Bonding apparatus, bonding method, and method for manufacturing semiconductor device
An apparatus includes a first and second stages. The first and second stages respectively hold a first and second substrates. The second stage being opposed to the first stage. A stress application portion applies a stress to the first substrate based on a first magnification value. A calculator calculates the first magnification value based on a flatness of the first substrate and a first equation. The first equation represents a relation between flatness of a third substrate, a second magnification value, and an amount of pattern misalignment between the third substrate and a fourth substrate bonded to the third substrate. A controller controls the stress application portion to apply a stress to the first substrate on the first stage based on the first magnification value while the first and second substrates are bonded to each other.
1. A bonding apparatus comprising:
a first stage configured to hold a first substrate;
a second stage configured to hold a second substrate, the second stage being opposed to the first stage;
a stress application portion configured to apply a stress to the first substrate based on a magnification value while the first and second substrates are bonded to each other;
a calculator configured to calculate the magnification value based on a flatness of the first substrate and a first equation, the first equation representing a relation between a flatness of a first learning substrate, the magnification value, and an amount of pattern misalignment between the first learning substrate and a third substrate bonded to the first learning substrate; and
a controller configured to control the stress application portion;
wherein the calculator calculates, for each flatness C of a plurality of the first learning substrate, a coefficient k from an equation | by using the magnification value S and an amount of pattern misalignment OL between the first learning substrate and the third substrate, where the equation 1 is OL=S+k×C, and
further calculates the first equation by substituting the calculated coefficient k into the equation 1.
2. The device of claim 1 , wherein
a plurality of the first equation is calculated for each flatness of a plurality of the first learning substrate, and
the calculator selects one of the first equation in accordance with the flatness of the first substrate and calculates the magnification value.
3. The device of claim 1 , wherein the first equation is a relational expression between the magnification value and the flatness that the amount of pattern misalignment to be zero.
4. The device of claim 2 , wherein the first equation is a relational expression between the magnification value and the flatness that the amount of pattern misalignment to be zero.
5. The device of claim 1 , further comprising:
a flatness measurement portion configured to measure flatness of the first learning substrate or the first substrate on the first stage;
a pattern-misalignment-amount measurement portion configured to measure the amounts of pattern misalignment by using a plurality of the first learning substrate and a plurality of the magnification values while each of the first learning substrate and the third substrate are bonded to each other; and
a storage configured to store the first equation.
6. The device of claim 1 , wherein the calculator performs spherical approximation for the flatness of the first learning substrate or the first substrate to calculate the first equation or the magnification value.