Cure Checking Simplified

Differential Scanning Calorimetry as a tool for powder coating quality control

Powder coatings represent a well-established technology for coating applications in many different industries. However, overproduction of standard powder coatings and more and more demanding product requirements are leading to increased efforts in new material development. In particular, a rising demand of formulations for the coating of polymeric substrates requires investment in material research as well as process development and optimization.

As the processing energy footprint becomes increasingly important, a further reduction of process energy consumption becomes necessary. The key to optimizing the use of established coating systems for industrial applications is therefore a thorough understanding of the main factors for controlling the production process. For thermosetting powder coatings in particular, a comprehensive characterization of the thermal influences is necessary.

Differential Scanning Calorimetry (DSC) offers a standardized analytical technique that is already well-established for paints and coating systems [1, 2, 3]. It is used during compound or resin development, for process-oriented material characterization and for quality control during production.

DSC technology simplified for the user

With the NETZSCH DSC 214 Polyma, a new level of DSC analysis has been reached (Figure 1): a holistic measurement concept was developed covering all necessary steps from sample preparation, programming and conduction of the DSC measurement to final evaluation of the resulting data.

For the first time in DSC analysis, AutoEvaluation permits autonomous evaluation of the thermal transitions. This can be done by just one click, still providing freedom for the user to adjust and improve the evaluated data. The algorithm evaluates in accordance to ASTM and ISO standards, starting with the main effects and subsequently going down to side effects. In contrast to conventional software, AutoEvaluation requires no operator with expert knowledge for data interpretation and creates a user -independent, sophisticated second opinion.

Identify is a software module that compares the thermal properties of a polymeric sample from a DSC-curve (glass transition, endo- and exothermal peaks etc.) with a measurement library [4, 5, 6].

It offers two main advantages, especially for quality control tasks: first, the Identify database recognizes an unknown polymer sample by its thermal characteristics and saves time-consuming literature research; second, users can extend the integrated NETZSCH database with his own DSC results. This software gives a unique opportunity for quality control by comparing polymers objectively, at delivery checks, after storage and drying or at the final product stage.

Figure 1: Typical measurement configuration for the NETZSCH DSC 214 Polyma (left) including an automatic sample changer (ASC), a sample preparation set and the SmartMode user interface (right).Figure 1: Typical measurement configuration for the NETZSCH DSC 214 Polyma (left) including an automatic sample changer (ASC), a sample preparation set and the SmartMode user interface (right).

Results at a glance

⇒ Five different commercial thermosetting powder coating samples were used to demonstrate the capabilities of the DSC 214 Polyma, a Differential Scanning Calorimetry (DSC) system for comprehensive material characterization as used in quality control.

⇒ The thermal transitions of these powder coatings were characterized. All showed a slightly different melting and curing behaviour. Relevant process parameters such as melting temperature, as well as application parameters such as glass transition temperatures, could be derived from the results.

⇒ A measurement library for quality control can readily be created using the NETZSCH Identify software module. This can easily assign a curve to the known database of powder coating types, making it a useful tool for automated quality control. The entire system requires minimal skill and experience on the part of operators for quality control functions.

Experimental procedure outlined

The DSC samples investigated here were prepared from five different, commercially available types of powder coating. They were arbitrarily named A10, B12, C8, C13 and E4, and represent standard powder coatings as used for example in the white goods industry.

For the experiments presented here, ca. 8.5 mg of powder sample were evenly distributed in an aluminum Concavus® DSC pan, and sealed with a pierced lid allowing gas exchange with the surrounding atmosphere (see Table 1).

For the DSC investigation, the NETZSCH DSC 214 Polyma was used with the Intracooler IC70 with a maximum operation range from -70°C to 600°C (see again Figure 1). The measurement method was set according to Table 2 with a purge gas flow rate of 40 ml/min nitrogen. The low thermal inertia furnace on this equipment allows high heating and cooling rates to be obtained. This can accelerate the acquisition of DSC results compared to conventional heat-flow DSC systems. Here, this advantage was used for the cooling segments with a cooling rate of 50 K/min from 300°C to 0°C. Therefore, the total measurement time for three consecutive heating at a rate of 20 K/min could be reduced by almost half to less than 45 minutes for each test.

The DSC was operated in SmartMode. This includes an optimized user interface, which requires a minimum of data input to start a measurement. In addition, the implementation of user-defined methods makes the setup of the automatic sample changer (ASC) very convenient and fast.

The Tau-R correction was applied to the analysis of the results in order to minimize artificial broadening of the measured DSC transitions caused by thermal inertia of sample and reference pan.

Table 1: Sample weigth documentation

 

SamplePan typeWeight (mg)
A10Al (Concavus®) + pierced lid8.58
B12Al (Concavus®) + pierced lid8.47
C8Al (Concavus®) + pierced lid8.40
C13Al (Concavus®) + pierced lid8.66
E4Al (Concavus®) + pierced lid8.48
"unknown"Al (Concavus®) + pierced lid8.50

Table 2: Temperature profile used for the DSC 214 Polyma experiments

 

SectionStart-Temp (°C)End-Temp (°C)Heating rate (K/min)Duration (min)
1. Cooling250500.5
1. Heating0100205
2. Cooling1000502
2. Heating03002015
3. Cooling3000506
3. Heating03002015
Total Duration43.5

Three heating runs characterize performance

Figure 2 shows the first, second and third heating of sample C8, performed according to the temperature program given in Table 2. In the first heating from 0°C to 100°C, an endothermal transition with a peak temperature at 70°C and an enthalpy of 9.19 J/g is found. This corresponds to a relaxation of the powder coating. As it is a non-reversible transition for thermosetting polymer powders, it is not to be mistaken with a melting transition.

The second heating from 0°C to 300°C was chosen according to the corresponding curing step which is applied in the coating production process. Here, a maximum temperature of 300°C can occur and the thermal behavior of the coating can be observed directly in the DSC. At 58°C in the second heating the glass transition of the uncured coating is observed. Further on at 123°C, a small melting transition is detected and the exothermal curing reaction of the sample is observed with a onset temperature at 173°C and a peak at 217°C. Around the maximum temperature of 300°C, no deviation of the DSC baseline and therefore no decomposition processes are found.

The third heating is again performed from 0°C to 300°C in order to detect irreversible changes as compared to the second heating curve and to confirm full cure of the sample. The glass transition with a midpoint temperature at 76°C is now significantly shifted to higher temperatures with respect to the second heating. This is due to the curing reaction, as greater crosslinking within the coating material leads to a higher glass transition temperature. At 123°C, a reversible melting peak as already seen in the second heating can be detected. This can be due to melting of an additive in the powder coating formulation, e.g. a wax component used to improve impact and processing properties of the sample.

Figure 2: DSC measurement of sample C8; depicted are the first heating (continuous line), second heating (dashed) and third heating (dotted); for better visibility, the y-axis of the respective curves is shifted by an arbitrary DSC offsetFigure 2: DSC measurement of sample C8; depicted are the first heating (continuous line), second heating (dashed) and third heating (dotted); for better visibility, the y-axis of the respective curves is shifted by an arbitrary DSC offset
Figure 3: First heating curves for all measured coating types A10, B12, C8, C13 and E4; the curves were evaluated using the AutoEvaluation software, displaying peak temperature and peak area; again the curves are arbitrarily stacked for better viFigure 3: First heating curves for all measured coating types A10, B12, C8, C13 and E4; the curves were evaluated using the AutoEvaluation software, displaying peak temperature and peak area; again the curves are arbitrarily stacked for better visibility.

First heating shows melt characteristics

Figure 3 shows a comparison of the first heating for the five different coating types investigated in this study. All curves were evaluated using the AutoEvaluation function of the analysis software. As already seen in the DSC curve for sample C8, all coatings show a pronounced relaxation peak with peak temperature ascending from B12 to C8. The evaluated enthalpies differ within a standard deviation of ca. 7.5% which is well above the uncertainty of the weighed sample masses (Table 1).Therefore, small but significant differences can be determined for the endothermal relaxation enthalpy of the measured samples. All results are summarized in Table 3.

As an additional finding, the relaxation transition for the thermoset coatings is finished before the heating segment end for all types. Therefore a temperature of 100°C is sufficient for a first step in the coating process in order to achieve uniform relaxation of the powder materials.

Table 3: Summary of the DSC results for all measured samples


1st heating

2nd heating

3rd heating
Sample nameRelaxation Peak (°C)Relaxation Enthalpy (J/g)Glass Transition (°C)Curing Peak (°C)Curing Enthalpy (J/g)Glass Transition (°C)
A10638.995219621.472
B12597.9651n.a.n.a.65
C8709.195821540.376
C13689.515721735.173
E46810.157n.a.n.a.65
"unknown"709.12n.a.n.a.n.a.n.a.

Second and third heating compares curing performance

In Figure 4 the DSC results for the second and third heating of all measured coating samples in a temperature range from 0°C to 300°C are summarized, again using the AutoEvaluation function. After powder melting in the first heating, a glass transition can be found for all samples between 51°C and 58°C. The order of the glass transition temperatures of the different coatings corresponds to the order of the melting transitions depicted in the first heating curves (see Figure 3).

Starting at temperatures from ca.110°C to 130°C, an exothermal curing reaction is observed for samples A10, C8 and C13. The curing of both samples C8 and C13 is very similar whereas the curing peak is shifted to higher temperatures by ca. 20K. No significant exothermal curing enthalpy is found for samples B12 and E4. It is assumed that this is due to the generally low curing rate of the curing reaction for the powder coatings investigated here. This is corroborated by the fact that all samples display a very broad curing peak distributed over a temperature range of more than 100K. Additionally, weak endothermal effects due to evaporation of the products released during polycondensation cannot be excluded. These could in turn cancel out with the exothermal curing enthalpy and result in a flat DSC signal.

Figure 4: Second (dashed, top) and third heating curves (dotted, bottom) for all measurerd powder coatings; the curves were are stacked with an arbitrary DSC offset for better visibility.Figure 4: Second (dashed, top) and third heating curves (dotted, bottom) for all measurerd powder coatings; the curves were are stacked with an arbitrary DSC offset for better visibility.

In the third heating (Figure 4), the glass transition can be found for all tested coating samples in a range from 65°C to 76°C, respectively. It can be observed that all glass transition temperatures shift to higher temperatures with respect to the second heating, with a maximum shift of 20K for sample A10. This can be taken as a clear sign of a crosslinking reaction in all samples, despite the fact that samples B12 and E4 display no significant curing peak. Furthermore, as no additional exothermal curing peaks were found in the third heating, full curing for all samples after previous heating to 300°C can be assumed.

As already indicated for sample C8, no significant sign of decomposition can be found for either both second or third heating after thermal treatment up to 300°C for all coatings.

Evaluation of unknown sample further tests DSC system

In a next step, a sample from the powder coating type C8 (“unknown”) was chosen to demonstrate the capabilities of the used DSC software for quality control: To demonstrate the Identify functionality, a library (“DSC_powder coatings”) was established from the first heating curves as seen in Figure 2 of all measured powder coatings. For that the curves only have to be marked and are loaded into the newly created library with one mouse click.

Figure 5: The DSC curve for sample “unknown” is correctly assigned to the powder coating type “sample C8_heat1” (left column), which is detected in the “DSC powder coatings”-library (right column), previously constructed from the first heatingFigure 5: The DSC curve for sample “unknown” is correctly assigned to the powder coating type “sample C8_heat1” (left column), which is detected in the “DSC powder coatings”-library (right column), previously constructed from the first heatings of all samples measured in this study.

In Figure 5, the first heating curve of the sample “unknown” is shown, measured under the same conditions as described above (Table 2). For the analysis, the newly constructed database is selected from the list of available libraries. The assignment of the sample is shown in the left column with a top entry of “sample C8_heat1”. From the curve comparison, a similarity value of 99.59% with respect to the test sample “unknown” is given, clearly indicating it a sample of coating type C8. Other entries are listed below with descending order of similarity. The similarity value is calculated according to the overall curve shape of the curve taking all evaluated parameters into account [4]. In that way the underlying Identify algorithm allows the comparison of a measurement curve with a few hundred database entries in a few seconds.

In a next step (not shown here) the procedure can be used in order to construct a fully automated curve recognition protocol e.g. for the qualification of incoming goods. The establishment of classes [4] allows sorting sample curves of incoming goods into “Pass” and “Fail”. The Identify algorithm can then automatically assign a newly measured DSC curve of a powder coating to either the “Pass” or the “Fail” class.

Published in:

European Coatings Journal 11 – 2015, www.european-coatings.com

Literature

[1] DIN EN ISO 11357: “Plastics – Differential Scanning Calorimetry (DSC)”, www.beuth.de, 2010

[2] WF. Hemminger, HK. Cammenga: “Methoden der Thermischen Analyse”, Springer, Heidelberg, 1989

[3] GW. Ehrenstein, G. Riedel, P. Trawiel: “Thermal Analysis of Plastics: Theory and Practice”, Hanser Gardner Publications, 2004

[4] A. Schindler: “Identify – How this New DSC Curve Recognition System Simplifies Polymer Characterization”, White Paper, Netzsch-Gerätebau GmbH, 2013

[5] E. Moukhina, A. Schindler: “Automatic Evaluation and Identification of DSC curves”, Presentation during International Symposium “Thermal Analysis and Calorimetry in Industry”, Berlin, 2014

[6] E. Füglein, E. Kaisersberger: “About the development of databases in thermal analysis”, J. Therm. Anal. Calorim. (2015) 120:23-31