Qualitative and quantitative characterization of outgassing from SU-8 Joost Melai a, * , Cora Salm a , Rob Wolters a,b , Jurriaan Schmitz a a Semiconductor Components, MESA + institute, University of Twente, Hogekamp 3138, P.O. Box 217, 7500 AE Enschede, The Netherlands b NXP Research, Eindhoven, The Netherlands article info Article history: Received 28 September 2008 Accepted 11 November 2008 Available online 21 November 2008 Keywords: SU-8 Outgassing Vacuum Mass spectrometry Gas chromatography Packaging Microsystems abstract SU-8 is often used as a structural material in Microsystems. In this work, the outgassing characteristics from such cross-linked SU-8 layers are studied using mass spectrometry and gas-chromatography tech- niques. With these methods the composition of the released matter can be identified, also the outgassing rate can be quantitatively characterized. A qualitative estimate of outgassing from SU-8 is given for cross- linked layers. The effect of hard-bakes is studied in situ by measuring at typical hard-bake temperatures. These tests indicate that a hard-bake is needed to provide good performance in UHV environments. Using gas-chromatography the outgassing rate from SU-8 is determined. The total outgassing rate is inversely proportional with time which further illustrates the effect of a hard-bake step. Ó 2008 Elsevier B.V. All rights reserved. 1. Introduction In microtechnology SU-8 [1] is widely used for making high as- pect ratio permanent structures. Many important material aspects have been investigated, as recently summarized in a review article [2]. Especially the lithographic patterning capabilities and mechan- ical aspects have been widely studied. Outgassing remained, to the best of our knowledge, undocumented. With some applications SU-8 structures are used in gaseous environments with a critical composition [3], other applications would have SU-8 structures operating in vacuum [4]. In these cases the impact of the SU-8 on the ambient is of vital importance. Precise characterization of and influence over outgassing from SU-8 is also important when SU-8 is used in wafer-scale packaging [5]. In this study, we present our investigations on the outgassing of SU-8 using time-resolved mass spectrometry (MS) and gas- chromatography. 2. Experimental The SU-8 layers under study are formed on top of bare 4 00 Si wa- fers. To prevent additional outgassing signals the wafers are first cleaned extensively in fuming HNO 3 at room temperature (10 min) and 69% HNO 3 at 95 °C (10 min) and subsequently rinsed in DI water. After this a short 2% HF dip is performed to remove na- tive oxide, the hot 69% nitric acid step is repeated to regrow a thin chemical oxide and the substrates are rinsed again. After a 10 min dehydration bake at 120 °C the SU-8 layer is deposited by spin coating at 3000 rpm, a primer is not used. The wafers are given a 20 min Soft-Bake at 95 °C, using a hotplate, to solidify the resist layer. Next a blanket UV exposure with a dose of 290 mJ/cm 2 is performed to initiate cross-linking of the entire layer. A post-expo- sure bake is performed for 10 min at 80 °C. All bakes are done using long ramp-up and ramp-down times as per instructions given by the supplier [1]. The wafers are diced into 2 cm 2 cm samples using a Disco DAD-321. Immediately after drying (using a N 2 spray gun) the samples are packaged in Al foil that is folded twice around the sample, the packaged samples are stored in sealed bags until they are measured. Some samples have received an additional 10 min hard-bake (HB) at 150 °C, also on a hotplate, prior to packaging. All initial measurements are performed on layers of SU-8 50 of 50 lm thick. Additional measurements on thinner layers of 2 and 10 lm have also been performed. For these SU-8 formulations with a lower viscosity are used (SU-8 2 and SU-8 5, respectively), the spin speeds are lower and the bake times are shorter. Two analysis techniques have been used. For MS analysis the SU-8 samples are placed in a vacuum chamber, a mass spectrome- ter (Pfeiffer QMS Prisma) is connected to monitor the gas flow coming out of the system. The specimen chamber can be elevated in temperature by resistive heating. Further measurements are performed using a combined gas- chromatography and mass-spectrometry technique (GC–MS). The samples are heated in an oven that is flushed continuously with 0167-9317/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.mee.2008.11.008 * Corresponding author. Tel.: +31 534894394; fax: +31 534891034. E-mail address: j.melai@utwente.nl (J. Melai). Microelectronic Engineering 86 (2009) 761–764 Contents lists available at ScienceDirect Microelectronic Engineering journal homepage: www.elsevier.com/locate/mee