Procedure and device for the inspection of a semiconductor chip before the assembly.

31-01-2013 дата публикации
Номер:
CH0000705370A1
Принадлежит: Kulicke & Soffa Die Bonding Gmbh
Контакты:
Номер заявки: 01-00-2011265
Дата заявки: 31-07-2011

[1]

Subject of the invention the available invention concerns the area of the automatic control engineering. It refers to a device, a unit and a procedure for manipulating chips, in particular semiconductor chips, in accordance with the generic term of the independent patent claims.

[2]

State of the art one the largest challenge in the semiconductor assembly is taking up, manipulating and processing very thin semiconductor chips.

[3]

The semiconductor chips are made available generally on a transparency (English, “carrier tape”), which contain generally a complete disk (English, “wafer”), which was cut into small chips; a procedure of which frequently as sort (English “dicing”) is designated. The semiconductor chips are therefore usually called discrete components (English “this”). The transparency is frequently the same foil, which carried the disk during the sort, and as sort foil (English “dicing tape”) is often designated. For discrete components nowadays thicknesses to down are usual to 25 over, whereby a continuous trend exists to a further reduction of the thicknesses. In the long-term planning of the semiconductor manufacturers already with thicknesses from CBM one proceeds.

[4]

Semiconductor assembly automats (English “the Bonders”) take up individual discrete component of the transparency, place this on a substrate or a further discrete component, where it is afterwards fastened. The individual components are in most cases fastened permanently; there are however also sample applications, where only a temporary attachment takes place. Frequently a temporary attachment is transformed in the following by an additional Heizoder pressing process into a permanent. Thin discrete components exhibit typically a laminating on the back of the disk (English “wafer backside lami nation”, WBL) or Übereiner wiring (English “film more over wire”, FOW), i.e. a detention film, which is applied on a unstruktrierten side of the disk or the discrete component. A procedure, which covers a diluting of the disks, an attachment of the detention film, an attachment on the transparency and a sort into individual chips, is usually called disk preparation. Laminating, which it permits due to its detention characteristics to fasten the discrete components can be planned either between the transparency and the half lyra disk, or on a surface, which away-points from the transparency.

[5]

Take up, placing and transport between a Aufnehmund Platzierposition can by means of individual chip manipulation unit in the chip assembly automat take place. Modern chip assembly automats however often cover a multiplicity at chip manipulation units: Generally taking up at a so-called disk table (English “wafer table”) takes place, supported from a zeroth chip manipulation unit, which is called ejector (English “the ejector”). The ejector facilitates the replacement of the discrete component of the transparency, for example by presses of the discrete component against a first chip manipulation unit, which is called Aufnehmeinheit. The Aufnehmeinheit takes up afterwards the discrete component in a Aufnehmschritt of the transparency and hands it to a second chip manipulation unit, which is called Platziereinheit.

[6]

The Platziereinheit places the discrete component on a goal position, where it is fastened. In some cases at least a third chip manipulation unit - a so-called data transfer unit - is intended, in order to hand the discrete component over of the Aufnehmeinheit to the Platziereinheit. An example is in WHERE 07,118,511 Al which hereby in full-extent by means of reference is included. In this way a discrete component on a substrate, for example a leader framework (English “lead frame”), a leader plate, a multi-layer plate etc., or on another discrete component, which was fastened on same meadow, can be fastened.

[7]

The production of very thin disks is very expensive compared with such with conventional thicknesses and without laminating. Sawing, taking up and manipulating very thin discrete components are subject to substantial yield losses. The largest part of the yield goes to the disk preparation, which lost a taking up of the discrete components or the following manipulation, which leads to damaged discrete components with typical damages during, jumped or splintered discrete components e.g. broken. Both placing and fastening take place comparatively reliably, and lead only to negligible losses.

[8]

In particular for stacked semiconductor assembly processes, with which two are fastened one on the other or more discrete components, an attachment of a damaged discrete component can have serious consequences: If such is fastened a damaged discrete component on a pile, this leads to the loss of all discrete components in the appropriate pile, already fastened, which is called package (English “package”). In an extreme situation it can come so for example to the loss a package from 15 stacked discrete components by fastening a damaged sechzehnten discrete component on this. On the basis of a yield of 99% for the Aufnehmschritt, one sinks packing yield which can be expected on 0.99A16, i.e. on 85% in this example, and on only 44% for a Aufnehmschritt with a yield of 95%.

[9]

Well-known inspection methods for detecting broken, jumped or splintered discrete components are accomplished before the Aufnehmschritt - which limited yield has - and are based generally in an illustration of a lit surface of the discrete component with following image processing. Such inspection methods have limited reliability due to small contrasts and widths of the saltuses, distortions in the discrete components and a confusableness between typical leap features and looking similar samples on the surface of the discrete components.

[10]

Until recently semiconductor assembly automats ordered no mechanisms or functionality for detecting leaps in the discrete components before the assembly and after the Aufnehmschritt. This is to be due to two reasons: First of all the typical “direct” Aufnehmund Platzierarchitektur does not permit use of semiconductor assembly automats more particularly (Aufnehmund placing) tools with integrated means for leap detection. Secondly leaps are difficult in the discrete components due to of them strongly varying manifestation both in the form and width to detect.

[11]

In the WO2011/018 375 Al, a procedure and a device were introduced for the inspection by chips before their assembly by means of an optical leap detection.

[12]

Short description of the invention it is therefore task of the invention, instead of - or additionally of an inspection before taking up a discrete component such after taking up, for example on an ejector, to permit, in order to guarantee that by a semiconductor assembly automat only intact discrete components on a substrate or a discrete component fastened before are fastened.

[13]

Furthermore the invention is to make an identification for damaged possible discrete components and permit it in such a way, to omit whose placement on a substrate or a pile of discrete components already fastened. Ideal way should be an inspection of a discrete component directly before the placement and following attachment on a substrate or a discrete component already fastened possible.

[14]

The aforementioned tasks are solved from a chip manipulation automat and a procedure to manipulating chips in accordance with the independent patent claims.

[15]

With a chip manipulation device and a procedure for manipulating chips in accordance with the available invention acoustic oscillations and/or ultrasonic vibrations in a discrete component become lively by means of an actuator, in particular a piezoelectric actuator. In a preferential arrangement of the procedure and the device the discrete component on the actuator is centered and fastened temporarily by means of negative pressure. While a suggestion frequency, i.e. an oscillating frequency of the actuator are varied, self-fashions of the discrete component become lively. Natural frequencies, which belong to the self-fashions, can be characterized and lain by means of a fashion number, dependent on a size of, thickness, etc. of the discrete component, usually in a range between 100 cycles per second and unite 10 kHz.

[16]

In accordance with an exemplary arrangement of the invention a chip manipulation device, in particular an assembly automat for discrete components to place to transport presented, which a supply station for chips to take up an assembly station for chips enclosure and one or more chip manipulation units, which are laid out in addition, a chip of the supply station to the assembly station and at an assembly position, whereby each chip manipulation unit is laid out in addition, the chip temporarily in a defined situation regarding the chip manipulation unit and whereby the chip manipulation device furthermore means cover for suggesting from oscillations, in particular acoustic oscillations and/or ultrasonic vibrations in the chip, while this by one chip manipulation units is held, as well as means to measuring the oscillations lively in the chip.

[17]

In a further preferential arrangement of the available invention a chip manipulation unit, in particular for use in a semiconductor chip assembly automat, presented, which is laid out in addition, a chip, in particular a semiconductor discrete component receiving and to a transfer point passing on at a Ubernahmeposition, and, in order to keep the chip temporary in a defined situation regarding the chip manipulation unit, as well as means covers means for suggesting oscillations, in particular acoustic oscillations and/or ultrasonic vibrations in the chip.

[18]

In a further preferential arrangement of the available invention a chip manipulation procedure with the following steps is presented:

[19]

a) Receive a chip (5), in particular a semiconductor discrete component, at an assumption position by means of a chip manipulation unit, b) temporary holding of the chip (5!. in a defined situation regarding the chip manipulation unit to hold, transport c) the chip (5) to a Ubergabeposition and a d) following releasing of the chip, whereby between step a) and step d) oscillations in the chip (5) becomes lively, and which are measured in the chip lively oscillations.

[20]

These and further tasks, advantages and characteristics of the invention become obvious from the following detailed description of preferential remark examples in connection with the designs.

[21]

Short description of the designs the invention is to be understood best on the basis the following detailed description in connection with the attached designs. Usual practice following the individual elements of the designs are not true to scale shown. Rather the dimensions of individual elements are arbitrarily increased or reduced for the sake of the clarity. The designs cover the following illustrations:

[22]

Fig. 1 fig. 2 fig. 3 fig. 4 fig.

[23]

shows a semiconductor assembly automat in accordance with an exemplary arrangement of the available invention.

[24]

a schematic representation of a preferential arrangement of a chip manipulation tool points to the use with the available invention.

[25]

a schematic representation of a further preferential arrangement of a chip manipulation tool points to the use with the available invention.

[26]

a schematic representation of a further preferential arrangement of a chip manipulation tool points to the use with the available invention.

[27]

a flow chart of a chip manipulation procedure shows in accordance with an exemplary arrangement of the available invention.

[28]

Detailed description of the invention fig. 1 shows a semiconductor assembly automat in accordance with an exemplary arrangement of the available invention. The semiconductor assembly automat covers a multiplicity of chip manipulation units: a zeroth chip manipulation unit, which ejector 91 one calls, loosens a discrete component 5, which with a WBLoder FOW Lamination can be provided, from a transparency 59 on a disk table off. Disk table and ejectors represent a chip supply station for the semiconductor assembly automat. As soon as the discrete component 5 is replaced, it is received by a first chip manipulation unit, which Aufnehmeinheit 92 one calls. Furthermore the semiconductor assembly automat covers a second chip manipulation unit, which Platziereinheit 94 one calls, and for transporting and placing the discrete component serves 5 to a goal position on a substrate 6 at a chip assembly station. The Aufnehmeinheit 92 is around a first axle swivelling and laid out to hand over discrete component the 5 to a third chip manipulation unit. Even if third chip manipulation unit, as data transfer unit 93 designation, meant is swivelling around a second axle and can in such a way discrete component 5 to the Platziereinheit 94 over hand. The data transfer unit 93 covers a chip manipulation tool 10, which covers at least one vacuum opening for temporary fastening of the discrete component 5. In addition the data transfer unit 93 covers a piezoelectric actuator, to induce which is designed for it, acoustic oscillations and/or ultrasonic vibrations in the discrete component 5 fastened to the chip manipulation tool 10. If the data transfer unit 93 at a transfer point is, as on the basis the pulled through line in fig. 1 represented, by the piezoelectric actuator acoustic oscillations in the discrete component 5 are induced. For the measurement of the oscillations induced in the discrete component 5 a laser-based rangefinder 7 is intended.

[29]

The semiconductor assembly automat covers besides a Steuerungsoder not shown in fig. 1 monitoring system as a check of the movements of the different chip manipulation units etc.

[30]

Fig. 2 points a schematic representation of a preferential arrangement of a chip manipulation tool to the use with the available invention.

[31]

The chip manipulation tool 10 covers a shank 103, by means of which it can be installed at a chip manipulation unit, in particular the data transfer unit 93. The piezoelectric actuator 101 is intended on the shank 103. From a flexible material, e.g. rubber, existing hose 102 is installed on the piezoelectric actuator 101. The hose has a vacuum connection 1021 for the connection to a source of negative pressure, in order to suck in the discrete component 5 temporarily against a vacuum opening formed at the upper end of the hose.

[32]

Fig. 3 points a schematic representation of a further preferential arrangement of a chip manipulation tool 10 " to the use with the available invention.

[33]

The chip manipulation tool 10 " covers again a shank 103, by means of which it can be installed for example at a chip manipulation unit, in particular the data transfer unit 93. The piezoelectric actuator 101 is again intended on the shank 103. With this execution form the hose is installed directly on the shank 103, the discrete component 5 can again against a vacuum opening formed at the upper end of the hose 102 be sucked in, and in position be kept so temporary. Dimensions of hose 102 and piezoelectric actuator 101 are in such a way selected that a small air gap between a discrete component sucked in at the upper end of the hose 102 and the piezoelectric actuator 101 results, which preferably exhibit a length between 25 and 500 IJm, verzugsweise 50 to 200 IJm.

[34]

Fig. 4 points a schematic representation of a further preferential arrangement of a chip manipulation tool 10 w to the use with the available invention. The chip manipulation tool 10 covers again a shank 103, by means of which it can be installed for example at a chip manipulation unit, in particular the data transfer unit 93. On the shank 103 a hollow piezoelectric actuator 101 is intended ", so that a vacuum opening at one of the shank 103 removed end of the hollow piezoelectric actuator 101 " is formed.

[35]

In the shank a Unterdruck-Zuführkana11031 is intended, in order to be able to supply to an inside of the hollow piezoelectric actuator 101 " a negative pressure. Preferably a layer from soft material, for example rubber, is to prevent at the end of the hollow removed from the shank 103 piezoelectric actuator 101 " over that the discrete component 5 with a comparatively hard material piezoelectric actuator 101 " comes into contact.

[36]

As previously mentioned, a laser-based rangefinder 7 can be used, in order to measure in the discrete component induced oscillations, for example by means of measurement of misalignments and/or bucklings of the discrete component 5, e.g. as function of the time. The misalignments and/or bucklings can be based on only one place, for example at only one corner of the discrete component 5. Preferably they are based on two or several places. In this way it is guaranteed that damages can be recognized independently of their position, and permitted besides to receive a reference where the damage could be. This is again helpful as guideline assistance for additional inspection methods for discrete components, which could be used.

[37]

Favourably also optical measuring procedures, which are based on Interferometrie, can be used, in order to measure misalignments and/or bucklings of the discrete component.

[38]

On the basis an entrance at the piezoelectric actuator 101 and a measurement of misalignments and/or bucklings, one or more frequency answer functions of the discrete component 5 are computed. These contain Amplitudenund concerning phase information a reaction of the discrete component 5 to suggestions for vibration. Generally the frequency answer functions exhibit various resonance points, which natural self-fashions of the discrete component 5 show in the available, special experimental structure.

[39]

If the discrete component 5 is damaged, if leaps, column, flakings or other defects are e.g. present, shows the frequency answer functions deviations in relation to the frequency answer functions of an intact, but otherwise identical discrete component 5th in a damaged discrete component 5 is mechanical rigidity and fashion form. Damaged discrete components 5 can be identified from there through to comparisons of a measured frequency answer function with a reference frequency answer function.

[40]

This is reached preferably through comparisons by frequency answer functions by means of a continuous frequency spectrum, preferably in a subrange between 1 cycles per second and 100 kHz. Alternatively this can be also only achieved by comparison of one or several individual resonances with that good discrete components.

[41]

Preferably a suggestion for sine wave with slowly sliding suggestion frequency is used. However preferably also a simultaneous suggestion could be used over a range by frequencies to make in particular in order a faster determination possible of the frequency answer function. In particular a suggestion would be conceivable by means of white noise, in order to at one time suggest far spectrum of frequencies.

[42]

Instead of measuring in the discrete component 5 induced vibrations based on measurements of misalignments and/or bucklings with optical means, can be e.g. used preferably also an acoustic receiver and/or an ultrasonic receiver, a microphone or a further piezoelectric element. In this way the vibration frequency can be determined directly without necessity for a measurement of misalignments and/or bucklings.

[43]

Preferably the oscillations induced in the discrete component 5 can be determined also by determination of an impedance or an impedance spectrum of the piezoelectric element.

[44]

Based on a distinction between damaged and intact discrete components an error handling control can permit it to 5 then to prevent a fastening of defective discrete components 5 to remove and these from the data transfer unit 93.

[45]

Fig. a flow chart shows 5 in accordance with certain exemplary arrangements the available invention. As evident to the specialist, certain steps in the flow chart can be added be omitted, additional steps, and be amended the sequence of the steps opposite the shown.

[46]

In the past description the discrete component 5 the data transfer unit 93 relative to by means of negative pressure in a defined position was held. This implied that the discrete component 5 is sucked against any kind bearing surface, in which one or more vacuum openings it is formed, or which is defined by one or more vacuum openings. Preferably the discrete component 5 can be held for repulsive forces also by contactlessly in position, for example by a combination of attraction, e.g. electrostatic, and, e.g.

[47]

Compressed air, to which the discrete component can be exposed in the proximity of a work surface of an appropriately arranged chip manipulation tool.

[48]

Even if the invention managing with reference to specific execution forms descriptive and illustrated is, this is not limited to these execution forms. Rather different modifications can be made by details within Schutzund of equivalence range of the patent claims, without from it a deviation from the invention results.

[49]

In particular the invention, although managing regarding semiconductor discrete components described, for each kind chip can be used, including each essentially flat panel from one or more materials, which is taken up by any kind of supply carrier, in particular a foil or a volume, and is placed at a goal position, in particular on a chip, a substrate, a foil, a volume or a storage element.



[50]

A chip handling apparatus, unit and method is presented. The chip handling apparatus comprises a chip supply station; a chip mounting station; and one or more chip handling units configured to pick a chip from the supply station, transport the chip to the mounting station, and place the chip at a mounting location; wherein each chip handling unit is configured to temporarily retain the chip in a defined position relative to the chip handling unit. The chip handling apparatus further comprises means for inducing sonic vibrations in the chip when retained by one of the chip handling units; and means for measuring the vibrations induced in the chip.



1. A chip handling apparatus, in particular a die bonder, comprising

a) a chip supply station,

b) a chip mounting station, and

c) one or more chip handling units configured to pick a chip (5) from the supply station, transport the chip to the mounting station, and place the chip (5) at a mounting location,

d) each chip handling unit configured to temporarily retain the chip (5) in a defined position relative to said chip handling unit, characterized in that

e) the chip handling apparatus further comprises

i. means for inducing vibrations, in particular sonic and/or ultrasonic vibrations, in the chip (5) when retained by one of the chip handling units

ii. means for measuring the vibrations induced in the chip (5).

2. The chip handling apparatus according to claim 1, further comprising discrimination means configured to determine whether the chip (5) is damaged based upon a measurement of the vibrations induced in the chip (5).

3. The chip handling apparatus according to one of the previous claims, characterized in that the apparatus comprises a first chip handling unit for picking the chip (5) from the supply station, a second chip handling unit (94) for placing the chip (5) onto the mounting location, and wherein the first chip handling unit is configured to hand the chip (5) to the second chip handling unit at a first hand over location.

4. The chip handling apparatus according to claim 1 or 2, characterized in that the apparatus comprises a first chip handling unit for picking the chip (5) from the supply station, a second chip handling unit for placing the chip (5) onto the mounting location, and a third chip handling unit configured to receive the chip from the first chip handling unit, and hand it to the second chip handling unit.

5. The chip handling apparatus according to claim 3 or 4, characterized in that the means for inducing vibrations in the chip (5) are configured to induce vibrations when the chip is retained by the second chip handling unit.

6. A chip handling unit, in particular for use in a semiconductor die bonder, said chip handling unit being configured to receive a chip (5), in particular a semiconductor die, at a takeover location and to hand over said chip (5) to a delivery location, said chip handling unit comprising

a) means for temporarily retaining the chip (5) in a defined position relative to the chip handling unit,

b) means for inducing vibrations, in particular sonic and/or ultrasonic vibrations, in the chip (5).

7. A chip handling process, in particular for a die bonding process, comprising the steps of

a) receiving a chip (5), in particular a semiconductor die, at a takeover location by means of a chip handling unit,

b) temporarily retaining the chip (5) in a defined position relative to the chip handling unit,

c) transporting the chip (5) to a delivery location and

d) subsequently releasing the chip (5), characterized in that

e) between steps a) and d), vibrations are induced in the chip (5), and

f) the vibrations induced in the chip (5) are measured.

8. The chip handling process according to claim 7, characterized in that in step e), sonic and/or ultrasonic vibrations are induced in the chip (5).

9. The chip handling process according to claim 7 or 8, characterized in that at the delivery location, the chip (5) is placed onto a substrate or onto another chip.

10. The chip handling process according to claim 7, 8 or 9, characterized in that at the delivery location, the chip (5) is handed to a further chip handling unit, in particular a place unit (94).

11. A chip handling process, in particular a die bonding process, comprising the steps of

a) picking a chip (5), in particular a semiconductor die, from a supply station,

b) transporting the chip (5) to a mounting station by means of at least one chip handling unit, said chip handling unit being configured to temporarily retain the chip (5) in a defined position relative to the chip handling unit

c) placing the chip (5) onto a mounting location,

characterized in that

d) in step b), vibrations are induced in the chip (5), and

e) the vibrations induced in the chip (5) are measured.

12. The chip handling process according to claim 11, characterized in that in step d), sonic and/or ultrasonic vibrations are induced in the chip (5).

13. The chip handling process according to one of claims 7 through 12, further comprising the step of determining whether the chip (5) is damaged based upon an analysis of the vibrations measured.