhesive is present on the "good" versus "suspect" materials as well as the continuity of the adhesive coating. Most investigators will check the adhesive coat weight only. It is possible to have two samples with a similar adhesive coat weight but vastly different adhesive coverage or coating quality. If a sample were taken from a low coverage area, the physical properties would deviate from those of a nominal or higher coverage area. A quick and simple test to check for uniformity of adhesive coating is a caliper (or thickness) test (provided the face stock or backing caliper is uniform throughout the sample or at the least, measurable). If coating weight and / or uniformity are found to be the issue, only verification is needed. If not, further investigations are warranted. A look at a general problem verification / analysis roadmap might look as follows:
If physical benchmarking methods have substantiated a significant difference between the “good” and “suspect” samples, the next step would be to analytically characterize the "good" and "suspect" products. The path to studying the adhesive analytically can be seen as two variants followed in one of two ways: (1) surface chemistry and (2) bulk chemistry. Proper selection of analytical methods is critical in obtaining the correct information.
The primary job of an adhesive is to bond or adhere two substrates together with some minimally acceptable level of strength in use. The action of the adhesive adhering to a substrate occurs on both surfaces. The benchmark testing, already performed, compared the adhesion of the "good" and "suspect" products under similar conditions to similar substrates. If a difference had been seen, the surface of the substrate is most likely not the contributor.
The problem may lie on the surface of the adhesive. There are a wide variety of surface analytical testing methods to aid the investigator. The most common and most widely used is ATR-FTIR (an acronym for Attenuated Total Reflectance Fourier Transform InfraRed) spectroscopy. This technique probes the surface of a sample (up to 2 microns deep) with infrared energy of varying wavelengths. To understand the technique involves we must consider a little theory.
Molecules are constantly in motion above absolute zero (-273.2° C). These motions can vary in nature from (i.e., vibration, rotation, stretching, wagging, etc.) and each molecule has a distinct frequency of motion determined by the masses of the atoms involved as well as the bonding of those atoms. Motions that result in a non-zero dipole moment in a molecule are said to be IR-active. This means is that the motion can be excited with wavelengths of IR energy. It is by monitoring what energies (i.e. wavelengths) of IR energy are absorbed by a sample, that the investigator is able to determine what atoms are present and how they are bonded (functionally grouped together). An example IR spectrum is shown in Figure 1-1 below:
Figure 1-1: Sample IR Spectrum [Silicone Fluid]
The y-axis indicates the amount of incident IR energy that passes through or reflects off of the sample. The x-axis indicates the distinct frequencies of IR energy the sample is exposed to in wave numbers (i.e. inverse wavelength). A decreasing transmission occurring at a distinct wave number correlates to absorption of that corresponding energy. Correlation charts are referred to for determine functionality assignment.
IR is usually performed on a bulk basis by passing IR energies through a solvated material or a solid IR-transparent pellet composed of, most likely, potassium bromide. Several drawbacks exist with this analysis including: insoluble products and the extensive sample preparation involved in producing a solid pellet. With the advancement of technology, the ability to perform IR analysis on the surface of a sample has been made possible. Crystals having favorable refractive indices (i.e. zinc selenide, diamond, etc.) they can be employed to incident IR energy to multiple spots (ATR) or one spot (Single Reflectance ATR) on the surface of a sample. One of the conditions a sample must meet to use ATR-FTIR is that it must be flexible enough to make good contact with the crystal surface used. Pressure sensitive adhesive coated materials meet these criteria due to their natural ability to wet-out out a surface making the method readily useable.
Another requirement of using ATR-FTIR is that the thickness of the adhesive should not be less than 2 microns (< 0.1 mils) because that is the penetration depth of the IR energy on the surface of the sample. Otherwise the information obtained would be more characteristic of the adhesive/ face stock interaction surface or possibly the face stock alone. By its very nature, ATR is a powerful tool for identifying IR-active surface impurities, like silicone. Silicone is an excellent example because of the strong extinction coefficients related to the motions of the silicon-oxygen/silicon-carbon bonds. Silicones are known to appear in the 1260 (± 5), 800 (± 10) , and 1000 ñ 1200 cm-1 ranges.
Potential drawbacks of IR spectroscopy include detection limits, the additive features of the spectrum, and the inability to absolutely distinguish between certain chemicals. The detection limit of IR has been noted to be between 2 and 3% by weight of the total analyzed sample. Therefore, any IR- active component in a sample at less than 2% by weight has little chance of showing appreciable signals compared to the vastly more intense signals of the background (other components). Secondly, IR is known to be an additive technique, which in its simplest term means that the IR spectrum of a mixture composed of IR-active components: A, B, and C will appear like an overlay of the IR spectra of components A, B, and C (provided they do not chemically react with each other).
While this presents obvious advantages for a compounded pressure sensitive adhesive in determining what types of components are present, at the same time it can create problems for a spectroscopy scientist trying to differentiate what signals belong to what components. This indicates why IR is more widely used as a quality tool to compare "good" and "suspect" materials rather than as a tool to ascertain absolute identification of materials or components. There are instances where IR is not powerful enough alone to distinguish between two chemicals or components. For example, the spectra of the same polymer at different molecular weights can look the same. Also comparing n-heptane to n-nonane may prove difficult. More in in-depth analysis analytical techniques coupled with IR spectroscopy can help minimize these problems.
A second surface surface-analysis technique involves the use of X-ray energy to identify what elements are present. This technique is referred to as x-ray Photoelectron Spectroscopy (XPS). In this technique X-rays of known energy displace inner-shell electrons of certain atoms (excluding hydrogen and helium). The kinetic energies of the expelled electrons are measured and coupled with the knowledge of the incident x-ray energy, to determine the binding energy of the expelled electrons. This binding energy is used to identify what atoms are present on the surface. The resulting survey spectrum shows the elements present on the surface of a sample. XPS is limited to analysis depths of 20 to 50 angstroms due to the poor penetrating power of the electrons that are expelled.
If analytical comparison of "good" and "suspect" sample surfaces does not show a chemical difference, the next logical step is to analytically study the bulk adhesive. A troublesome aspect of this type of analysis is how the "good" and "suspect" samples are supplied. More than likely, the samples will be finished constructions with adhesive coated onto a carrier or face stock with potentially and laminated to a release liner. Therefore to analyze the bulk adhesive, the carrier or face stock needs to be removed from the adhesive. Removal is commonly performed in one of two ways: (1) solvating of the adhesive off of the carrier or face stock or (2) microtoming of the adhesive from the face stock. The preferred method would be microtoming of the adhesive since there may be coatings on the face side of the adhesive-coated material that could also solvate off leading to false identifications.
The starting point in any bulk analysis is to determine roughly how many components are present in the finished material. One of the quickest analytical methods to accomplish this is with a thermal analysis technique known as Thermal Gravimetric Analysis (TGA). A sample of known mass is heated at a specified rate in an inert atmosphere (nitrogen) from room temperature up to a maximum of 1000° C. The mass of the sample is continually monitored as the temperature is raised. Under these conditions, decompositions of individual components in the sample occur as evidenced by decreasing mass at certain temperatures due to thermal decomposition of the individual components in the sample. This technique so powerful because it reveals both qualitative (how many decomposable components are present) as well as quantitative information on the sample. Knowing the initial sample mass and the mass of the sample after a component degrades the investigator can get a relative weight percentage of that component. Therefore, a simple analysis of both "good" and "suspect" samples can reveal relative compositional differences as long as the experimental conditions are similar. A sample TGA plot is shown in Figure 1-2 below:
Figure 3: Sample TGA Thermogram
The plot shows the change in weight (y-axis) with respect to increase in temperature (x-axis). The obvious drawback of this technique is that it is destructive. Once the sample is analyzed, the sample has been irreversibly degraded. However, the sample mass needed to perform the test is in the tens of milligrams and can usually easily be spared.
When a relative idea of how many components are present in the sample has been determined, we can devise a methodology for separation of components using various solvents or with a chromatographic method. Once the components are solvated, techniques such as solvated FTIR can be performed. A more involved and powerful technique used to analyze solvated materials, called Fourier Transform Nuclear Magnetic Resonance (FTNMR), is a more detailed step to distinguish between chemicals with similar IR spectra. FTNMR gives information on the nuclear level for a select few atoms. The two most common nuclei studied are 1H (hydrogen) and 13C (carbon-13). The power of FTNMR lies in the ability to distinguish between hydrogen or carbon nuclei bonded to different chemical types of atoms. Therefore, structural information can be obtained on a sample. Like FTIR, FTNMR is an additive technique leading to an increased ability to ascertain differences. However, the complexity of the interpretations increases exponentially with the number of components in a sample.
FTNMR can be performed on only two types of samples: (1) solvated or liquid products or (2) solid powders. Thus FTNMR analysis of formulated and compounded pressure sensitive adhesives is only possible if the adhesive can be solvated. Choice of solvent is critical in so that it does not cause spectral interference with the sample dissolved. In most cases deuterated (2H) solvents are used in place of hydrocarbon based solvents. These solvents can be costly based on the complexity of the structure. FTNMR involves placing a sample into a magnetic field and monitoring the absorption of radio frequencies. 1H and 13C nuclei, when exposed to a magnetic field, possess split nuclear spin states. The splitting between those spin states depend on what chemical environment the nuclei are in (i.e. surrounded by electrons, other similar nuclei, etc.) and the magnitude of the external magnetic field. The position of the signal is normalized to a ppm value in order to factor out the effects of different magnetic field strength instruments leading to a direct correlation of signal position and chemical environment no matter what magnetic strength instrument is employed. Again, comparison between "good" and "suspect" solvated adhesives can show possible differences in the samples.
In addition to the above, other, more specialized analytical techniques are available to the investigator. These techniques offer a good foundation to build from to compare "good" and "suspect" samples. The correlation of the vast amount of analytical data to the physical benchmarking data can offer a complete picture on identification of a problem and possible sources of resolution. What follows is a step-by-step study of an adhesive issue making use of the "roadmap" discussed herein:
Case Study #1
End-user E has purchased pressure sensitive adhesive label stock from Vendor V over a 5-year period. Vendor V obtains the label stock from Manufacturer M. End-user E has seen tighter release characteristics and label failures with their last lot of label stock purchased.
Step #1: Problem Verification
The End-user is asked to submit samples of known "good" lot(s) of label stock along side the "suspect" lot(s). The label stocks are multi-purpose and therefore choice of a substrate is arbitrary. Stainless steel was chosen as a standard surface for testing. Release and subsequent adhesion (Test Method PSTC-4B) as well as probe tack (Test Method ASTM D 2979) were evaluated. The following results were obtained:
Table 1a: Physical Comparative Testing of "Good" and "Suspect" Labels
|
|
"Good" |
"Suspect" |
|
Release (PSTC-4B): aged 24 hr @ 70 °C / 0.25 psi: g/in |
||
|
n |
3 |
3 |
|
AVG |
3.5 (± 0.7) |
670.4 (± 98.0) |
|
MOF |
A |
A |
|
180 ° Peel Adhesion (PSTC-4B): lbs/in |
|
|
|
n |
3 |
3 |
|
AVG |
4.26 (± 0.09) |
0.61 (± 0.04) |
|
MOF |
A |
A |
|
Probe Tack (ASTM D 2979): g/cm2 |
|
|
|
n |
5 |
5 |
|
AVG |
140 (± 8) |
65 (± 7) |
|
MOF |
A |
A |
The data clearly indicates a discrepancy between the two lots of labels. Release is over 100 times greater in the case of the "suspect" material. Peel adhesion to stainless steel after aging of the label stocks for 24 hours at 70º C under pressure shows an 85% decrease comparing the "good" to the "suspect". Probe tack is 54% lower in the case of the "suspect" label stock compared to the "good" sample. Therefore physical testing has verified that a problem exists. However, there is not enough information from the data gathered to identify where the problem lies.
Step #2: Physical Characterization of Problem
Now that a statistically significant difference can be seen between "good" and "suspect" label stocks in a controlled laboratory environment, possible sources of the discrepancy can be investigated. Adhesive coat weight and caliper are then determined for the two labels:
Table 1b: Physical Comparative Characterization of "Good" and "Suspect" Labels
|
|
"Good" |
"Suspect" |
|
Adhesive Coat Weight (TLMI Coat Weight): g/m2 |
||
|
AVG |
16 |
16 |
|
Caliper: mils |
|
|
|
AVG |
4.76 (± 0.04) |
4.76 (± 0.07) |
As can be seen from the data gathered, there was no statistically significant difference between the adhesive coating weight or adhesive caliper between the two label stocks.
Step # 3: Analytical Characterization of Problem
The "good" and "suspect" labels were shown to be different by physical testing and the difference was not seen to be due to adhesive coating weight and/or caliper. The first area of investigation for analytical work should be the adhesive surface. ATR-FTIR spectroscopy was performed on multiple spots of each label adhesive surface. The individual stacked spectra of each spot analyzed are shown in Figures 1-3 and 1-4:
Figure 1-3: Stacked Overlay ATR-FTIR Spectra of Multiple Spots On Surface of "Good" Adhesive
The spectra in Figure 1-3 appear very similar to each other.
Figure 1-4: Stacked Overlay ATR-FTIR Spectra of Multiple Spots On Surface of "Suspect" Adhesive
The spectra in Figure 1-4 also appear similar to each other.
For comparative reasons, each group of 4 spectra was averaged and stacked in Figure 1-5:
Figure 1-5: Stacked Overlay Average ATR-FTIR Spectra of "Good" and "Suspect" Adhesive Surfaces
Upon initial investigation, the spectra appear similar. A closer look at the spectra below 1700 cm-1 reveals differences as shown in Figure 1-6:
Figure 1-6: Zoomed In Region of "Good" and "Suspect" Adhesive ATR-FTIR Spectra Overlay
These differences are not readily seen unless a direct subtraction of the average ATR-FTIR spectrum of the adhesive surface from the "good" label is made from the spectrum of the adhesive surface of the "suspect" label:
Figure 1-7: ATR-FTIR Difference Spectrum Of "Good" Label Adhesive From "Suspect" Label Adhesive
The boxed regions are indicative of possible silicone presence.
Step #3-1: Verification
ATR-FTIR testing of the surfaces of the "good" and "suspect" labels showed possibly more silicone present on the "suspect" label adhesive surface. XPS was used as a check for detection of quantitative differences in percentage of silicon atoms in the regions analyzed. Figures 1-8 and 1-9 represent the XPS elemental survey scans of a region of the adhesive surface of "good" and "suspect" labels respectively:
Figure 1-8: XPS Elemental Survey Scan of "Good" Label Adhesive Surface
Oxygen, carbon, and silicon atoms were detected on the surface of the "good" label adhesive denoted as O1s, C1s, Si2s, and Si2p respectively.
Figure 1-9: XPS Elemental Survey Scan of "Suspect" Label Adhesive Surface
The same three elements were detected on the surface of the "suspect" label adhesive denoted by the similar signals. To obtain an average surface percent atom composition, three spots were analyzed on the surface of the "good" label adhesive:
Table 1c: Percent Atom Composition of Adhesive Surface On "Good" and "Suspect" Labels
|
Atomic Percent |
|||
|
|
Carbon |
Oxygen |
Silicon |
|
"Good" Label Adhesive |
77.0 (± 1.0) |
21.3 (± 0.4) |
1.8 (± 0.7) |
|
"Suspect" Label Adhesive |
67.4 |
23.5 |
9.1 |
To confirm the result, fuchsine dye was used to wipe the surface of the liner after label removal on both "good" and "suspect" samples. Areas on the "suspect" label liner indicated removal of silicone by absorption of the dye into the base liner paper.
Conclusion
"Good" and "suspect" labels showed substantial differences in physical properties. Physical characterization of the two products did not reveal a significant difference. Surface analytical testing showed possible presence of silicone on the adhesive surface through ATR-FTIR. XPS percent atomic survey showed a substantially higher level of silicon atoms on the surface of the "suspect" adhesive. Dye stain/anchorage testing of the liners showed that silicone was poorly anchored on the release liner to which the "suspect" labels had been laminated. Therefore, the silicone presence detected is hypothesized to be due to silicone transfer from the liner to the adhesive surface in the "suspect" lot. Thus explaining the diminished adhesive properties in the "suspect" material.
Case Study #2
Manufacturer Y purchases a water-based, high shear pressure sensitive label adhesive from Company D to use in production of permanent labels. In mid-production, a switch in the lots of adhesive used was made. Sporadic failure complaints came in for poor adhesion and flagging of the labels.
Step #1: Problem Verification
Manufacturer Y submits failing and non-failing label products for testing. Stainless steel was chosen as a standard surface for testing. Peel adhesion (Test Method PSTC-101F) as well as probe tack properties (Test Method ASTM D 2979) were evaluated:
Table 2-1: Peel Adhesion and Probe Tack Data For Failing and Non-Failing Labels
|
|
||
|
|
Non-Failing Labels |
Failing Labels |
|
Peel Adhesion (PSTC-101F):lbs/in |
||
|
Number |
3 |
3 |
|
AVG |
2.74 (±0.01) |
1.16 (± 0.04) |
|
MOF |
FD |
A |
|
Probe Tack (ASTM D 2979): g/cm2 |
|
|
|
Number |
5 |
5 |
|
AVG |
178 (±21) |
66 (±7) |
|
MOF |
A |
A |
The data clearly shows a discrepancy between the two label products with face stock stock-tearing bonds in one case and clean clean-release peel in the other. There is also an approximate 3-fold difference in tack between the failing and non-failing products.
Step #2: Physical Characterization of Problem
Now that a statistically significant difference can be seen between the label stocks in a controlled laboratory environment, possible sources of the discrepancy can be investigated. Adhesive coat weight and caliper are determined for the two label stocks:
Table 2-2: Physical Comparative Characterization of "Good" and "Suspect" Labels
|
Non-Failing Failing Labels Labels |
||
|
Total Caliper (ASTM D 3652 modified): mils |
||
|
Number |
7 |
5 |
|
AVG |
5.45 (±0.07) |
5.39 (± 0.04) |
|
Adhesive Coat weight (TLMI Coat Weight Test - modified):g/m2 |
||
|
AVG |
22.0 |
21.6 |
There was no statistically significant difference in the data obtained.
Step # 3: Analytical Characterization of Problem
The failing and non-failing labels were shown to be different by physical testing and the difference was not seen to be due to adhesive coating weight and/or caliper. To investigate surface differences, ATR-FTIR spectroscopy was performed on multiple spots of each label adhesive surface.
The individual stacked spectra are shown in Figures 2-1 and 2-1:
Figure 2-1: Stacked ATR-FTIR Spectra of Non-Failing Label Adhesive Surface
There appears to be no significant differences between the three regions other than the signals at ~ 2300 cm-1 attributed to carbon dioxide presence from the atmosphere.
Figure 2-2: Stacked ATR-FTIR Spectra of Failing Label Adhesive Surface
Again, the spectra appear similar to each other. To directly compare the two samples, the average spectra were stacked together and presented in Figure 2-3:
Figure 2-3: Stacked Average ATR-FTIR Spectra of Failing and Non-Failing Label Adhesives
Some signal intensity differences can be seen between the two spectra that could be due to either contact pressure of the sample with the ATR crystal or concentration differences in the formulation. Signal location and structure seems to be similar between the two samples.
Step #4: Bulk Analysis
To help ascertain whether the intensity differences seen are due to quantitative formulation differences or the analysis procedure, TGA was performed on the two adhesives after removal from the face stock. The thermo grams are presented in Figures 2-4 and 2-5 respectively:
Figure 2-4: TGA Thermo gram of Non-Failing Label Adhesive Sample
The adhesive appears to have two major decomposable components. Relative quantification of the components leads to a ratio of approximately 70:30. The corresponding thermo gram for the failing adhesive follows:
Figure 2-5: TGA Thermo gram of Failing Label Adhesive Sample
There still appears to be two components with relatively similar decomposition rates and temperatures. However, the ratio between the two components seems to be shifted from approximately 70:30 to approximately 80:20. Therefore, the signal intensity differences seen with ATR-FTIR were indicative of a formulation difference between the two adhesives. Further analysis would be needed to determine the identity of the two components.
Conclusion
Physical comparative testing confirmed differences in the failing and non-failing label stocks in peel adhesion and tack properties. Caliper and adhesive coating weight were not seen to be statistically different between the two labels. ATR-FTIR showed similar signal location and structures with slight differences in some signal intensities. Possible causes were rationalized to be experimental error or formulation differences. TGA bulk adhesive analysis was performed on the failing and non-failing samples with differences seen in the ratio of two major components. Therefore the root cause of the problem is most likely related to the adhesive formulation differences between the two labels.
These two Case Study examples presented two of many quality issues that can likely be resolved by use of the troubleshooting "roadmap" presented. A thorough investigation of quality issues not only determines what the differences are between "good" and "bad" product but also identifies potential root causes. By applying this "roadmap" design, one can accomplish both goals.
Coupled with common physical tests, certain analytical techniques can be run to form such a “roadmap” to the root cause identification of quality issues. Two of the many possible techniques that can be used were presented here: Thermo-gravimetric analysis (TGA) and Infrared spectroscopy (IR). These techniques are good starting points for analytical testing.
The two techniques offer different types of information about pressure sensitive products and materials. There are additional advanced analytical techniques available to study pressure sensitive adhesives and related coatings. However, these two basic techniques offer a good starting point for coupling analytical results to physical data. It can be seen how information relating to the structural makeup or physical proportions of adhesive components can assist in determining reasons for lack of product performance or product quality. While routine physical tests like peel, tack and shear can identify that you have a problem; it is analytical testing that can help identify the root cause of the problem.
Chemsultants International, Inc. 9079 Tyler Blvd. Mentor, OH 44060 Phone: 440.974.3080 Fax: 440.974.3081About Us Contact Us Site Map Career Opportunities Locations
