USAGE OF SIMS AND AUGER TEST METHODS IN THE PRODUCTION OF MULTILAYER PRINTED CIRCUIT BOARDSt

The quality and reliability of multilayer boards are determined by the adhesion strength between the copper sheets and the epoxy-glass laminates. The adhesion properties of copper foil may be improved by mechanical or chemical roughening. The most efficient method is, however, to oxidize the copper surface.


INTRODUCTION
Multilayer boards are built up from simple etched boards, by means of hot lamination with "pre-preg" (special epoxy glass laminate).The crossection of a multilayer board is illustrated in Figure 1.Conductive layers may be signal, earth or simple layers.
Adhesion of the copper foil is characterized by the peel strength.The adhesive strength between the mechanically or chemically roughened copper foil and the "pre-preg" is generally insufficient.The peel strength is as low as 0.3 to 0.7 N/mm, This drilled plated and resin filled hole /drilled and plated hole

FIGURE
Cross section of a multilayer board.
Paper originally presented at the 5th International Spring Seminar on Electrotechnology held at Prenet, Czechoslovakia, 1-4 June, 1982.
has been recognised already by producers of copper cladded laminates, inducing them to oxidize the copper foil surface contacting the plastic board.This oxide layer provides an adequate adhesion, endowing the treated copper foils with peel strengths of 1.4 to 1.7 N/mm.Copper foil oxidation has become general also in the production of multilayer boards.The etched board is oxidized in an electrolyte containing sodium chloride.In principle, the chemical process seems to obey a formula given by W.S. DeForset and H.V. Conelly.
2 Cu + NaC1Oz + 2 HzO 2 Cu(OH)z + NaC1  CuO + H Cu + H,O (3) a statement confirmed by A.A. Fedulova, E.P. Kotov and E.R. Javics.This reaction disconnects the copper oxide and the epoxy resin, and the forming water is a catalyst to epoxy resin decomposition above 200C.Thus, if the formula (Equation 1) were true, delamination would be much more intense.stronger than unoxidized foil.Thus the copper surface was assumed to be coated by Cu20, as, in the present case, secondary valence bonds were possible.Oxide layers were made using different preparation parameters and peel strengths were determined.To test the oxide layer on the best boards by thermogravimetry, DTG and TG diagrams of CuO and CuzO were recorded (Figure 2).Also mixes ofboth oxides in different proportions were tested (Figure 3).The temperature differences between the DTG peaks were too small to distinguish the two oxides.(These tests were made at the Department of General and Analytical Chemistry, Technical University, Budapest).
Oxide layers made by us were tested by the SIMS and SAM methods.SIMS tests were made in Balzers equipment.These tests gave no absolute results on the surface layer but pointed to many impurities on the copper foil surfaces even ifuntreated, (e.g.Na, Mg, A1, Si, Fe, C1, etc.).
SAM tests were made in an equipment made by Physical Electronics, at a basic vacuum of2.10 -7 Pa.Tests were made with a primary energy of 5 keV, an electron beam of 8 m diameter and 1/.tA current intensity.The electron multiplier had a voltage of 1150 to 1500 V, and a twentyfold electronic amplification.Spectrum records were made with a time constant of sec.Argon gas supplied for ion sputtering had a partial pressure of 5.10 -4 Pa.The ion current had an intensity of to 10/x/L Auger electron spectra were recorded on various surface positions of the test samples.Thereafter the samples were bombarded with argon ions to remove surface layers ofvarious thickness, in order to record Auger energy spectra at different depths.The tests involved the recording of surface absorbed current pictures (ABS pictures).

RESULTS
Surfaces exhibited white and grey spots.Impurity quantities at two different depths from the untreated copper foil surfaces have been compiled in Table I., and at two different depths from the oxidized copper foil surface in Table II.
In the case of the oxidized copper foil, copper to oxygen ratios by weight show the oxide layer to be Cu20.In addition, the tests furnished interesting data on the layer impurities.According to L.J. Ritchie, the sodium carbonate content of the electrolyte is a maximum at 4 g/1.The carbon content of the top layer of the copper oxide layer may be assumed to be released by the Na2CO3.The relatively high C1 content is due to C1 in the original copper.
Thus the test results confirm the suitability of SIMS and SAM tests for development worlc FIGURE 2DTG and TG diagrams for CuO and Cu20.
oxide formation.According to P. Schuessler, CuO reacts with hydrogen released by the epoxy resin above 180C:

3 DTG
FIGURE 3    DTG and TG diagrams for mixed CuO/CuO oxides.