Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
~2S~39
6082-188D
This application is a division of our Canadian
patent application Serial No. 467,015 filed November 5, 1984.
This invention relates to apparatus for calibrating
a capacitance height gauge applied in a reticle position
detection system for electron beam lithography apparatus and
to calibration apparatus for capacitance height gauges.
There is a great need for high accuracy height
measuring gauges for a wide variety of applications. One such
application is in determining the position of a reticle in an
electron beam lithography apparatus.
Electron beam lithography is rapidly becoming the
method of choice for exposing ultrahigh accuracy reticles in
the production of very large scale integrated circuits. An
electron beam lithography apparatus typically includes an elec-
tron beam optics housing positioned over a vacuum chamber in
which a reticle is installed. The reticle is a glass plate
covered by a layer of chromium, with a layer of electron beam
resist deposited over the layer of chromium. The reticle is
mounted on a stage which moves the reticle in the X and Y
~0 directions under the control of the control system while the
electron beam writes on, or exposes, the beam resist layer to
produce the desired circuit pa-ttern on the reticle. The con-
trol system not only moves the reticle to the desired X, Y
coordinate for the stage position being exposed, bu-t also,
controls the beam deflection angle to control the point on the
reticle at which the electron beam strikes.
In the past one of the more difficult problems in
~;Z5~35~
6082-188D
this art has been to determine the precise position of the
reticle with respect to the electron beam optics. The precise
position of the reticle must be determined in order to proper-
ly deflect the electron beam in order to accurately write on
the reticle. It is extremely important that any system used
to determine the position of the reticle be vacuum compatible,
compact, and noncontacting.
In addition, accurate methods for calibrating
capacitive height gauges are needed. In the past, capacitance
gauges have been calibrated by means of mechanically measured
distances, such as by means of micrometer measurements.
The invention described and claimed in the parent
application No. 467tOl5 comprises a highly accurate capa-
citance height gauge which is applied in the presently pre-
ferred embodiment in a reticle position detection system for
an electron beam lithography apparatus.
The capacitive height gauge circuitry includes a
hybrid circuit substrate which carries four measuring capacitor
circuits and two reference capacitor circuits. The driven
plates of the four measuring capacitors are disposed on the
bottom of the substrate opposite to the object surface to be
~s~
1 measured. The driven plates of the two reference capacitors
2 are disposed on the top surface of the substrate under caps
3 which support oppositely disposed grounded capacitor plates at
4 a nominal distance from the driven plates. The object surface
comprises the grounded plate for the four measuring
6 capacitors. The reference capacitors provide voltage
7 regulation to ensure that a stable signal drives the measuring
8 capacitors. The reference capacitors also set zero points for
9 1 the measurements from the four measuring capacitors. These
10 ¦ zero points are set at the nominal distance from the measuring
11 ¦ capacitor sensors. The four measuring capacitors provide
12 ¦ readings with respect to the object surface to determir.e the
13 position of the object surface with respect to the nominal zero
14 points. This information can be used by a suitable control
system to manipulate the object surface or tools used to work
16 on the surface in response to the position information as
17 appropriate to the task being performed.
18
19 In the presently preferred embodiment, the capacitance
height gauge is employed as an integral part of a reticle
21 position detection system for an electron beam lithography
22 apparatus. The "normal angle" of deflection of the electron
23 beam is first calibrated with respect to a calibration plate.
24 The four measurement sensors of the capacitance gauge are then
utilized to detect the position of the calibration plate and to
26 input this position information into the electron beam control
27 system to define the plane which the calibration plate lies in
28 which is denoted as the calibrated plane" for the reticle.
2g ~he reticle is then moved under the electron beam optics.
Rea~1~ngs ~re ta~en from the four measuring sensors of the
31
32 - 3 -
~2S~3~
6082-188D
capacitance gauge to detect -the position of the reticle. This
reticle position information is then also input to the control
system. The control system determines the devlation, if any,
of the reticle from the calibrated plane and appropriately
adjust the deflection angle of the electron beam in response
to the detected deviation to write at the desired point of
the reticle with a very high degree of accuracy.
In addition, a calibration apparatus for capacitance
height gauge is disclosed which is far superior to the present
mechanical measurement methods.
Thus, in accordance with a broad aspect of the in-
vention, there is provided an apparatus for calibrating a
capacitance height gauge having a capacitive sensor, comprising:
a base; a laser interferometer assembly supported by said base;
an electrically conductive mirror movably supported on said base
opposite to said laser assembly; a means for supporting said
capacitance gauge with said sensor oppositely disposed with
respect to said mirror; and a means for moving said mirror with
respect to said laser interferometer, wherein said moving means
~0 positions said mirror a desired location from said sensor such
that said sensor forms a capacitor with said mirror, and wherein
said laser interferometer assembly is utilized to determine
the distance between said sensor and said mirror, further com-
prising a means for obtaining an electrical reading from said
capacitor comprised of said sensor and said mirror while said is
positioned at said desired location.
Having described the invention in its presently
preferred embodiment in brief overview, the advantages, features
-- 4 --
~S~ 39
6082-188D
and novel aspects of the invention will become apparent from the
more detailed description of the invention which follows taken
in conjunction with the accompanying figures, in which:
Figure lA shows a schematic overview of the capa-
eitance height gauge.
Figure lB shows a bottom view of the hybrid circuit
substrate.
Figure lC shows an elevational view of the hybrid
cireuit substrate of the eapacitance height gauge overlying the
object surface.
- 4a -
~Sq3~l3~ ~
1 ¦ Figure 2 shows the oscillator circuitry and other
2 ¦ circuitry driving the reference capacitor CR f 1 including
3 ¦ the reference capacitor feedback loops which stabilize the
4 ¦ driving signal of the gauge.
5 I
6 1 Figure 3 shows an elevational view of the linearization of
71 the electric field lines of the driven plate of reference
81 capacitor CRef 1 under the influence of the guard ring.
91 `'
10l Figure 4 shows a simplified circuit diagram of the
11¦ measuring capacitor circuitry and measurement signal generation
3 circuitry for the ~X sensor.
14 ¦ Figure 5 is a schematic diagram showing the environment of
15 ¦ the capacitance height gauge as applied in a reticle position
16 ¦detection system for an electron beam lithog~aphy apparatus.
17 l
18 ¦ Figure 6 shows the calibration of the electron beam using
19 la calibration plate.
20 ~ .. .
21 ¦ Figure 7 shows the adjustment made by the control system
22 ¦ to the deflection angle where the reticle is positioned above
23 ¦ the calibrated plane.
24 1
25 ¦ Figure 8 shows a top view of the hybrid circuit substrate
26¦ of the capacitance height gauge.
271
28 Figure 9 is a simpli~ied circuit diagram of the
29 temperature control circuitry of the present invention.
31
32 - 5 -
:~5~
6082-188D
Figure 10 shows an elevational view of the capa-
citance height gauge calibration fixture.
Figure 11 shows a plan view of the fixture.
Figure 12 shows a perspective view of the inter-
ferometer cube support member, wi-th the capacitance height gauge
substrate secured to i-ts forward face.
Figure 13 shows a partially schematic, functional
block diagram of the calibration system.
Figure 14 shows a functional block diagram of the
calibration fixture electronics.
Figure 15 is a schematic diagram showing the test
set-up for the capacitance height gauge.
DESCRIPTION OF THE PRESENTLY PREFERRED EMBODIMENT
OF THE INVENTION
Figure lA shows a schematic diagram of the capa-
citance height gauge circuitry.
The signal source circuitry 10 includes an oscillator
15 which provides a sinusoidal signal. The sinusoidal signal is
split between a pair of voltage regulating circuits 20, 24.
~0 The upper signal goes through a phase 1 transformer 30 while
the lower signal goes through a phase 2 transformer 32 to
generate 180 out of phase signals. Both of these driving
signals are fed to hybrid circuitry carried on a substrate
(later described) which overlies the object surface being
measured. The hybrid circuitry is comprised of reference
r ~25V~39 ~-
1 capacitor circuitry 40, 42 for CRef 1 and CRef 2~ and
2 measuring capacitor circui try 50, 52, 54, 56 for the C~x,
3 C+y~ C X and C_y measuring capacitors. The hybrid
4 eircuit substrate 110 is shown in Figs. lB, lC and 8. Figs. lB
and lC shos~ that only the ~X, +Y, -X and -Y sensors 112, 114,
6 116, 118 and their associated guard rings 120, 122, 124, 126
7 are disposed on the bottom of the substrate opposite to the
8 object surface to be measured. The remainder of the bottom
9 surface is ground plane. Fig. lC shows that both reference
Ref. 1 and CRef 2 are disposed on the top of
11 Ref, 1 and CRef 2 are comprised of driven
12 plates 130 and oppositely disposed ground plates 134, whieh are
13 disposed on the internal surfaee of the eaps 140 at a fixed
14 nominal distanee from driven plates 130. See Fig. 3. The
1~ measuring capacitors are eomprised of ~X, ~Y, -X, -Y driven
16 sensor plates 112, 114, 116, 118 and a grounded plate eomprised
17 of the objeet surfaee 145 to be measured. See Fig. lC.
18 CRef 1 and CRef 2 inelude feedback loops 60, 62 to
19 regulate the voltage of the oscillator to provide a stable
signal to the measuring eapaeitors 50, 52, 54, 56. See
21 Ref. 1 and CRef 2 circuits 40, 42 also
22 assist in providing nominal zero points for each of the
23 measuring eapaeitors as will later be explained more fully.
24 The outputs of ~neasuring eapacitor eireuits 50, 52, 54, 56 are
eoupled to assoeiated measurement signal generating
26 eireuitry 70, 72, 74, 76 to provide signals to the control
27 system 80 indicating the position of the objeet surfaee 145
28 with respeet to thefour sensors 112, 114, 116, 118.
29
31
32 - 7 -
~25~)~3~ ('
1 The C+x and C+y circuits 50, 52 and C_x and C y
2 circuits 54, 56 are driven by 1~0 out of phase signals to
3 ensure that there is no net current flow through the object
4 surface 145 to ground. This feature may be important where
object surface 145 is a reticle in an electron beam lithography
6 apparatus as will be explained later on.
8 Having described the capacitive heiyht gauge in brief
9 overviëw, reference is now made to Fig. 2 which shows the
signal source circuitry 10 and the circuitry 40 associated with
CRef 1 in greater detail. This circuitry provldes a clean,
12 symmetric, sinusoidal driving signal to the measuring
13 capacitors as will now be explained.
14
With reference to Fig. 2, a TTL programmable
16 oscillat~r 150 generates a high frequency square wave signal
17 which is converted by the tuned circuit 152 into a sinusoidal
18 si~nal. The signal undergoes a unity voltage gain, current
19 amplification at amplifier 154 and is split at node 156 between
2210 the two symmetrically arranged circuits shown in the upper and
lower half of the figure. To simplify the description of the
22 circuitry, only the upper portion of the circuitry will be
23 described initially. The signal flows upwardly through R
24 and through the FET 1 to ground. FET 1 together with Rl
comprise a voltage divider. The resistance of FET 1 varies
27 with its gate signal as will later be explained more fully.
From node 158 the signal goes through a voltage amplification
28 at amplifier 160 and another unity voltage gain, current
29 amplification at amplifier 162. ~he signal is then input over
a transmission line 1~4 to tl~e prima~y coiL of a step-up
31
32 - 8 -
,~ lZ5~3~ (
transformer 170 which is tuned to the oscillator frequency by a
2 ¦ variable capacitor 176 connected in parallel with the secondary
3 coil 174. The transformer 170 acts as a band pass filter and a
4 low distortion, geo~etrically symmetric, sinusoidal signal at
the oscillator frequency is thereby produced and transmitted
6 over a transmission line 180 to the hybrid substrate
circuitry 40 associated with reference capacitor CR f 1 and
8 the measuring circuitry 50, 52 associated with measuring
9 capacitors C+x, C+y~ respectively. The circuit 40 for
CRef 1 will be described first. Ignoring diodes D3 and
11 D4 and the associated guard ring 136 for the moment, the
12 signal is split at node 190 and passes through DC blocking
13 capacitors Cl and C2. Diodes Dl and D2 are connected
14 between capacitors Cl and C2 in a half wave rectifier
arrangement- Reference capacitor CRef 1 is connected
16 between diodes Dl and D2. The plates of CRef 1 are
17 separated by a fixed distance which is equal to the "nominal
18 distance". The nominal distance is the desired distance
19 between sensor plates 112, 114, 116, 118 of the measuring
capacitors and object surface 145. The driven plate 130 of
21 CR f 1 is supported on the upper surface of substrate 110
22 while grounded plate 134 is supported on the interior surface
23 of ceramic cap 140 which is secured to substrate 110 over
2~ driven plate 130 as shown in Fig. 3. Inductors Ll and L2
are connected between Cl and Dl, and C2 and D2,
26 respectively, as shown. Filter capacitors C3 and C4 are
27 connected between Ll and L2, respectively, and ground. The
28 signal from coil Ll leaves t~e hybrid circuit via a shielded
29 coaxial cable 194 and is grounded through a coil L3. The
signal from coil L2 leaves the hybrid circuit via a shielded
31
32 9 _
~- 125~)13~ (
2 coa~ial cable 196 and is connected through a coil L4 to the
l inverting input of transresistance amplifier 200.
31 Transresistance amplifier 200 has a grounded noninverting input
41 and a feedback resistor R3 as shown.
51
61 Diodes Dl and D2 are connected in a half wave
8 ¦rectifier arrangement, as noted, so that one diode is "off"
9 Iwhile the other diode is "on" and vice versa. As the
Isinusoi~dal input signal increases from its maximum negative
10 ¦value to its maximum positive value, Dl is forward biased and
2 ¦conducts to charge capacitor CRef 1~ Conversely, as the
signal falls from its maximum positive value to its maximum
13 negative value, D2 is forward biased to drain the charge on
capacitor CRef 1' We use the current produced in L2 by
the discharging of capacitor CRef 1 to regulate and
16 stabilize the signal supplied to the measuring capacitors C~x
17 and C+y as will become apparent. We have grounded the signal
18 passing through Ll in that for the purpose of voltage
regulation we need only to monitor the L2 current.
21 As CRef 1 discharges through D2, C2 blocks all DC
23 current and passes the AC component so that a very small DC
l current passes through coil L2. L2 is a large coil on the
225 order of 470 microhenries to block the AC component of the
current. ~ny small remaining AC component of the current is
27 filtered out by filter capacitor C4. The small DC current
passes from coil L2 through coil L4 and is converted by
28 transresistance ampl1fier 20~ into a voltage representative of
29 the current throu~h L2- The current through L2 which is
30 input to transresistance amplifier 200 is multiplied by the
31
32
- 10 -
~S~ 39 ('
1 value of resistance R3 to generate the output volt~ge. This
2 1 output ~oltage is then imposed across resistor R4 and
3 produces a current IR flowing to the right out of node 204.
4 1 A ~7 volt reference voltage is imposed on the resistor R5 to
51 produce a current flowing left to right into node 204 in
61 Figure 2.
71
81 It has been previously mentioned that the gate signal to
9 ¦ FET l controls the drain to source resistance of the FET. If
10¦ we assume as an initial condition that just before start-up the
11 ¦ circuit power is "on" but the oscillator is "off", the 7 volt
12 ¦ reference voltage will be applied across R5 with zero voltage
13 ¦ being applied across R4. Consequently, a maximum current
14 will be applied through R5 to the inverting input of
differential amplifier 210 and differential amplifier 210 will
16 have its maximum negative output of -15 volts, for example,
17 which will shut FET l "off" and present a maximum resistance to
18 ground. Consequently, the current flow through, and voltage
19 drop across, Rl will be the minimum and the voltage at
node 158 will be a maximum~ If we assume that R5 = 3SK ohms
21 and that ~Ref l = 7 volts, then IR = 200 microamperes flowing
22 through R5 into the inverting pin just before start-up.
23
24 Just after start-up this maximum voltage at node 156 will
produce maximum current flow through L2 and to the right from
226 node 204 through R4. The current flow through R4 will
initially exceed the 200 microampere current through R5 which
28 will make the output of di~ferential amplifier 210 become less
29 nega~ive and rise to -lO volts, for example. The resistance of
FET ~ will accordingly drop proportionately reducing the
31
32
~5~3
1~ ~
1 ¦ voltage at node 158 and therefore the current through L2 and
2 ¦ R4. This cycle will repeat itself until the current flowing
3 ¦ from node 2~4 through R4 equals the 200 microampere current
41 flowing through R5 into node 210. At this point zero current
51 will flow into the inverting input of differential
6 ¦ amplifier 210 and the system will consequently lock.
7 l
8 ~ We know IR is constant at 200 microamperes as calculated
10 ¦ above. To make IR equal to 200 microamperes, if we assume R4 =
11 ¦ 20K ohms, a voltage of V = I R = (-200 microamperes)
12 ¦ (20K ohms) = -4 volts must be imposed across R4 to lock the
13 system. Consequently, the output of transresistance
14 amplifier 200 must be -4 volts at this point. We have already
noted that the output of transresistance amplifier 200 is the
16 product o~ R3 and the input current I . Consequently, I
17 VTRlR3, and if we assume that R3 = 50K ohms, then IL
18 -4 volts/50K ohms = -80 ~icroamperes. Therefore the system
19 stabilizes when the current through L2 = -80 microa~peres for
the illustrative parameters used. Note that this
21 -80 microampere current through L2 corresponds to the
2~ capacitance of CRef l which has a plate separation of the
23 nominal distance. The current through L2 is denoted the
24 reference capacitor signal for CRef l- The CR f
circuit 40 together with the signal source circuit lO, thus,
26 together stabilize the driving signal supplied to the C+x and
27 C+y measuring capacitor circuits 50, 52. If the capacitance
28 of CRef l varies, the voltage of th~ driving signal will
29 correspondingly vary; however, the current flowing thro~lgh L2
will remain stable and unchan~ed.
31
32 - 12 -
~5~ 3~ ~_
1 Note that the DC loop for the current through L2,
2 comprising the reference capacitor signal, flows from ground
3 g Ref. 1' D2~ L2~ L4~ R3 and into the output
4 of transresistance amplifier 200 which has a very low
5 impedance, and therefore, functions as a ground to complete the
6 loop. The composition of this loop is important as will later
7 become apparent.
9 Before describing the circuits for the meas~lring
10 capacitor C+x and C+y~ mention should be made of the guard
11 ring circuitry of the circuit which has to this point been
12 ignored. The guard ring is simply an annular capacitor
14 plate 136 insulated from the driven plate 130 of CRef 1 and
disposed between the driven plate 130 and the substrate 110 as
15 best shown in Figure 3. The guard ring 136 is disposed
16 opposite to the grounded plate 134 and comprises the driven
17 capacitor plate of the capacitor, Cguard 1 Guard ring 136
18 encircles and overlaps the driven plate 130 of C~e~ 1 As
shown in Figure 3, guard ring 136 linearizes the electric field
20 lines at the edges of driven plate 130 of CR f 1 to
21 eliminate the fringe effects of the field lines which would
22 otherwise occur at the plate edges. By linearizing the field
~3 lines of CRef 1' guard ring 136 equalizes CRef 1 with the
24 measuring capacitors C+x and C+y as will be better
25 explained later on.
26
27 Another important function of the guard ring 136 is to
28 shield CRef 1 from stray capacitances which would otherwise
329 be present betwe~n the dri~en plate of CRef 1 and other
31
3~ _ 13 -
~ZS~3~
I ch rged or cc~nd~lctive surfaces in tl~e vicinity. These stray
2 capacitances could add to the current flowing from CR f
3 and lead to measurement inaccuracies if not shielded.
Guard ring 136 is driven through half-wave rectifying
6 diodes D3, D4 in the same way as driven plate 130 is driven
7 through diodes Dl, D2. D3, D4 are identical to and
8 provide the same voltage drops as Dl, D2, and accordingly,
9 the potential of the guard ring 136 at all times matches the
10 ¦ potential of driven plate 130. By maintaining this voltage
match, current flow between guard ring 136 and driven plate 130
12 is prevented and the field lines at the edge of the driven
13 plate 130 are maintained perpendicularly to grounded
14 plate 134. Capacitors Cl and C2 also act as DC blocking
15 ¦ capacitors for the guard ring circuit.
16 l
17 ¦ The stabilized signal produced by the CRef 1 circuit is
18 ¦ used to drive the C x and C+y circuits 50, 52 as will be
19 ¦ now explained. The C+x circuitry 50 will be described as
20 ¦ representative of the operation of the C+y circuitry 52 to
21 ¦ avoid duplicity in explanation.
22 l
23 ¦ C+x measuring circuitry 50 and measurement signal
24 ¦ generation circuitry 70 are shown in Figure 4. Again, the
25 ¦ guard ring 120 and associated diodes D12 and D13 will be
26 ¦ ignored initially. Measuring circuitry 50 of Figure 4 is
27 ¦ identical to the referenced capacitor circuit 40 as a
28 ¦ comparison of Fig. ~ with Fig. 4 will reveal. Conse~uently, as
29 ¦ the driving signal increases from its maximum negative value to
30 ¦ its maximum positive value, Dlo charges capacitor C+x which
31
32 I - 14 -
~l25~3~
1~ ~
2 is comprised of sensor plate 112 supported on the bottom of
3 substrate 110 as the driven plate, and object surface 145 as
the grounded plate. See Fig. 2C. As the signal falls from its
4 maximum positive value to its maximum negative value, Dl1
conducts to drain charge from C~x. This alternate switching
76 on and off of Dlo and Dll causes small equal and oppositely
valued DC currents to flo~ through Llo and Lll as follows:
93 When Dlo is conducting, current flows from the output of
transresistance amplifier 220 which is effectively grounded due
10 to its very low impedance, through R12, L12, L1o, Dlo
11 and C~x back into ground to complete the
13 loop Thus the current through Llo, IL
14 left in Figure 4 and is considered to be positive. When D
conducts, current flows from ground through C+x, Dll,
1~ Lll, L13, R13 and into the low impedance output of
17 transresistance amplifier 230 to complete this loop. Hence,
18 the current througll Lll, IL ~ is left to right in Figure 4
19 and is considered to be negative. Note that these loops are
20 symmetric with respect to one another in that each of the loops
21 are comprised of equivalent components. Consequently, the
22 current values through Llo and Lll while opposite in
23 polarity will be equal in magnitude. The equivalency in
24 magnitude is also the result of the fact that the driving
25 signal is a geometrically symmetric sinusoidal signal as
26 previously mentioned. These small DC currents comprise the
27 measuring capacitor signal for C+x- ~e previously noted that
28 the DC current loop for the reference capacitor signal through
29 L2 flowed from ground through CRef 1~ D2, L2, L4 and
30 R3, and back into the output of transresistance
31 mplifier 200. We a~o n~ted that the 80 microampere current
32
- 15 _
:~5~ 3~ ~
1 through L2 corresponds to a separation between the plates of
2 CRef 1 of the nominal distance. Furthermore, we know that
3 the driven plate 112 of C~x is identical to the driven
4 plate 130 of CRef 1 and is fitted with an identical guard
ring 120 to linearize field lines at the edge of plate 112.
6 The effect of the linearization of the field lines at the edges
7 of driven plates 112, 130 is to equate the grounded plates 145,
8 134 of the capacitors C+x, CRef 1 by making their
9 dimensions effectively identical. The driven plates 112, 130
for both capacitors are 1/4 inch in diameter. ~he grounded
11 plate for C~x is object surface 145 which is effectively an
12 infinite plate. The grounded plate for CRef 1 is the
13 interior surface 134 of the cap 140 having a diameter of 1/2
14 inch. Since the field lines at the edges of driven plates 112,
130 of both of these capacitors are vertical, however, the
16 effective diameter of both of the grounded plates 145, 134 is
17 1/4 inch. Hence, due to the effect of the guard rings 136,
18 120, at a plate separation of the nominal distance C+x is
19 electrically identical to CRef 1' Consequently, assuming
that the separation between the sensor 112 and object
21 surface 145 is the nominal distance, we will get equal and
22 opposite 80 micxoamp currents through Llo and Lll since the
23 DC current loops for Llo and Lll are identical to the DC
24 current loop for L2. As will become apparent, this
80 microamp current level, as set by CRef 1 circuit 40 and
26 signal source circuit 10, and a reference signal (later
27 explained), is the "zero point" for the height of sensor 112
28 above object surface 145,
2g
31
32 - 16 -
. ~2~3~3
1 To demonstrate that the 80 microa~p current level
2 corresponds to the zero point for measuring capacitor C+x, as
3 a first condition let us assume that the object surface 145 is
4 lying at precisely the nominal distance below the sensor 112.
If this is the case, the current through Llo will be
6 +B0 microamps as discussed above. In Fig. 4, this current will
8 flow into measurement signal generation circuit 70.
9 Consequently, if R12 = 50K ohms, the inverting input to
transresistance amplifier 220 will be +80 microa~peres and its
output will be +4 volts at node 235. We have a fixed reference
11¦ voltage VRef 2 of 7 volts injecting a "reference signal"
121 current into node 235 through the resistor Rlo which is a
34 ¦43.75K ohm resistor here. A -160 microampere current is
15 Igenerated by this VRef 2 reference voltage through Rlo,
land assuming R12 - 50K ohms, this current imposes a
16 ¦(50K ohm)(-160 microampere) = -8 volt potential at node 235.
17 ¦The net potential at node 235 is therefore 4 volts - 8 volts =
18 1-4 volts. This -4 volt potential is present at the inverting
¦input to differential amplifier 240. The equal and opposite
21 -80 microampere current through Lll is converted by
l transresistance amplifier 230 to an output voltage of
23 VTR4 = IL R13 = (-80 microamperes) (50K ohms) = -4 volts,
24 assuming R13 = 50IC ohms. This -4 volt signal is imposed at
the noninverting input to differential amplifier 240.
26 Differential amplifier 240 subtracts the -4 volt signal at the
invertin~ input from the -4 volt signal at its noninverting
27
input to get a zero output. This zero output corresponds to
28 the "zero point" for the ~X sensor in that it indicates to the
29
control system ~0 that the sensor is the nominal distance above
31
~ _ 17 -
~ ~5~39 (-
2 the ob~ect surface. The VR f 2 reference signal together
with the driving signal generated by CRef 1 circuit 40 and
3 signal source circuit 10 combine to set this zero point.
4 Note that the overall operation of the circ~itry is to compare
the measuring capacitor signal from C+x circuitry 50 to the
6 reference signal supplied by VRef 2- The VRef 2
7 reference signal imposed at node 23~ is representative of the
)3 nominal plate separation. Measuring signal generating
circuitry 70, thus, compares the measuring capacitor signal
with a nominal plate separation reference signal to determine
12 the deviation of the plates of C+x from the nominal
13 distance. Here, since we have assumed that the C+x plates
14 are separated by the nominal distance, we generate a zero
deviation signal from circuitry 70.
16 Assume now that the object surface 145 is positioned
17 closer than the nominal distance such -that the capacitance of
18 C+x increases and +82 microamperes flow through Llo while
-~2 microamperes flow through Lll. Transresistance
21 amplifier 220 will have an output of VTR3 = (+82 microamperes)
(+50K ohms) = 4.1 volts. ~t node 235 the potential will be
23 4.1 - 8 = -3.9 volts. Thus, -3.9 volts will be present at the
inverting input to differential amplifier 240.
24
Transresistance amplifier 230 on the other hand will have
226 an output of -4-1 volts which will be present at the
noninverting input to differential amplifier 240. Differential
28 amplifier 240 will subtract the difference between -4.1 volts
29 and -3.9 volts = -.2 volts and multiply that value by a gain of
50, for example, to o~taiD an output 24~ of -10 volts. This
31
32 - 18 _
:~S~3~3~ (
1 ¦ -10 vo~t output will be interpreted by the control system to
2 indica~e that t~e object surface is a scaled distance closer to
3 the +X sensor than the nominal distance.
5Again, the circuitry compares the 82 microampere measuring
6 capacitor signal with the VR f 2 signal representative of
7 nominal plate separation. The output generated by circuitry 70
8 is, therefore, representative of the deviation in plate
9 separ'ation of the plates of measuring capacitor C+x from the
nominal distance.
11
12 ¦The circuitry for the C+y capacitor is identical to, and
13 ¦ operates in exactly the same manner as, the circuitry just
14 ¦ described for capacitor C+x. Thus, circuit 40 for CRef
15 ¦ and circuit 10, together with the VRe~ 2 reference signal,
16 ¦ set the same nominal zero point for the +Y sensor 114, and in
17 ¦ exactly the same way.
18 l
19 ¦ The other two measuring capacitors C X and C y are
20 ¦ driven via the lower half of the circuit shown in Figure 2.
21 ¦ With reference to Figure 2, an identical signal to the signal
22 ¦ passing from node 156 upwardly through Rl, passes from
23 ¦ node 156 downwardly. The signal drives identical circuitry,
24 ¦ the only exception being that the secondary coil 184 of
25 ¦ transformer 186 is wound in a direction opposite to its primary
26 ¦coil 182, whereas the coils 172, 174 of transformer 170 are
27 ¦ both wound in the same direction. The result is that a
28 ¦ 180 degree out of phase e~ual and opposite signal drives
29 ¦ CR f 2~ C X and C y~ than drives CRef 1~ C+X and
30 ¦ C+y~ C~f 2 othexwise provides voltage regulation
31
32
- 19 -
~ ~ lZSS)13!~ ~
1 ¦ feedback in exactly the same way as described above to generate
2 ¦ a geometrically symmetric and stable sinusoidal driving signal
3 ¦ for the measuring circuits 54, 56, of C X' C y~ The
4 CRef 2 eircuit 42 and signal source circuit 10 produces
equal and opposite 80 microampere current levels through the
6 C ~, C y eireuits 54, 156 when the -X and -Y sensors 116,
7 118 are separated the nominal distance from the object
8 surfaee 145 in exactly the same manner as was described with
9 reference to the CRef 1 circuit 40.
11 By driving the C+x and C+y circuits 50, 52 with an
12 equal but opposite signal to the signal which drives the C X
13 and C y eircuits 54, 56, a zero net current is passed through
14 the objeet surfaee 145 to ground by the measuring sensors at
any g~ven point in time. This eonstant maintenanee of a zero
16 eurrent flow through object surface 145 prevents object
17 surfaee 145 from developing a voltage potential with respeet to
18 ground due to its impedance to ground. This ensures that
19 objeet surfaee 145 remains at ground potential even if it has a
finite nonzero impedance~ It also prevents the object
21 surfaee 145 from being slight~y charged due to dieleetrie
22 absorption or other effeet. Where the objeet surfaee 145 is a
23 retiele in an eleetron beam system, the maintenanee of zero net
24 eurrent to ground may be advantageous as will later be
explained. In applieations where it is not neeessary to
26 maintain a net zero eurrent flow from the retiele to ground,
27 t~le use of a two phase system would not be required and all
28 measuring eapacitors eould be driven from a single phase RF
29
31 ~
~ 20 -
- ~25(~139
1 signal. Consequently, the capacitance height gauge circuit of
2 the present inve~tion is not intended to be limited to a two
3 phase system such as the one disclosed.
Whether or not a two phase system is used, the four sensor
6 measurements, after being processed by the associated
7 measurement signal generation circuitry, are input to the
8 control system 80 to indicate the position of the object
9 surface 145 with respect to the four nominal zero points. The
control system can then, in response to this information,
11 ¦ either manipulate the position of the object surface, or the
12 position of the tool being used to work on the object surface,
13 for example, as appropriate to the task being performed.
14
Having described the capacitive height gauge of the
16 present invention in its general form, it will now be explained
17 as an integral part of a system for detecting the position of a
18 reticle 250 in an electron beam lithography apparatus. The
19 many advantages and special features of the gauge will become
more apparent from the description of the use of the gauge in
21 this application.
22
23 Fig. 5 shows the environment of the capacitance height
24 gauge as applied in an electron beam lithography apparatus. In
Fig. 5, the gauge is shown as comprised of the hybrid circuitry
26 on the substrate llO and the remainder of the circuitry which
27 is denoted as simply "capacitance gauge electronics" 254. The
28 hybrid circuit substrate llO is secured to the bottom of the
29 electron beam optics housing 258 in an overlying relationship
with respect to the chro~e~on-glass reticle 250. The
31
32 - 21 -
125~1l 39
1~ reticle 250 is supported on a stage 260 in the vacuum
21 chamber 262 of the apparatus for movement in the X and Y
31 directions under the control of the control system 80. A
41 layer 266 of electron beam resist is deposited over the
51 chromium layer 268 of the reticle as shown. The electron
61 beam 270 passes through a central aperture lll in the
71 substrate llO and patterns the beam resist layer 266 with the
81 desired VLSI circuit pattern as the reticle 250 is advanced in
91 the X or Y directions.
10 1
11 ¦ In addition to positioning the reticle 250, the control
12 ¦ system 80 controls the deflection angle of the electron beam to
13 ¦ ensure that it strikes the desired point on the reticle with
14 ¦ the high degree of precision required of VLSI circuitry~
1~ I
16 ¦ Given the high precision required in this art, prior to
17 ¦ utilizing the electron beam to write on the reticle 250, the
18 ¦ beam 270 must first be calibrated with respect to the curre~t
19 ¦ environmental conditions. A calibration plate 280 such as that
20 ¦ shown in Fig. 6 is used for this purpose. The calibration
21 ¦ plate 280 is secured to the stage 260 adjacent to the
22 ¦ reticle 250, and directly below the electron beam optics
23 ¦ housing 258. A conductive layer 282 covers the calibration
24 ¦ plate 2807 A grid 284 of very tiny nonconductive points 286 is
25 ¦ formed into layer 282. The grid 284 is comprised of a
26 ¦ multitude of equally spaced and symmetrically arranged
27 1 conductive points 286. Point 290 is centered with respect to
28 ¦ the centerline of optics housing 258 by movement of the stage
29 ¦ to specified X, Y coordinates. Point 290 is denoted the
30 ¦ "origination point." With the cap gauge measuring circuitry
3~ ~
32 I - 22 -
~- lZ5(J13~ (-
1 turned off, the electron beam 270 is turned on and is
2 sequentially advanced by the control system from one
3 nonconductive point on the grid to the next. The control
4 system 80 senses each sequential nonconductive point and stops
5 when it strikes the desired point 294. The deflection angle A
6 which is necessary to cause the beam ~o strike precise
7 point 294 is recorded by the control system. Point 294 is
8 located a specified distance "d" from origination point 290
9 since the geometry of grid 284 is known. The electron beam 270
is now turned off and the cap gauge circuitry is turned on to
11 take calibration plate position readings from the four
12 sensors 112, 114, 116, 118. The associated measurement signal
13 generation circuitry for each sensor determines the position of
14 the calibration plate 280 with respect to the nominal zero
15 point or the sensor and enters this information into the
16 control system 80. A best fit plane is calculated from the
17 four points using well known mathematics. This plane defines
18 the "calibrated plane" for the reticle 250.
19
Once the calibrated plane has been determined, stage 260
21 is moved to position reticle 250 under electron beam optics
22 ousing 258 as shown in Fig. 7. The point 298 on the
23 reticle 250 directly aligned with the center line of the
24 housing is again denoted as the origination point. The
oint 300 which is the precise distance d from point 298 is the
26 esired point for writing on reticle 250. The reticle outlined
27 by dotted lines in Fig. 7 is located precisely in the
28 alibrated plane, and hence, the beam 270 would strike
29 oint 300 on the dotted reticle using the normal deflection
ngle A. The actual ~osition of the reticle in Fig. 7,
31
321 - 23 -
lZ50139
1 ; however, is illustrated by reticle 250 outlined in solid
2 lines. In this position the beam 270 would write at point 302
3 if normal deflection angle A were used.
41
~o avoid such an error, prior to writing on the reticle,
6 reticle position readings are taken by the four sensors 112,
7 114, 116, 118. These readings are input to the associated
8 measurement signal generation circuits and then to the control
9 system. A best fit plane is calculated for the four points.
This plane is known as the "reticle plane." The control system
11 compares the reticle plane with the calibrated plane and
12 determines the deviation. In thi 5 case, the control system
13 will determine that the reticle is a precise distance above the
14 calibrated plane and will adjust the electron beam deflection
15 ¦ angle~from normal deflection angle A to a precise corrected
16 ¦ deflection angle B to ensure that the beam writes at the
17 ¦ precise point 300 notwithstanding the fact that the reticle 250
18 1 is positioned above the calibrated plane. The electron beam
19 ¦ then writes or patterns the segment of the circuit included
20 ¦ within this stage position and advances the reticle to the next
21 ¦ stage position. The reticle position readings are again taken
22 1 and processed to determine the deviation of the reticle plane
23 ¦ from the calibrated plane and appropriate adjustments are made
241 to the beam deflection angle to account for any deviation.
251 This cycle is repeated at each stage position until the reticle
26¦ is completely patterned, or a chanye in environmental
27 ¦conditions warrants recalibration of the electron beam.
28
29 1
30 1 24 -
32
~;~5~3~3~
(~ (
1 Having described the capacitance height gauge as applied
2 in the reticle position detection system disclosed, various
3 advantageous features of the invention will now become more
4 apparent. Fig. 8 shows the top surface of the hybrid circuit
substrate 110. As was mentioned previously, all of the hybrid
6 circuitry except for sensor plates 112, 114, 116, 118 and the
7 associated guard rings 120, 122, 124, 126 are supported on the
8 upper surface of the substrate. Thus, the reference
9 circuits 40, 42 and identical measuring capacitor circuits 50,
52, 54, 56 are supported on the top of the substrate. Each of
1 these circuits is comprised of four capacitors, four pin diodes
12 and two coils, all of which are discrete components
13 interconnected by signal lines deposited on the substrate. The
14 four circuits 50, 52, 54, 56 feed through tc the sensors 112,
lS 11~, 116, 118, respectively, and their associated guard
16 rings 120, 122, 126, 128. The two circuits 40, 42 feed
17 directly to their respective driving plates 130 and guard
18 rings 136 on the upper surface of the substrate 110.
19
The hybrid circuitry carried oh the top of the substrate
21 also includes the resistance heating strips 310, shown in
22 Fig. 8, which can be silk screened on the substrate. These
23 strips 310 together comprise the resistor RStrips shown in
24 the simplified circuit diagram of Fig. 9 which shows the
substrate temperature control circuitry of the invention. The
26 function of these resistive strips is to heat the substrate to
27 a temperature of a few degrees above room temperature and
28 thereby maintain control over the temperature of the hybrid
29 circuit componen~s at all times. With reference to Fig. 9, a
rheostat 312 is used to set the desired temperature of the
31
32 - 25 -
~- lZ5~39 ~ .
1 substrate 110. This temperature setting is input as a
2 Icorresponding voltage, ~, volts for example, on the inverting
3I pin of differ~ntial ampliricr 314. A substrate temperature
41 sensor 3~ scaled to one microampere per degree kelvin outputs
5 ¦ a current proportional to the substrate temperature. This
6 ¦ current signal is input to the noninverting input of
7 ¦ transresistance amplifier 318. The inverting input of the
8 ¦ transresistance amplifier 318 is provided with an offset
9 ¦ signal,' VRef 4, to convert the output of amplifier 318 to
10 ¦ degrees Celsius. The amplifier output is applied to the
11 ¦ noninverting input of differential amplifier 314 wherein it is
12 compared with the rheostat setting. Where the two inputs are
13 equal, the output voltage of the amplifier provides the
14 required voltage drop across the resistance heating elements
Rstrip to maintain the substrate at the current temperature.
16 If the substrate temperature drops below the desired
17 temperature then the output voltage of differential
18 amplifier 314 decreases to increase the voltage drop across the
19 resistive strips and thereby increase the substrate
temperature. Conversely, if the substrate runs too hot the
21 voltage drop across the resistive strips is reduced to cool the
22 substrate. The resistive strips 310 are evenly distributed
23 about substrate 110 to prevent temperature gradients.
24
By regulating the temperature of the substrate in this way
26 the identical components of the reference and measuring
27 capacitor circuits are maintained at the same operating
28 temperature and the accuracy of the gauge is protected from
29 temperature drifts. Moreover, in that both the reference and
measuring capacitoss are in the same "measuring environment,"
32 _ ~6 -
~S~l3~
1 the dielectric between the plates of all of these capacitors
2 undergoes the sa~e environmental ehanges, such as changes in
3 temperature and humidity. Consequently when environmental
ehanges do occur in the vacuum chamber, resulting in a change
in the capacitance of the reference and measuring capacitors,
6 the ehanges are automatically aecounted for by the reference
7 eapaeitor eireuits whieh maintain the eurrent flow through the
8 referenee and measuring capacitor circuits at a stable level
9 regardless of environmental changes.
ll Consequently, current flow through the measuring
12 eapaeitors varies only with plate separation in that
13 environmental effects are minimized. Another important
14 advantage of the hybrid eircuit embodiment disclosed is that
the referenee and measuring capacitor circuits are not only
16 identical, and subject to the same environmental influences,
17 but in addition, they are subjected to the same environment
18 throughout their operating life. Hence, the eireuit components
2 tend to age together at the same rate and in the same manner
O whieh further ensures the accuracy of the gauge over time.
222
23 A further important aspect of the invention is the fact
that the four diodes used in the reference and measuring
24 eapaeitor eircuits are ultra low capacitance PIN diodes. These
25 ¦ diodes function as extremely small capacitors in their reversed
276 ¦ biased state, and hence, do not contribute to the current
28 ¦ flowing through the measuring and reference capacitors. In
29 ¦ addition, the eapacitance of these diodes remains very small
30 ~ even during wide temperature drifts.
3~
3' - ~7 -
~25~3~
2 Havina describe~ the invention in its presently pr ferred
embodiment, the manner in which the outputs of the four sensors
3 are correlated with plate separation wi~l now be explained.
Figs. 10 and 11 show the calibration fixture 500 on which
6 the sensors are calibrated. Fixture 500 has a base 502 and
three adjustable supporting legs 504. Legs 504 can be
8 threadably adjusted, for example. Base 502 supports a laser
table 506, which in turn supports a laser 508. A vertically
10 upstanding member 510, supported by base 502, supports a laser
12 receiver 512. Laser receiver 512 is secured to member 510 by
screws 14, for example. A support block 516, supported by base
13 502, supports an interferometer cube support member 520, which
14 is best shown in Fig. 12. Support member 520 supports the
15 interferometer cube, or prism, 522, behind a verticle wall
16 524. Wall 524 has a large aperture 526 through which the laser
17 beam projects as will later become apparent. Four internally
18 threaded posts 528 project orthogonally from the front side of
wall 524, as shown. The cap gauge substrate 110 is threadably
20 secùred by screws 530 to the posts 528. Each screw is
21 counter-sunk within a precisely-machined washer 532. By
22 mounting the substrate 110 in this way, it is intended that the
23 substrate's surface 113 will lie in a plane substantially
24 perpendicular to the path of the laser beam. Note that the
2~ central aperture 111 of substrate 110 aligns with the aperture
27 526 of wall 524 and also permits the passage of the laser
beam. It is noted that the portion of the calibration fixture
29 so far described, comprised o~ the laser 508, the laser
receiver 512, and the interferometer cube 522, together with
laser electronics (later described) are all commercially
31 availab~e a.~ a pac~age from Hewlett-Packard as a laser
32
28
5V~
2 ¦ interferometer, Model No. HP5501A, having a resolution of 5
l nanometers. These components together comprise the laser
3 interferometer assembly and can measure precise distances in a
4 manner later described.
6 Note that all components of the laser assembly are rigidly
7 secured to the fixture 500. The portion of the calibration
8 fixture 500, which will next be described, comprises an
9 assembly which movably positions a mirror, or dummy reticle,
0¦ with respect to the laser assembly.
11 1
12~ A support table 540 supports an air bearing 542. Air
13 bearing 542 supports a longitudinally directed shaft 544 for
14 ¦ verv low friction rectilinear movement. End 546 of shaft 544
15 ¦ rigidly supports a mirror support member 548, which in turn
16 supports the mirror 550 in a vertical orientation, directly
17 opposite to the sensors 112, 114, 116, 118 of cap gauge
18 substrate 110. The mirror surface is electrically conductive
19 in that it functions as one plate of a measuring capacitor, as
will be later described. A drive magnet 552 is supported by
21 shaft 544 on the opposite side of air bearing 542 from the
22 mirror 550. Magnet 552 is reciprocally driven by voice coil
3 544, which encircles shaft 544 and is rigidly supported by
24 table 540. A velocity damping magnet 556 is secured at the end
558 of shaft 544. A velocity sensing coil 560 encircles the
26 shaft 544 adjacent to magnet 556 and is rigidly supported by
27 table 540. Drive coil 544 is energized in a manner later
28 described to drive the mirror 550 to the desired location with
29 respect to the sensors 112, 114, 116, 118. Velocity damping
magnet 556 moves with shaft 544, and as it moves, it generates
31
32 - 29 -
39
1 ¦ a voltage across the velocity sensing coil 560. This voltage
21 is used as a ~elocity damping signal in a manner which will
31 also be later described.
41
51 In order to provide vibration isolation for the
61 calibration fixture 500, the support legs 504 can rest upon a
7¦ massi~e granite block 562, which in turn can be supported by
8 ¦ air bags 564. In addition, the entire fixture can be installed
9 ¦ in a temperature controlled environment 566.
10 I
11 ~ Having described the basic structure of the calibration
12 ¦ fixture, its operation will now be described with reference to
13 ¦ the partially schematic, functional block diagram shown in
14 ¦ Fig. 13.
15 I ,
16 As noted above, the laser interferometer assembly,
17 comprised of the laser 508, interferometer cube 522, laser
18 receiver 512, and laser electronics 570, is commercially
19 available as a package to very precisely measure small
distances. In this case, the laser interferometer assembly is
21 used to determine the distance between the sensors 112, 114,
22 116, 118 and the surface of the mirror 550. The interferometer
23 assembly is used in the present invention in that it is the
24 best "yardstick" available for determining the distance between
the sensors 112, 114, 116, 118 and the mirror 550.
26
27 Briefly, the laser beam generated by laser 508 passes
28 through interferometer cube 522 and is reflected off of mirror
~9 550 back to the laser receiver 512. The laser receiver 512
indicates to laser electronics 57~ the position of mirror 550.
32
- 30 -
5~3~3~
l~Las electronics 57~ is advised~ in a ~anDe~ later described,
2 of the desired position of the mirror and utilizes the input
3 from laser receiver 512 to determine the differential distance
4 between the actual position of mirror 550 and its desired
position. Laser electronics 570 then outputs this differential
6 distance as an error signal. The use of this error signal by
7 the present apparatus to control the position of mirror 550
8 will be described shortly.
With reference to Fig. 13, to begin the calibration
11 procedure, microprocessor 574 instructs the calibration fixture
12 electronics 576 to energize drive coil 554 to drive mirror 550
13 into abutment with the four machined washers 532, which are
14 secured to substrate 110, as previously described. The
thickness of washers 532 is very accurately machined so that
16 the mirror 550 is positioned 20 milliinches from the sensors
17 112, 114, 116, 118. Microprocessor 574 now tells laser
18 electronics 570 that the mirror should be moved an additional
19 40 milliinches away from the sensors 112, 114, 116, 118. Laser
receiver 512 tells the laser electronics 570 where the mirror's
21 present position is, and laser electronics 570 generates an
22 error signal representative of the distance between actual
23 position of the mirror 550 and its desired location. This
24 error signal is input to calibration fixture electronics 576,
which in response thereto, energizes drive coil S54 to drive
26 mirror 550 back towards the desired position. As the mirror
27 ¦ moves away from the sensors 112, 114, 116, 118, the laser
28 ¦ receiver 512 constantly monitors its position and the error
29 ¦ signal generated by laser electronics 570 is correspondingly
30 ¦ reduced as the mirror 550 moves closer and closer to the
31
32 I - 31 -
~Z5(3~3~
2 de red location. As the mirror moves, the velocity damping
magnet 556 moves with respect to the velocity sensing coil 560,
3 and a voltage is generated across the coil 560 in proportion to
4 the speed of the magnet 5560 This velocity signal is input to
the calibration fixture electronics 576 as a velocity damping
6 signal to prevent oscillation of the mirror about the desired
7 location in a manner more fully described later on. Once the
8 mirror reaches the desired position, a "zero error signal" is
9 output from laser electronics 570, and calibration fixture 576
controls drive coil 554 to stop mirror 550 at the desired
11 location. At this time, laser electronics 570 indicates to
12 microprocessor 574 that mirror 550 has reached the desired
13 location which is 60 milliinches from the sensors 112, 114,
14 116, 118 in this initial situation. Microprocessor 574 now
takes voltage readings from the four sensors 112, 114, 116, 118
16 and stores this information for each sensor for this
17 initialization point of 60 milliinches. Microprocessor 574
18 then directs laser electronics 570 to move the mirror to a
19 first sample point 5 milliinches, for example, from the
initialization point. Laser electronics 570 again generates an
21 error signal representative of the 5 milliinch distance between
22 the actual position of the mirror and the new desired location
23 and this error signal is again used by calibration fixture
24 electronics 576 to drive mirror 550 to the new desired
location. When the location is reached, a zero error signal is
26 again generated by laser electronics 570 and in response
27 thereto calibration fixture electronics stops the mirror 550
28 and microprocessor 574 again samples the four sensors 112, 114,
29 116, 118. The voltage reading for each sensor 112, 114, 116,
118 is stored by microprocessor 574 for this new location. The
31 mirror is the~ mo~ed i~ accordance with the control program of
32
- 32 _
~5~
(
1 ~ micro~rocessor 574 to the next sample point by the same
_ I procedure and these sensors are again sampled at this new
3 ¦ location. Obviously, sensor readings for any desired number of
4 ¦ sample points can be obtained in this way. Microprocessor 574
5 ¦ stores this information in the form of a table, listing the
6 I voltage readings for each sensor at each sample point. In this
7 ¦ way, the sensors can be calibrated over a range of plate
8 ¦ separation distances.
9 '
Having described the overall operation of the calibration
11 system of Fig~ 13, calibration electronics 576 will now be
12 described in somewhat more detail with reference to Fig. 14.
13
14 As shown in Fig. 14, the error signal is delivered to a
15 ¦digital-to-analog converter 580 in the form of a 12-bit digital
16 ¦word directly from laser electronics 570. The analog signal
17 ¦produced by digital-to-analog converter 580 is then scaled by
18 ¦amplifier 582 to a 0-10 volt scale before being delivered to
19 ¦polarity determining circuit 584. Laser electronics 570 also
20 ¦outputs a "direction bit" which is delivered to polarity
21 ¦determining circuit 584, as shown. The direction bit indicates
22 ¦the polarity of the error signal. That is, it indicates which
23 ¦side of the desired location mirror 550 is currently located on
24 Iso that mirror 550 can be driven in the proper direction. The
25 direction bit is processed by polarity determining circuit 584
27 to assign the proper polarity to the error signal before it is
input to amplifier 586. Amplifier 586 then scales the error
28 signal between 0 to +10 volts, if the assigned polarity is
29 positive, or between ~ to -10 volts, i~ the assigned polarity
30 is negative. The err{)r signa~ is then input to position gain
31 amplifi~r 5~ qrhe position gain is adjustable by posi tion
32 - 33 -
~2~f~3~ ~
sain adjusting circuit 590, which may, for example, comprise a
2 ¦ potentiometer positione~ in the feedback loop of amplifier
3 1 5~3~. Additional damping is provided by velocity coil 560. The
4 ¦ signal generated by coil 560 is input to velocity gain
~ ¦ amplifier 592. A velocity gain adjusting circuit 594 is
6 ¦ provided to permit adjustment of the gain characteristics of
7 ¦ amplifier 592, and Iray also comprise a potentiometer positioned
8 ¦ in the feedback loop of amplifier 592. The gain adjusting
9 ¦ circuits 590 and 594 are adjusted together to stabilize the
10 ¦ circuit and prevent oscillation of mirror 550, or ringin~,
11 ¦ about the desired location. The signals from amplifiers 588
12 ¦ and 592 are summed before finally being filtered at noise
34 ¦ filter 596, and then finally amplified at power amplifier 598
I to attain the necessary power level to drive the drive coil 554.
15 I
16 ¦ The entire calibration fixture 500 is contained in a
17 ¦ temperature-controlled environment which may also be under the
18 ¦ control of microprocessor 574. Likewise, the cap gauge
19 ¦ substrates temperature control circuit shown in Fig. 9 may also
20 ¦ be under the control of microprocessor 574. Microprocessor 574
21 ¦ will generally maintain the environmental temperature a few
22 ¦ degrees cooler than the substrate temperature so that
23 ¦ temperature gradients in the substrate are avoided. The
24 ¦ environment and substrate temperatures may be varied by
25 microprocessor 574 according to its control program to generate
26 sets of calibration data at various temperatures. ~hen the
27 capacitance gauge is then later used in an electron beam
28 apparatus~ for example, the appropriate set of calibration data
29 for the temperature at which the appara~us will be used can be
30 utili2ed.
31
32 - 3~-
~5~3~
-- (
2 An important aspect of the present invention is that after
a particular capacitance gauge has been out in the field for a
3 period of time, it can be recalibrated on calibration fixture
4 550. For recalibration, the substrate is simply reinstalled in
the fixture 500, and run through the same calibration procedure
6 as just described. The results of the calibration process will
7 indicate whether the sensor readings have varied with time or
8 have remained the same.
9 .,
It should be noted that the intialization point described
11 ¦ above may not be exactly 60 milliinches from the sensors, but
12 ¦ rather is nominally 60 milliinches from the sensors 112, 114,
13 ¦ 116, 118. The principal reason for this is that, although the
14 ¦ washers 532 are finally machined to a 20 milliinch thickness,
15 ¦ in tha't they are a mechanical part, it is expected that there
16 ¦ will be some actual variation from an exact 20 milliinch
17 dimension. Consequently, when the mirror is first placed into
18 abutment with the washers 532, it may or may not be precisely
19 20 milliinches from the sensors 112, 114, 116, 118. Therefore,
when it is backed up 40 milliinches by the microprocessor 514,
21 the initialization point will only be a nominal 60 miliiinches
22 from the sensors rather than necessarily an exact 60 milliinch
23 distance from the sensors. As the mirror is moved in 5
24 milliinch increments, however, the incremental distances will
be separated by precisely 5 milliinches to the accuracy of the
26 laser interferometer assembly, which is far more accurate than
27 any mechanical measuring technique. Consequently, although the
28 initialization point may not be exactly 60 milliinches from the
29 sensors, we know that the subsequent test points can be stepped
off in precise 5 milliinch increments to the accuracy of the
31 laser inter~erometer. ~reover, it ;s these "difference
32 - 35 -
~5g~3~
1 readings" between the various points, which are of most concern
2~ to us. This is particularily true in the present applicatiOn~
3 for example, wherein the ca?acitance gauge is being used to
41 measure the difference between the distance to the referenee
51 plane and the distance to the reticle. Such distance
61 differenees ean be measured very aceurately by the instant
71 apparatus, and aeeordingly, fine adjustments ean be made to the
8 eleetron beam apparatus to eorrect for the aetual position of
9¦ the retiele.
101
1¦ An alternative manner in whieh the outputs of the four
2¦ measuring sensors are correlated with plate separation will now
13 ¦ be explained with reference to Fig. 15.
14 I
15 ¦ The capacitance gage substrate 110 is rigidly affixed in a
16 ¦ test fixture 330 opposite to a dummy reticle 334. A laser
17 ¦ interferometer 338 is also supported on fixture 330 and
18 ¦ projeets a laser beam through opening 111 in substrate 112
19 ¦ towards reticle 250 which reflects it back towards the laser.
20 ¦ Reticle 250 is supported by a member (not shown) which is
21 ¦ supported in an air bearing (not shown) for low friction
22 ¦ rectilinear movement with respect to substrate 110. The
23 ¦ reticle support member is reciproeally moveable by means of
24 ¦ reticle positioning mechanism 342. The fixture 330 is housed
25 ¦ in a temperature eontrolled enelosure on a vibration isolating
26 ¦ granite block whieh rests on air bags. The fixture temperature
27 ¦ ean be controlled by means of temperature control
28¦ electronics 346, fixture oven 350, and fixture thermometer 35~.
29
311
32l ~ 36 -
3~
2 According to the testing procedure, the microprocessor 360
directs the laser electronics 364 to move the reticle to a
3 specified distance from the sensors of substrate llO. Laser
4 electronics 364 relays the command to test fixture
electronics 368 which initiates reticle positioning
6 ¦mechanism 342 to move reticle 334 in the direction specified.
71 As reticle 334 moves towards substrate llO, laser
81 interferometer 33~ and laser electronics 364 monitor its
l progress and laser electronics 364 commands test fixture
10¦ electronics 368 to stop the movement of reticle 334 once the
11 ¦specified distance between reticle 334 and substrate llO is
12 ¦ achieved. Velocity sensor 372 damps the movement of
13 ¦ reticle 334 to prevent oscillation about the desired point.
14 ¦Microprocessor 360 now samples the substrate sensors via the
15 ¦capacitance gage electronics 376. The temperature of cap gage
16 ¦ electronics 376 can be controlled by the oven 380,
17 ¦thermostat 384, and temperature control electronics 346. In
18 ¦that microprocessor 360 knows the distance between the
¦substrate sensors and the reticle, since the laser
20 ¦interferometer has measured the distance, it can correlate the
21 ¦ output of the sensors with that distance. Microprocessor 360
22 ¦ takes sensor readings at various distances and temperatures and
23 ¦ stores the data generated.
2241
l Note that various reticle positioning mechanisms could be
26 ¦ used. For example, two voice coil/magnet assemblies could be
271 mounted coaxially on the centerline of the moving parts. The
28¦ magnets could be attached to the moving air bearing slide, with
291 the voice coils f~xed ~n position~ One coil/magnet set could
30 ¦ be used ~or v~loc~ty ~ensor damping ~nd the other for movement
31 ¦ of the reticle.
32
- 37 -
3~
,
2 ¦ The data accumulated from the sensors can be fit to a
3 ~ curve defined by the equation:
41 Z(v) = k/v-vo + Zo ~ Av +Bv + Cv,
5 l
61 where "Z(v)" is plate separation "Z" as a function of cap gage
71 output voltage "v" for the sensor, with K, vO, Zo, A, B and C
8¦ being constants which can be determined by the ~icroprocessor
for the particular cap gage being tested.
11 The Z(v) = k/v-vo -~ Zo portion of the equation can be
121 derived from Maxwell's equations to give the theoretical
13 function. The polynomial Av + Bv + C was empirically
14 found by the inventors herein to be necessary to reduce errors
to thé noise level.
16
17 Consequently, the microprocessor tests the cap gage and
18 calculates the particular values of the constants for the gage
19 being tested to provide a formula relating cap gage output
voltage to plate separation. This formula can be used by the
21 control system of the electron beam llthography apparatus to
232 calculate the calibration plate points necessary to define the
calibrated plane and the reticle position points necessary to
24 define the reticle plane. Having defined the planes, the
control system can determine the deviation of the reticle from
26 the calibration plane and adjust the deflection angle of the
27 electron beam accordingly.
28
29 - 38 -
32
~ ~LZ~3~ ( .
2 Having disclosed the presently preferred embodiments of
l the invention, many modifications and variations thereof will
31 be obvious to those skilled in the art. Particularly, others
41 will recognize that the calibration fixed apparatus disclosed
51 for calibrating capacitance height gauges may be applicable to
6¦ the calibration of other types of capacitive height gauges than
71 the one particularly disclosed herein. Accordingly, the
8 ¦ present invention is intended to be limited only by the scope
9 ¦ of the appended claims.
13
14
17
18
19
2.1
22
23
24
26
27
28
2g I
311
321 - 39 -