DE102004052039B4 - A method of determining the reflow depth during reflow of a metal layer and substrate for use in such method - Google Patents

Abstract

Method for determining the melting depth during melting of a metal layer (5, 6) by means of a laser beam (8), in which one or more layer (s) (3, 4) is / are present under the metal layer (5), which are The application of a test radiation emits / emits a respective layer-specific discharge radiation that is different from the discharge radiation of the metal layer, the discharge radiations differing in that they have different characteristic maximum wavelengths (λCu, λchar (4), λchar (3)); - During the melting, the melting area (7, 7 ', 7' ') is subjected to a test radiation (10), and - the melting of the layer (s) (5, 4, 3) from the reflected test radiation (11) by detection the layer-specific characteristic discharge radiation (λchar, Cu, λchar (4), λchar (3)) is recognized and the melting depth is deduced from this, the metal layer (5, 6) and the layer (s) (3, 4) being made from the same starting material and wherein the layer (s) (3, 4) is / are doped differently in layers to generate different characteristic discharge radiations (λchar, Cu, λchar (4) λchar (3)).

Description

The invention is in the field of the production of electrical circuits, in particular for power semiconductor modules, in which a metal layer applied to a carrier (substrate) is melted for assembly and / or contacting purposes. For this purpose, the metal layer is at least partially by heat input -. B. by means of high-energy radiation - heated to its melting point.

Because of the usually small dimensions and thus small amounts of material to be melted metal, the dosage of the heating intensity or heating time is of considerable importance. For reliable installations or electrical connections, a predetermined melting depth of the metal layer must be reproducibly achieved. A too low melting depth can lead to incomplete or faulty connections, while too great a melting depth or excessive heating can lead to thermal stresses and mechanical stresses - in the worst case, to the destruction of the circuit.

From the DE 41 38 157 A1 For example, a method for determining the thickness of a coating is known. In this case, a pulsed laser beam is directed onto the surface of a coated main body and evaporated with each laser pulse, a certain amount of material from the coating of the body, so that a plasma is formed. Each laser pulse deepens the recess created by evaporation. If the coating is removed locally down to the base body, then there is a significant change in the plasma spectrum, which indicates that the coating was locally removed up to the base body. From the number and energies of the laser pulses that were required to reach this stage, it is concluded that the thickness of the coating.

The present invention is therefore based on the object of specifying a method with which the melting point can be determined reliably reproducible during the melting of a metallization. It also raises the task of specifying a suitable substrate.

This object is achieved according to the invention by a method according to claim 1 and with respect to the substrate by a substrate according to claim 4. Embodiments and developments are the subject of the dependent claims and exemplary embodiments.

The method for determining the melting depth during the melting of a metal layer provides that the metal layer to be bonded or contacted is applied to one or more layers. This layer layer (s) is characterized or distinguished by the fact that they each have a layer-specific, different to the metal layer (and possibly to the other layers) different characteristic discharge radiation with respect to the metal layer and possibly to the other layer layers upon exposure to a test radiation emit or emit. The metal layer may also comprise a comparatively small metallized area, e.g. B. is used for load current decrease in a power semiconductor module.

h:

Plancksches Wirkungsquantum h = 6,2·10^–34 Js

f:

Frequenz der reflektierten Strahlung

λ_char:

charakteristische Wellenlänge der reflektierten Strahlung (Entladungsstrahlung)

c:

Lichtgeschwindigkeit

During the melting, the melting area is exposed to a test radiation. The proportion of the test radiation which is reflected back to an evaluation device is also referred to below as reflection radiation in the context of the invention. The reflection radiation contains a characteristic of the respective reflective material discharge radiation. The frequency or wavelength of the discharge radiation is dependent on the so-called work function. This in turn is determined by the irradiated, reflective material. Generally, the work function W _a and the frequency or wavelength of the discharge radiation for a given material are as follows: W _{a (material)} = h · f = h · c / λ _{char (material)} . (1) With:

H:

Planck's constant h = 6.2 · 10 ^-34 Js

f:

Frequency of the reflected radiation

λ _char :

characteristic wavelength of the reflected radiation (discharge radiation)

c:

Speed of Light

It can be seen from this that the wavelength of the discharge radiation is characteristic of the reflective material or its work function. Depending on the molten metal layer, one or more specific discharge radiation is / are therefore detectable in the reflection radiation.

Thus, when a layer consisting of a metal melts, a characteristic work function W _a or maximum wavelength λ _{char, Me} results according to the equation W _{a, Me} = h · f = h · c / λ _{char, Me} . (2)

For example, when the metal is copper (Cu), the corresponding characteristic work function will be hereafter referred to as W _{a, Cu} and the characteristic wavelength is denoted by λ _{char, Cu} .

If at least one further layer layer under the metal layer to be melted is achieved, i. H. If this layer layer begins to melt, a characteristic maximum wavelength of the discharge radiation, which is different from the characteristic wavelength of the copper layer, can be detected as a function of its material composition.

According to a preferred embodiment of the invention, it is provided that the metal layer and the layer layer (s) are produced from the same base material, the layer layer (s) being doped differently in layers to produce different characteristic discharge radiation.

If copper is used as the base material (starting material) in a preferred realization of the process, this can be applied in layers, for example, to produce a layer-specific characteristic discharge radiation. B. doped with aluminum and magnesium.

This would result in a first layer layer whose base material (copper) is doped by introduction of a dopant μ (eg aluminum) and which has a work function W _{a, μ} , according to the equation W _{a, μ} = h · f = h / λ _{char, μ} (3) a characteristic discharge radiation having the wavelength λ _{char, μ} , which is superimposed on the remaining radiation components in the reflection radiation.

From the reflection radiation, therefore, the successive melting of the layer layer (s) is detected by the fact that, depending on the melting depth in the reflection radiation different radiation components (namely, the layer-specific characteristic discharge radiation) are detected. From this it is concluded in a simple way to the current melting depth. The evaluation device for analyzing the reflection radiation or its spectral components may be a per se known spectrometer.

An essential aspect of the present invention is therefore that the reflection radiation contains material-specific characteristic discharge radiation having characteristic wavelengths.

An advantage of the invention is that the melting depth determination is performed in real time during the current reflow process. Thus, the measurement result can act directly on the control of the reflow process. In this way, extremely precise desired melting depths can be set reproducibly and reliably, thus optimizing the heat input. As a result, a secure connection can be ensured by sufficient melting and yet the material load can be minimized.

In a manufacturing technology preferred embodiment of the method according to the invention it is provided that the same laser device is used for melting the metal layer and for generating the test radiation.

The invention further relates to a substrate for use in a method according to any one of the preceding claims.

According to the invention, this substrate has a carrier with a plurality of metallic layers of a base material, each layer having an individual doping such that it emits a layer-specific characteristic discharge radiation when exposed to a test radiation. The base material is preferably copper.

More preferably, the doping of the layers is selected such that the respective layer-specific work function increases towards the uppermost layer. This advantageously enables an improved resolution in the spectral analysis of the contained radiation components (discharge radiation).

The invention will be further explained by way of example with reference to a drawing; show it:

1 a substrate used in carrying out the method according to the invention,

2 Construction and beginning of the process according to the invention,

3 a process state in which a desired melting depth is reached, and

4 a process state in which the desired melting depth is exceeded.

1 shows a substrate 1 that as a base or carrier 2 has a ceramic plate. There are three layers on it 3 . 4 . 5 applied, the base material, for example, each copper. But there are also other metals as basic material conceivable. The lowest (ie the base) layer layer 3 consists of magnesium-doped copper and has, for example, a thickness of 100 microns. The layer layer formed thereon 4 It is made of aluminum doped copper and has a thickness of for example, 50 microns. The upper metal layer on top 5 is made of copper and is 150 μm thick. On this layer is an example of an even thicker copper layer area 6 trained, the z. B. in an interconnection of a power semiconductor module as a load connection 6a can serve. In principle, it is also possible to provide further correspondingly individually doped lower layer layers, as a result of which the accuracy or resolution can be further increased in the melting depth determination to be described in detail below.

As explained above, the different doping effects the layers 3 . 4 . 5 different work functions. According to the relationship between the work function and the characteristic wavelength, when the layers are irradiated with a test radiation (eg laser radiation), discharge radiation each having a specific, layer-specific wavelength λ _{char (3)} , λ _{char (4)} , λ _{char, Cu results} . The choice of material or doping is chosen so that the work function of the lowest layer layer 3 to the top layer 5 respectively. 6 increases. As a result, an improved spectral resolution of the discharge radiation or the spectral components of the reflection radiation is possible.

2 shows a structure for carrying out the method according to the invention and the situation at the beginning of the procedure. You can see that in 1 shown substrate 1 with the metal layer to be melted 5 . 6 and the underlying layers 3 . 4 , An area 7 is with a laser beam 8th a laser device 9 applied. The laser beam brings highly concentrated and locally limited high energy into the layer material, whereby it is heated above its melting point, that is melted. Part of the laser radiation serves as test radiation 10 in that a part of the radiation on the layer 5 . 6 is reflected and as reflection radiation 11 to an evaluation device 12 is thrown back. The evaluation device 12 is a common glow discharge spectrometer (GDOS), this may have a detection range of 100 nm to 1000 nm wavelength. The detection or output signal 14 of the spectrometer 12 can serve as a control signal for the control of the laser device. A particular advantage of this embodiment is that for the actual melting and for the test radiation the same radiation source, namely here the laser device 9 , is used.

The melting can be used for laser welding to z. B. a dashed line indicated connection line 15 at the load connection 6a to fix.

In the in 2 As shown, the reflow process has only begun; the already molten area 7 is still completely in the metal layer 5 respectively. 6 , Therefore, in the reflection radiation, the discharge radiation of copper can be detected as the maximum wavelength. This is in the lower part of the 2 represented, in which the intensity I is plotted against the detected wavelength λ schematically. The maximum wavelength detected here is the characteristic wavelength of copper, λ _{char, Cu} . This indicates that the target reflow depth has not yet been reached, because the underlying layer serving as an indicator of an advanced reflow depth 4 not yet exposed and therefore can not generate any reflection yet.

In the in 3 - using the same reference numerals as in 1 and 2 for identical or identical elements - the situation shown is the reflow process in the area 7 ' advanced. It can be seen that the material of the (doped) copper layer layer 4 from the laser beam 8th has been reached and melted. Therefore, the reflection radiation contains 11 ' both (further) the discharge radiation λ _{char, Cu of} the layer 5 respectively. 6 as well as the discharge radiation λ _{char (4) of} the layer layer 4 , Since now - as in the lower part of the 3 shown in the intensity-wavelength diagram - the two specific characteristic, layer-specific wavelengths λ _{char (4)} and λ _{char, Cu} are detected by the spectrometer, it can be concluded in a simple manner that a melting depth up to the layer 4 is reached. In the exemplary embodiment was assumed (see. 1 ) that the metal layer 5 made of copper is 150 microns thick. If this is the desired minimum reflow depth of the layer layer 5 is, so that the achievement of the desired melting point is displayed and the melting process can be terminated here. At least a portion of the layer thickness (in this example 100 μm) of the layer remains as the residual melting depth 4 ,

4 shows - using the same reference numerals as in 1 to 3 for identical or similar elements - the farther in the range 7 '' advanced melting process. Here is the material of the layer layer 4 in the melting area 7 '' completely melted and the laser beam 8th reaches the layer position 3 , Therefore, the reflection radiation contains 11 '' both (further) the discharge radiation λ _{char, Cu of} the layer 5 respectively. 6 and the discharge radiation λ _{char (4) of} the layer layer 4 as well as the discharge radiation λ _{char (3) of} the layer layer 3 , Since now - as in the lower part of the 4 shown in the intensity-wavelength diagram J / λ - the specific characteristic, layer-specific wavelengths λ _{char (4)} , λ _{char (3)} and λ _{char, Cu)} are detected by the spectrometer, this indicates in a simple manner that the melting depth to to the layer position 3 enough. Thus, according to the exemplary embodiment (assuming a desired melting depth of 150 μm), it can be seen that the desired melting depth with the actual melting depth of at least 250 μm (sum of the thicknesses of the layers 4 and 5 ) is exceeded.

It can also be seen in the intensity-wavelength diagram J / λ that λ _{char (4)} <λ _{char (3)} ; thus the work function of the layer 4 is greater than that of the layer 3 , which allows for improved spectrometric resolution.

Instead of indicating the exceeding of the desired reflow depth, the layer layer could 3 and optionally further underlying layer layers are used to an even finer resolution of the determination of the melting depth. The principle of the invention is thus obviously not limited to three differently doped layers.

With the described method can by simple means during the laser welding of the layer 5 . 6 continuously - quasi online - the current melting depth can be precisely monitored and controlled with relatively little effort.

The invention has been described above with reference to preferred embodiments. In particular, here are examples of the different layers 3 . 4 . 5 . 6 called certain materials. However, the inventive principle is not limited to the use of said materials but extends to any combination of materials of the different layer layers, wherein the top layer layer 5 and or 6 of the carrier 2 which is to be melted, preferably formed from a metal or an alloy.

LIST OF REFERENCE NUMBERS

I

intensity

λ

detected wavelength

λ _{char, Me}

wavelength

λ _{char, Cu}

wavelength

λ _{char (3)}

wavelength

λ _{char (4)}

wavelength

W _{a, Me}

work function

W _{a, Cu}

work function

1

substratum

2

carrier

3

layer sheet

4

layer sheet

5

layer sheet

6

Copper layer region

6a

load connection

7

Area

7 '

Area

7 ''

Area

8th

laser beam

9

laser device

10

probing

11

reflected radiation

11 '

reflected radiation

11 ''

reflected radiation

12

evaluation

14

output

15

connecting cable

Claims (5)

Method for determining the melting depth during the melting of a metal layer ( 5 . 6 ) by means of a laser beam ( 8th ), in which - under the metal layer ( 5 ) one or more layer (s) ( 3 . 4 ) which, when exposed to a test radiation, emits / emits in each case a layer-specific, different discharge radiation of the metal layer characteristic discharge radiation, wherein the discharge radiation differs in that they have different characteristic maximum wavelengths (λ _Cu , λ _{char (4)} , λ _{char (3)} ); During the melting of the melting area ( 7 . 7 ' . 7 '' ) with a test radiation ( 10 ), and - from the reflected test radiation ( 11 ) the melting of the layer layer (s) ( 5 . 4 . 3 ) is detected by detecting the layer-specific characteristic discharge radiation (λ _{char, Cu} , λ _{char (4)} , λ _{char (3)} ) and from this the melting point is concluded, the metal layer ( 5 . 6 ) and the layer layer (s) ( 3 . 4 ) are produced from the same starting material and wherein the layer layer (s) ( 3 . 4 ) is doped differently in layers to produce different characteristic discharge radiation (λ _{char, Cu} , λ _{char (4)} λ _{char (3)} ). Method according to claim 1, in which - the reflected test radiation ( 11 ) by means of a spectrometer ( 12 ) is monitored. Method according to one of the preceding claims, in which - for melting the metal layer ( 5 . 6 ) and for generating the test radiation ( 10 ) the same laser device ( 9 ) is used. Substrate for use in a method according to any one of the preceding claims, wherein - on a support ( 2 ) several metallic layers ( 5 . 4 . 3 ) are applied from a base material, each layer ( 5 . 4 . 3 ) has an individual doping, which is selected such that the respective layer-specific work function to the uppermost layer ( 5 ) and that upon exposure to a test radiation ( 10 ) emits a layer-specific characteristic discharge radiation (λ _{char, Cu} , λ _{char (4)} , λ _{char (3)} ). A substrate according to claim 4, wherein - The base material is copper.

Images