iMaps Announcements 06/2024

In-situ investigation of the cross-linking process of epoxy resins for the detection of fluctuations in encapsulation processes of electronic components

Abstract : Epoxy resins are frequently used as encapsulation materials for the encapsulation of electronic components in electronic packaging technology. The chemical cross-linking process and the associated curing of these resins during processing have a considerable influence on the production and ultimately on the quality of the components. A Fourier transform infrared sensor (FT-IR) was integrated into the tool of a molding press in order to measure in-situ infrared spectra and thus chemical changes during the process. An evaluation method of the infrared spectra is presented, for which the temporal progress of the crosslinking is determined by calculating integrals of relevant wavenumber ranges per time step and a subsequent normalization of the integral curve. A comparison with conventional dielectric analysis (DEA) shows a similar curve progression, whereby the FT-IR sensor can be used to measure changes that are not visible with DEA, particularly for high degrees of crosslinking.

1 Introduction

In the current electronic component industry, the encapsulation process of electronic components often proves to be an indispensable step that protects the components from external influences, isolates electrical contacts and defines the shape of the package [1]. This process is carried out by means of a molding process, whereby the predominant methods are transfer molding and compression molding and epoxy resin molding compounds (EMC) are usually used as the material [2]. The encapsulation process is challenging due to fluctuations in the production batches or changes in the material properties and the associated adjustment of specific process parameters [3]. These factors have a significant influence on both the processing and the quality of the end product [3]. The selection of process parameters is primarily based on empirical data, as the understanding of the process is limited. By using different measurement methods, the cross-linking process of the EMC is systematically recorded, which enables a deeper understanding of the process [4].

2. fundamentals

In order to determine the chemical crosslinking state of the EMC, dynamic differential calorimetry is usually used, in which the heat flow released when the material is heated is measured and correlated with the degree of crosslinking [5]. In order to detect the temporal development of the chemical cross-linking during the ongoing process, DEA has become established in recent years [6]. Here, a sinusoidal voltage is applied and the response signal is measured, whereby ion mobilities are recorded that run in the opposite direction to the ion viscosity and correlate with the degree of crosslinking of the EMC [7]. Another in-situ measurement method is Fourier transform infrared spectroscopy (FT-IR). This method uses light in a wavenumber range between 600 cm-¹ and 1900 cm-¹ (mid-infrared spectroscopy) [8]. The light source is a spectrometer whose light is guided to the material during the process using optical fibers, where it is partly absorbed and partly reflected and returned to the spectrometer. The reflected light is used to generate spectra that are characteristic of the respective material composition and the current chemical crosslinking state [9]. In this way, the cross-linking reaction can be detected by changes in the chemical bonds of the molecules [10][11].

Materials and methodology

In this study, a single-layer substrate material with a thickness of 200 μm and a highly filled, pre-compressed EMC molding compound with a filler content of about 90 % are used. A release film made of LM-ETFE is used to protect the mold and for subsequent demolding of the component. The encapsulation process used here is known as local pressure molding [12]. The material is processed at a mold temperature of 170 °C, an internal mold pressure of 50 bar, a piston feed rate of 0.2 mm/s and a cycle time of 900 s. In order to monitor the degree of cross-linking of the EMC during the process, dielectric sensors and an FT-IR sensor are implemented in the mold, which measure in direct contact with the EMC. A detailed description of the measurement setup is explained in [8].

4. investigation of the curing process of epoxy resins 4.1 In-situ infrared spectra

A total of 115 usable FT-IR spectra can be recorded during an encapsulation process with a cycle time of 900 seconds. Figure 1 shows every tenth recorded FT-IR spectrum for a molding test. The transmission is plotted against the wavenumbers, with bands forming that are characteristic of the material composition of the EMC.

The first significant shoulder, which represents the epoxy band, can be seen at wavenumber 916 cm-1. This shoulder disappears during the encapsulation process as the epoxy ring is broken up during cross-linking. Another characteristic band can be seen at wavenumber 1006 cm-1, which represents the silica filler. Due to the high filler content, very low transmission values are formed for this band. In addition, during the cross-linking process, a band is formed at the wavenumber 1230 cm-1, which represents the reaction of the phenolic resin, which reacts with the epoxy resin as a hardener. Furthermore, characteristic bands can be recognized at the wavenumbers 1450 cm-1, 1508 cm-1 and 1610 cm-1, which represent a C-C double bond that can occur both in the resin and in the hardener of the EMC. Initial results have already been published in [8].

Fig. 1: FT-IR spectra measured at different times during an encapsulation process

4.2 Evaluation of chemical cross-linking from infrared spectra

The recorded FT-IR spectra are used to monitor the cross-linking of the epoxy resins during processing. The chemical cross-linking process becomes visible in the FT-IR spectra by measuring changes in the existing bands. However, there are also wavenumber ranges that do not change with increasing cross-linking. For this reason, it is important to identify the relevant wavenumber ranges that show purely systematic changes during chemical crosslinking. To identify these wavenumber ranges, the differences of the infrared spectra are formed in relation to the first recorded spectrum, with every tenth difference plotted over the wavenumbers in Figure 2. The three areas with the largest differences are located around the bands 916 cm-1, 1230 cm-1 and 1450 cm-1, which can be associated with the chemical reaction of the epoxy resin with the phenolic resin. In order not to consider only a single band, three wavenumber ranges are defined around the three bands: from 848 cm-1 to 1036 cm-1, from 1196 cm-1 to 1254 cm-1 and from 1444 cm-1 to 1508 cm-1. Within these ranges, the transmission values of the spectra decrease with increasing cross-linking. In order to be able to draw conclusions about the change in cross-linking over time, the dimension must be reduced, as a spectrum rather than a scalar value is available for each point in time. For this reason, the integral is calculated for each recorded FT-IR spectrum within the previously defined wavenumber ranges and the individual integrals are added up. In this way, the spectrum is reduced to one value per time point and the change in the integral can be plotted over time. This change is representative of chemical changes in the process due to the cross-linking process. Figure 3 (blue) first shows the evaluated integrals for an encapsulation process. It can be seen that the value of the integral decreases continuously up to a cycle time of 400 s until it finally stagnates. If no change is visible in the integral, it is assumed that the cross-linking of the EMC is complete. In order to obtain the cross-linking curve, the integral curve is normalized between zero and one(Fig. 3 in red).

4.3 Comparison with dielectric analysis

In recent years, DEA has become established for the in-situ measurement of crosslinking, whereby the time course of the crosslinking can be determined by normalizing the dielectric signal [6]. Figure 4 shows the cross-linking of the EMC evaluated with the DEA (blue) and using the FT-IR sensor (red) over time. The FT-IR sensor shows the first measurement point at a cycle time of 21 s, while the DEA is only capable of measurement from a cycle time of 37 s. Both cross-linking curves therefore increase with time until a plateau is finally reached. The crosslinking measured with the FT-IR sensor is higher at the beginning compared to the DEA. The detection of low degrees of crosslinking appears to be more accurate with the FT-IR sensor, as the measuring principle is used to measure purely chemical changes and the signal is not influenced by melting effects compared to DEA. At a cycle time of 82 s, the cross-linking curve of the DEA finally intersects the cross-linking curve of the FT-IR signal before the dielectric curve reaches the plateau at a cycle time of approx. 255 s. Here, the cross-linking evaluated on the basis of the infrared signal only shows a degree of cross-linking of 84%. The infrared sensor continues to show changes in the degree of cross-linking up to a cycle time of 450 s. This indicates that the FT-IR sensor can be used to detect chemical changes in the high degrees of crosslinking that cannot be resolved with the DEA.

Fig. 4: Evaluated temporal change in the degree of crosslinking of the vEMC with DEA (blue) and FT-IR (red)

5 Summary and outlook

The presented method offers a new method for in-situ monitoring of the chemical crosslinking of epoxy resins in encapsulation processes using infrared spectra measured during the process. By calculating integrals of relevant wavenumber ranges of the infrared spectra per time step, an integral over time curve is determined that is representative of chemical changes in the process. Subsequent normalization of the integral curve leads to the crosslinking curve, which can supposedly resolve the crosslinking of the epoxy resin more sensitively compared to conventional DEA. Changes in the cross-linking that cannot be resolved with DEA can be measured, especially at high degrees of cross-linking. This not only offers the possibility of optimizing the process, but also of detecting individual chemical reactions as sub-processes of the entire cross-linking process during the encapsulation process.

Calendar of events

IMAPS Germany - Your association for packaging and packaging technology

IMAPS Germany, part of the International Microelectronics and Packaging Society (IMAPS), has been the forum in Germany for all those involved in microelectronics and packaging technology since 1973. With almost 300 members, we essentially pursue three important goals:

• we connect science and practice
• we ensure the exchange of information among our members and
• we represent the point of view of our members in international committees.

Imprint

IMAPS Germany e. V.
Kleingrötzing 1,
D-84494 Neumarkt-St. Veit

1st Chairman: Prof. Dr.-Ing. Martin Schneider-Ramelow, Director of the Fraunhofer Institute for Reliability and Microintegration (IZM)
This email address is being protected from spambots. You need JavaScript enabled to view it.

Treasurer
(for questions about membership and contributions): Ernst G. M. Eggelaar, This email address is being protected from spambots. You need JavaScript enabled to view it.

You can find detailed contact information for the board members at www.imaps.de
(Board of Directors)

Literature

Sasajima, H.; Watanabe, I.; Takamoto, M.; Dakede, K.; Itoh, S.; Nishitani, Y.; Tabei, J.; Mori, T.: New Development of Epoxy Molding Compound for Encapsulating Semiconductor Chips, Springer, (2009)
Komori, S.; Sakamoto, Y.: Development Trend of Epoxy Molding Compound for Encapsulation Semiconductor Chips, Springer, (2009)
Kaya, B.; Kaiser, J.; Becker, K.; Braun, T.; Lang, K.: Evaluation of the Dielectric Cure Monitoring of Epoxy Molding Compound in Transfer Molding Process for Electronic Packages, in IEEE Proceedings of the European Microelectronics Packaging Conference (EMPC), Friedrichshafen, Germany, September 14-16, 2015.
Krahnbühl, D.; Hood, D.; Rogozinski, J.; Meyer, A.; Neag, M.: Monitoring the changing state of a polymeric coating resin during synthesis, cure and use, Progress in Organic Coatings, Vol. 35, (1999), No. 1-4, pp. 101-107
Gerami, G.; Bagheri, R.; Darvishi, R.: Investigation of isothermal and dynamic cure kinetics of epoxy resin/nadic methyl anhydride/dicyandiamide by differential scanning calorimetry (DSC), Journal of Thermal Analysis and Calorimetry, vol. 137, (2019), pp. 575-582
Franieck, E.; Fleischmann, M.; Hölck. O.; Kutuzova, L.; Kandelbauer, A.: Cure Kinetics Modeling of a High Glass Transition Temperature Epoxy Molding Compound (EMC) Based on Inline Dielectric Analysis, Polymers, Vol. 13, (2021), No. 11, 1734
Lee, H.: The Handbook of Dielectric Analysis and Cure Monitoring, Lambient Technologies LLC, Cambridge, USA, (2017)
Niegisch, C.; Haag, S.; Braun, T.; Hölck, O., Schneider-Ramelow, M.: In-situ infrared spectroscopy for chemical analysis in electronic packaging processes, In IEEE Proceedings of the 9th Electronics Systems-Integration Technology Conference (ESTC), Sibiu, Romania, September 13-16, 2022.
Bienz, S.; Bigler, L.; Fox, T.; Maier, H.: Spektroskopische Methoden in der organischen Chemie, 9th ed. Stuttgart: Georg Thieme Verlag, (2016)
Cholake, S.; Mada, M.; Raman, R.K.; Bai, Y.; Zhao, XL; Rizkalla, S.; Brandyopadhyay, S.: Quantitative analysis of curing mechanisms of epoxy resin by Mid- and ear-Fourier Transform Infra Red Spectroscopy, Def. Sci. Journal, Vol. 64, (2014), No. 3, pp. 314-321
Salzmann, M.; Blößl, Y.; Todorovic, A.; Schledjewski, R.: Usage of Near-Infrared Spectroscopy for Inline Monitoring the Degree of Curing in RTM Processes, Polymers, Vol. 13, (2021) No. 18, 3145, pp. 1-13
Schaaf, U.; Kugler, A.: Device and method for the production of components at least partially covered with a casting compound, (DE 10 2017 216 711 A1 2019.03.21). German Patent and Trade Mark Office, (2019)