JP2814603B2 - Adhesion judgment method of multilayer printed wiring board - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention uses an ultrasonic impulse to detect an uncured portion or a peeled portion of an adhesive layer in judging the adhesion of a multilayer printed wiring board manufactured by bonding a composite material. The present invention relates to a non-destructive method for judging adhesion by using the method.

(Prior art)

The demand for multilayer printed wiring boards, on which IC chips and electronic components can be mounted and three-dimensionally electrically wired, can be stored in a small space with high reliability in electronic circuits, is increasing in demand. .

As shown in FIG. 5, a cross-sectional view of this multilayer printed wiring board is formed by forming a copper conductive wiring 14 on a board made of a composite material containing glass fibers 11 in an epoxy resin 12. It is created by laminating using a system adhesive 13.

One of the problems that can occur in the manufacturing process of a multilayer printed wiring board is poor adhesion such as generation of an uncured portion of an adhesive layer that adheres each substrate and delamination due to mixing of foreign matter. Occurrence of poor adhesion causes disconnection of wiring formed on the surface of each layer, destruction of a joint with an electronic component to be mounted, and the like, thereby lowering the reliability of a manufactured printed wiring board. For this reason, it is necessary to detect poor adhesion portions such as uncured portions and peeled portions of the adhesive layer.

As a means for detecting these, a pulse mode ultrasonic flaw detection method has been used conventionally. One example will be described with reference to FIG. An electric signal obtained by the broadband pulse generator 1 is applied to the piezoelectric transducer 2, and an ultrasonic pulse 4 having a narrow width on the time axis of about 10 MHz is transmitted.
It is transmitted to the subject 7 via an ultrasonic wave propagation liquid (not shown). The transmitted ultrasonic pulse is reflected by the exfoliation portion 8 when the exfoliation portion 8 is present on the subject surface, the subject bottom surface, and the subject. The reflected wave is received by a piezoelectric transducer and converted into an electric signal. When this electric signal is observed by voltage-time with an oscilloscope 5 (see FIG. 6B), the reflected signal a from the surface of the subject and the multilayer A reflection signal c from the bottom surface of the printed wiring board is observed. If there is a poor adhesion portion such as the adhesive peeling portion 8, the reflection signal b is observed during that period. Further, the depth (see d in FIG. 6A) of the peeled portion is obtained from the time difference t between the reflected signal a from the subject surface and the reflected signal b from the peeled portion 8.

[Problems to be solved by the invention]

However, it is extremely difficult to use the above-described pulse mode ultrasonic inspection for judging the adhesion of a multilayer printed wiring board for the following reasons.

First, 10MHZ used in pulse mode ultrasonic testing
Ultrasonic pulse of high frequency causes ultrasonic absorption by epoxy resin and scattering by glass fiber inside,
Reflected waves from poorly adhered parts such as peeled parts and uncured adhesive parts become weak enough to be undetectable, or ultrasonic pulses are reflected by conductive wiring made of copper foil, resulting in deeper information. Can not be obtained. Further, even if it is possible to transmit through the conductive wiring by using a low-frequency ultrasonic pulse, it is difficult to measure the depth of the poor adhesion portion because the reflected ultrasonic pulse spreads on the time axis.

Second, in the pulse mode ultrasonic inspection method, when a poor adhesion portion exists near the surface of the multilayer printed wiring board, the reflected wave from the multilayer printed wiring board surface and the reflected wave from the poor adhesion portion overlap, so that the poor adhesion occurs. Parts cannot be detected,
In some cases, it is difficult to measure the depth.

As described above, the conventional method is not a method suitable for judging the adhesion of a multilayer printed wiring board using a material that is strongly absorbed by ultrasonic waves or scattered by glass fibers inside.

[Means for solving the problem]

The present invention relates to a method for judging the adhesion of a multilayer printed wiring board manufactured by bonding a composite material, wherein an ultrasonic impulse having a frequency of 5 MHz or less is transmitted from a piezoelectric transducer, and the multilayer printed circuit is transmitted through an ultrasonic wave propagating liquid. A step of receiving the ultrasonic impulse reflected from the multilayer printed wiring board with a piezoelectric transducer and converting it into an electric signal; a step of obtaining a frequency distribution of reflected wave intensity from the electric signal; outputting on a frequency distribution Is a step of obtaining a frequency group in which the minimum is obtained, comparing the obtained frequency group with a frequency group in a good adhesive layer, and when both are greatly different, there is an uncured portion or a peeled portion inside the multilayer printed wiring board And a step of judging the adhesion of the multilayer printed wiring board.

[Detailed Description of the Invention]

In the present invention, as shown in FIG. 1, an electric signal generated by a broadband pulse generator 1 is input to a piezoelectric transducer 2 so that an ultrasonic impulse 4 of 5 MHZ or less is transmitted, and an ultrasonic wave propagating liquid (shown in FIG. ) Perpendicularly to the surface of the multilayer printed wiring board 3 via The reflected ultrasonic impulse 4 is received by the same piezoelectric transducer 2 as the transmission, and is further converted into an electric signal by the piezoelectric transducer 2 and frequency analysis is performed by the spectrum analyzer 6, and the reflected ultrasonic impulse 4 is reflected from the multilayer printed wiring board 3. Obtain the frequency distribution of the intensity of the sound impulse. Further, a digital oscilloscope and a controller may be used instead of the spectrum analyzer 6.

The inside of the multilayer printed wiring board 3 will be described with reference to FIG. 2 (a). The spectrum intensity-frequency distribution of the reflected wave obtained from the portion having good adhesion (see a and c in FIG. 2 (a)) FIG. 2 (b) for FIG. 2 (a) and FIG. 2 (d) for FIG. 2 (a) for (c) in FIG. 2 (a)) have a minimum reflected wave output a at a frequency interval ΔF expressed by the following equation. .

V is the propagation speed of the sound wave, and D is the thickness of the multilayer printed wiring board.

This is because the reflected component A of the ultrasonic impulse 4 from the surface of the multilayer printed wiring board and the reflected component B from the bottom of the multilayer printed wiring board cause interference.

The frequency of the ultrasonic wave used at the time of measurement is sufficient for an ultrasonic impulse of 5 MHz or less at which a reflected wave component having a frequency whose wavelength is equal to or less than half of the substrate thickness D is output. This is because pulse mode ultrasonic testing requires a frequency at which the wavelength inside the substrate is less than one-tenth of the thickness of the substrate, whereas this method uses much lower frequency ultrasonic waves.

In addition, the frequency used is relatively low at 5 MHZ or less for conductive wiring such as copper foil inside a multilayer printed wiring board, or for the presence of minute structures such as glass fiber, so that ultrasonic impulse Is less susceptible to scattering and reflection.

Furthermore, if the adhesive of the multilayer printed wiring board is also sufficiently cured, the adhesive layer is generally very thin, and its acoustic impedance is close to the impedance of a substrate such as epoxy. It has been confirmed that it is not reflected.

However, when there is a peeling portion (see (b) in FIG. 2 (a)) inside the multilayer printed wiring board, the frequency distribution of the intensity of the reflected wave differs. FIG. 2 (c) is a spectrum intensity-frequency distribution measured at the position shown in FIG. 2 (a), b. Since the ultrasonic impulse is almost completely reflected at the separation part, the depth of the separation part The reflected wave intensity minimum b is shown in the same frequency group.

Further, when an uncured portion of the adhesive exists, more complicated reflection occurs inside the multilayer printed wiring board. Since the uncured part of the adhesive is close to liquid and has a very different impedance value from the base material of the multilayer printed wiring board, strong reflection occurs at this uncured part and the uncured part of the adhesive has passed through the uncured part. The component of the ultrasonic impulse is reflected at the bottom of the multilayer printed wiring board, and is reflected again at the uncured portion of the adhesive and interferes, thereby obtaining a different frequency group from that of the cured adhesive layer.

Further, since a peeled portion or an uncured portion of the adhesive is generated only in the adhesive layer, the depth of the poor adhesion portion does not become smaller than the thickness d of the uppermost layer of the multilayer printed wiring board (see FIG. 2A). From this, it is sufficient that the piezoelectric transducer used has a frequency distribution up to the minimum frequency interval ΔF of the reflected wave output corresponding to the thickness d of the uppermost layer.

Further, as shown in FIG.
The ultrasonic impulse 4 transmitted from 2a is obliquely incident on the multilayer printed wiring board surface 3 via an ultrasonic wave propagating liquid (not shown), and the broadband ultrasonic impulse 4 is incident on the multilayer printed wiring board 3, The reflected wave can be received by another piezoelectric transducer 2b, and the frequency distribution of the reflected wave intensity from the multilayer printed wiring board 3 can be obtained.

When the ultrasonic impulse 4 is obliquely incident on the surface of the multilayer printed wiring board at the portion where the adhesive layer inside the multilayer printed wiring board is completely cured (see (a) and (c) in FIG. 4 (a)), it is called a plate wave. Elastic waves are excited in the multilayer printed circuit board for specific frequency components, and the elastic waves propagate through the multilayer printed circuit board while re-emitting the energy into the ultrasonic wave propagating liquid.

In the case of a substance that strongly absorbs ultrasonic waves, such as a multilayer printed wiring board, the energy is consumed while propagating in the substrate, so that the amount of energy re-emitted into the ultrasonic wave propagation medium is small. For this reason, the frequency distribution of the reflected wave intensity has a minimum a in the frequency group that excites the elastic wave. FIG. 4 (b) is a spectrum intensity-frequency distribution measured in FIG. 4 (a), and FIG. 4 (d) is a spectrum intensity-frequency distribution measured in FIG. 4 (a). is there.

This phenomenon is observed as a stronger phenomenon because the elastic wave propagates to a region where the receiving piezoelectric transducer cannot receive.

The frequency f for exciting the elastic wave is determined by the following equation.

fxD = C (2) where D is the thickness of the multilayer printed wiring board, and C is a constant determined by the angle of incidence of the ultrasonic impulse on the multilayer printed wiring board, and It can be obtained from the physical constants of the multilayer printed wiring board by elasticity theory or actual measurement experiments.

Next, a change in the frequency distribution of the reflected wave intensity in the case where the adhesive layer has a peeled portion and an uncured portion of the adhesive will be described.

Even in the case where there is a peeling portion, the elastic wave is excited as described above by obliquely inputting the ultrasonic impulse. However, in this case, the frequency at which the elastic wave is excited can be obtained by substituting the distance between the surface of the multilayer printed wiring board and the peeled portion into the second equation. Therefore, the value of the frequency group in which the reflectance intensity has the minimum value becomes large in the area where the separation has occurred.

Next, it has been confirmed by numerical calculation based on the theory of elasticity that even when an uncured portion of the adhesive occurs, a group of minimum reflectance frequencies different from that in the case of good adhesion is obtained. In general, when an uncured portion, that is, a liquid is present in a multilayer printed wiring board, as shown in FIG. 4 (a), the difference in impedance between the solid and the solid is large. The ultrasonic impulse 4 is strongly reflected at the boundary of the part, and furthermore, only the longitudinal wave sound propagates in the liquid, so that the elastic wave is excited in the multilayer printed circuit board, so that the adhesive is hardened. Then, a spectrum intensity-frequency distribution (see FIG. 4 (c)) of the reflected wave intensity which is different from the case of FIG.

However, in the detection of the uncured portion of the adhesive, the measurement in which the ultrasonic impulse is obliquely incident on the surface of the multilayer printed wiring board has a higher output on the frequency distribution of the reflected wave than the measurement by the vertical incidence. It is suitable for detection because the minimum frequency group moves greatly.

In this way, the frequency group that excites these elastic waves is affected by the pulse-mode ultrasonic flaw detection method against the effects of conductive structures made of copper foil or minute structures such as glass fibers inside the multilayer printed wiring board. Therefore, the method is effective because the ultrasonic impulse is hardly affected by scattering and reflection. Furthermore, if the adhesive layer formed when bonding a large number of printed wiring boards using the adhesive is sufficiently cured, the adhesive layer is generally extremely thin, and its acoustic impedance is the impedance of a substrate such as epoxy. , It has been confirmed that it is not greatly affected by the frequency distribution.

As described above, it is possible to detect a poor adhesion portion such as a peeled portion or an uncured portion of the adhesive inside a large number of printed wiring boards from a change in the output minimum frequency group on the frequency distribution of the reflected wave intensity.

[Example 1] Adhesion of a multilayer printed wiring board manufactured by bonding 10 printed wiring boards with a thickness of 2 mm was determined. The printed wiring board is a composite material containing glass fibers in an epoxy resin, and the adhesive is made using a two-component epoxy-based adhesive.

As shown in FIG. 1 (a), a broadband electric signal was generated by a broadband pulse generator 1, applied to a piezoelectric transducer 2, and an ultrasonic impulse 4 was vertically incident on a multilayer printed wiring board 3. As an ultrasonic wave propagating liquid (not shown), liquid Freon, which is an inert liquid, was used so as not to damage copper electric wiring. The reflected ultrasonic impulse from the multilayer printed wiring board 3 is applied to the same piezoelectric transducer 2.
And the output signal was passed through a spectrum analyzer 6 to obtain the frequency distribution of the reflected wave output. As shown in FIG. 2 (a), when frequency distributions were observed at several points from (a) to (c), as shown in (b), the frequency group causing the output minimum was extremely small (see FIG. 2 (c)). It was found that b) in the parentheses) exists in this multilayer printed wiring board. The velocity of longitudinal ultrasonic waves in this multilayer printed wiring board is about 3300m /
From the equation (1), it was found that when the bonding was completely performed, the minimum frequency group of the reflected wave appeared at 0.82 MHz in the observable frequency region using the first equation. The output minimum frequency other than b was 0.80 MHZ, and the presence of the 10 μm adhesive layer and the minimum of the reflected wave output based on the glass fiber were not observed. However, the lowest frequency group of reflected wave output was observed at 3.3MHZ in B. Since this cannot be explained by the change in the thickness of the multilayer printed wiring board, it was inferred that the peeled portion was present at the lower part of B and the depth was 0.5 mm using the first formula. Later, when a cross section was observed, it was found that a peeled portion was present in the adhesive layer having a depth of 0.5 mm.

Example 2 The adhesion of a multilayer printed wiring board manufactured by bonding two printed wiring boards with a thickness of 1 mm and two printed wiring boards was determined. PWB its longitudinal acoustic velocity of a composite material containing glass fibers in epoxy resin 3300 m / s, shear wave velocity is 1600 m / s, a specific gravity of about 2.0 g / cm ³ on average, the adhesive It was prepared using a two-part mixed epoxy type.

As shown in FIG. 3A, a piezoelectric transducer 2a for transmitting ultrasonic waves and a piezoelectric transducer 2b for receiving ultrasonic waves.
Were made to face each other at an angle of 30 degrees from the normal line of the surface of the multilayer printed wiring board. The apparatus was set so that the frequency distribution of the reflected waveform from the receiving piezoelectric transducer 2b was determined by the spectrum analyzer 6. As shown in FIG. 4 (a), when the frequency distribution of the reflected wave output was measured from the surface of the multilayer printed wiring board A to the surface C, the elastic wave was observed when the thickness of the multilayer printed wiring board was set to 1 mm except for B. The excited frequency was 1.85 MHZ, which was very close to 1.9 MHZ. However, in the case of B, the frequency distribution of the reflected wave output having a minimum in different frequency groups was obtained. For this reason, as a result of performing cross-sectional observation by destructive inspection, it was found that the adhesive was uncured due to an error in the compounding operation.

〔The invention's effect〕

Since the present invention is configured as described above, in a manufacturing process of a multilayer printed wiring board manufactured by bonding a composite material, or in an inspection process thereof, non-destructive disconnection of wiring, reduction in product reliability is reduced. It has become possible to detect a poor adhesion portion such as an uncured portion of the adhesive or a peeled portion, which is a cause.

[Brief description of the drawings]

FIG. 1 is an explanatory view of an embodiment in which an ultrasonic impulse is vertically incident on the surface of a multilayer printed wiring board, FIG. 2 (a) is a cross-sectional view of a multilayer printed wiring board (two layers) having a peeling portion, FIG. 2 (b) is a graph showing the spectrum intensity-frequency distribution at (a) in FIG. 2 (a), and FIG. 2 (c) is the spectrum intensity at (b) in FIG. 2 (a).
FIG. 2 (d) is a graph showing the spectrum intensity-frequency distribution in FIG. 2 (a), and FIG. 3 is an oblique incidence of the ultrasonic impulse on the surface of the multilayer printed wiring board. FIG. 4 (a) is a sectional view of a multilayer printed wiring board (two layers) having an uncured portion in an adhesive layer, and FIG. 4 (b) is FIG. 4 (a). FIG. 4 (c) is a graph showing the spectrum intensity-frequency distribution in FIG. 4 (a), and FIG. 4 (d) is a graph showing the spectrum intensity-frequency distribution in FIG. a) Spectrum intensity in medium C
FIG. 5 is a graph showing a frequency distribution, and FIG. 5 is a sectional view of a multilayer printed circuit board. FIG. 6 is an explanatory diagram of a conventional pulse mode ultrasonic flaw detection method. 1 Broadband pulse generator 2, 2a, 2b Piezoelectric transducer 3 Multilayer printed circuit board 4 Ultrasonic impulse 5 Oscilloscope 6 Spectrum analyzer 7 Subject 8 Stripping section 9 … Uncured area 10… Hardened area 11… Glass fiber 12… Epoxy resin 13… Adhesive layer 14… Conductive wiring

────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Katsumi Ohira 1-5-1, Taito, Taito-ku, Tokyo Toppan Printing Co., Ltd. Examiner Katsumi Enari (56) References JP-A-64-91056 (JP, A ) Patent 2650344 (JP, B2) JP 7-9418 (JP, B2) (58) Fields investigated (Int. Cl. ⁶ , DB name) G01N 29/00-29/28

Claims (1)

(57) [Claims] 1. A method for judging the adhesion of a multilayer printed wiring board manufactured by bonding a composite material, comprising the steps of: (a) transmitting an ultrasonic impulse having a frequency of 5 MHz or less from a piezoelectric transducer, and (B) receiving an ultrasonic impulse reflected from the multilayer printed wiring board by a piezoelectric transducer and converting it into an electric signal; and (c) converting reflected electric wave intensity from the electric signal into an electric signal. A step of obtaining a frequency distribution, (d) a step of obtaining a frequency group having a minimum output on the frequency distribution, and (e) comparing the frequency group with a frequency group in a good adhesive layer, and the two were greatly different. Judging that there is an uncured portion or a peeled portion inside the multilayer printed wiring board. Method.