semi合集-English.pdf - 第1416页

SEMI E30-1103 © SEMI 1992, 2003 90 calibration. The need for calibrati on is a non-linear function of the number of wafers processed in that chamber. A status variable exists which contains the number of wafers processed…

SEMI E30-1103 © SEMI 1992, 2003

A.7 Examples of Limits Monitoring

A7.1 Introduction

A7.1.1 Four limits monitoring examples are included

below to help clarify the use of limits and to illustrate

typical applications. The first example shows how to

apply limits to boolean values. The second illustrates

application of several limits to a floating point variable

in a classical control zone style. The third example

shows an integer counter variable used to prompt for

equipment maintenance.

A7.2 Examples

A7.2.1 Example 1 –– Valve Monitoring

A7.2.1.1 The ACME Shine-Um-Rite Model 13

includes a sump which contains the chemical agent

used to clean bare wafers. A chemical feeder system

serves to refill the sump when the level drops below a

certain level. The fill is accomplished via an on-off

value driven by sensors in the sump. Facilities must be

informed of the proportion of the time the valve is open

(approximates usage) and any time the value remains

open for more than 5 minutes (valve likely broken).

A7.2.1.2 To implement this requirement, a limit was

defined for the Boolean status variable which contains

the current state of the valve (i.e., 0 = Closed, 1 =

Open). See Figure A7.1 for illustration. LIMITID1 was

defined with UPPERDB = LIMITMAX = 1 (Open) and

LOWERDB = LIMITMIN = 0 (Closed). As a result,

any time the valve opens, a collection event is

generated with TransitionType = 0 and when the valve

closes, a collection event is generated with

TransitionType = 1. An event report containing the

DVVAL LimitVariable was attached to each collection

event and reporting for the event was enabled.

NOTE: Boolean values are defined as 0 =

False/Closed/Off and any value > 0 = True/Open/On ––

never depend on a value of 1.

Figure A7.1

Valve Monitoring Example

A7.2.2 Example 2 –– Environment Monitoring

A7.2.2.1 ACME also makes a Model 2 Stepper. The

environmental control system of this equipment is

designed to hold the internal temperature relatively

constant, but is sensitive to large changes in the

external environment, opening of access doors, etc. To

ensure that processing conditions are appropriate, the

internal stepper temperature is monitored to ensure it

remains in a safe operating zone (within “Shutdown”

limits). In addition, a second set of limits are used

within the Shutdown limits to bound the “Normal”

operating range. Frequent excursions from the normal

range into the “warning” range will prompt service on

the environmental control system. The target

temperature range is specified as 98–100° , the

shutdown limits as 95–103° .

A7.2.2.2 Event reports are desired when the internal

temperature moves outside of the normal operating

zone into a warning zone (above or below), when the

temperature moves back into the normal operating zone

from the warning zones, and when the temperature

moves out of the warning zones into the shutdown

zones. Furthermore, temperature fluctuations of 0.5°

should not trigger multiple event reports.

A7.2.2.3 Probably the most intuitive use of the limits

monitoring capability is in establishing normal,

warning, and shutdown zones for a particular

equipment variable. Limits may be combined to provide

such a scenario. The method is described below and

illustrated in Figure A7.2. Please note that in the figure,

limits are denoted as solid lines for simplicity, with

deadbands indicated using the ± notation.

Figure A7.2

Environment Monitoring Example

A7.2.3 Example 3 –– Calibration Counter

A7.2.3.1 Another ACME equipment is the multi-

chamber Duz-It-All Model 7. This machine includes

redundant chambers to increase throughput. One

particular chamber on this equipment requires periodic

SEMI E30-1103 © SEMI 1992, 2003

calibration. The need for calibration is a non-linear

function of the number of wafers processed in that

chamber. A status variable exists which contains the

number of wafers processed since the last calibration

was performed. Maintenance is definitely required after

every 8 cycles, but the machine must be checked after 5

and 7 cycles to determine whether early calibration is

necessary. This checking may be done by examining

certain other equipment status variables.

A7.2.3.2 To meet this need, three limits are defined for

the counter variable. Three limits were set, at 5, 7, and

8. Deadbands are set to zero, since chattering is not a

problem. All the pertinent information is placed in an

event report which is attached to the CEID for the

limits of the counter to negate the need for further

message exchange. Event reports are generated as each

limit is reached (one zone transition each), and when

the counter is reset following calibration (one, two, or

three zone transitions referenced in one report). Figure

A7.3 illustrates this example. Note that disabling the

report upon counter reset (downward transitions) is not

possible.

Figure A7.3

Calibration Counter Example

A7.2.4 Example 4 –– Derived Variables

A7.2.4.1 The flagship of the ACME line is the new

HotDog Furnace. This is a vertical furnace which

exposes wafers to a variety of temperatures during

processing. The temperature profile during the run is

critical to the process and is typically contains a number

of plateau’s at different levels during the run. The

owner wishes to monitor the temperature and be alerted

whenever the actual temperature profile differs from the

ideal by > 0.5° . The derived variable was created to

provide a steady target range during the run, no matter

what the desired temperature range happened to be.

Deviation from “ideal temperature profile” was chosen

as the new variable to be monitored. The equipment

already had access to the profile for the run, which

described the desired temperature at a given time into

the process. The manufacturer added a calculation each

time the actual temperature was sampled, subtracting

the ideal temperature from the actual. They provided as

status variables the actual temperature, the ideal

temperature, and the new “deviation from profile”

variable. One limit was activated and set to 0.5 degrees

and a second set to –0.5 degrees (each with a deadband

width of 0.05). Thus, when the temperature deviation

from setpoint exceeds ± 0.5, an event is generated

containing the current desired temperature and the

actual temperature. For good measure, additional data

was added, providing time since start of run to

document the precise point in the process that the

problem occurred.

A7.2.4.2 In order to achieve the desired behavior, the

host defines four monitoring limits. Two of the limits

establish the target zone. These are responsible for

reporting transitions from normal to warning zones in

either direction. The other two limits establish the

transitions between the warning zones and the error

zones. The difference between UPPERDB and

LOWERDB for each limit is 0.5. This may also be

expressed as limit ± 0.25. Combining limits does not

change the way the equipment treats limits monitoring,

but rather builds a method of interpreting limits from

the host’s point of view.

A.8 Process Parameter Modification for

Process and Equipment Control

A8.1 Introduction

A8.1.1 In many equipment control applications there is

a need for a GEM host to modify one or a small set of

process parameters associated with a recipe. The

number of parameters modified, frequency of

modification (e.g., wafer-to-wafer, batch-to-batch, etc.),

range of modification, etc., is largely a function of the

equipment control application. Utilizing GEM, at least

two methods are envisioned for modifying process

parameters on a tool. With the first method,

“Equipment Constants” can be used to relate process

parameters of the updated recipe. Equipment Constants

can also be used in a mode where they relate suggested

modifications to process parameters from the stored

recipe, i.e., the constants contain only the + /-

differential from a nominal value. The former mode is

preferred because it better ensures data integrity

between the controller and tool. With the second

method the entire recipe could be downloaded, but this

results in an enormous amount of communication

overhead. Note that, in all cases, the Equipment

Constants do not replace the process parameters inside

a recipe, but are associated with (e.g., linked to) these

parameters to relate modifications. The remainder of

this application note provides a description of how

process parameter modification can be implemented

using existing GEM capabilities. The method may be

used in a GEM compliant system provided that the

specific GEM capabilities described are supported.

SEMI E30-1103 © SEMI 1992, 2003

A8.2 Equipment Constants

A8.2.1 Incremental process parameter modification for

process and equipment control can be supported over a

GEM interface by using the Equipment Constants GEM

capability (see Section 4.5). With this capability, each

process parameter (or process parameter at a step) that

can be modified, e.g., for purposes of process control,

is associated with an equipment constant. Using the

Equipment Constant GEM scenarios (see Section 4.5.5)

the host can (1), send process parameters or parameter

modifications, (2), request current values of modifiable

recipe parameters, (3), retrieve name lists of equipment

constants associated with modifiable parameters, and

(4), be informed by the equipment when an operator

changes one of the modifiable process parameters.

A8.2.2 The equipment constants should represent the

actual values of the process parameters with which they

are associated. Depending on the equipment operation

and control application, the equipment constant could

represent the actual value of a process parameter at a

recipe step, or over the entire recipe. The equipment

constants could also be utilized to represent the

differentials of process parameters from nominal

values. However it is important to note that, when

using differential values to relate process parameter

modifications, any loss of synchronization between

equipment and host could result in an incorrect

assessment of the value of a process setpoint by the

host. Note also that, upon system startup, and whenever

the appropriate process parameters are modified, the

equipment constants should also be modified as

necessary to always reflect the (absolute or relative)

values of the associated process parameters.

A8.2.3 In order to maintain synchronization between

equipment and host, it is recommended that equipment

constants associated with recipe parameters be applied

to only override the currently selected and active recipe.

A selected recipe is considered to be active whenever

the equipment is in the “PROCESSING” state and the

recipe is the currently selected recipe (process

program). If multiple recipes are utilized during one

process event, e.g., cluster tool scenario, it is

recommended that separate equipment constants be

utilized for each recipe/process parameter pair.

A8.2.4 Note that timing and traceability issues

associated with utilizing the equipment constants

capability (for process control) are application specific

and beyond the scope of this application note.

Equipment that provides for recipe parameter overrides

though setting of Equipment Contstants should also

provide additional Equipment Constants for the set that

includes the name of the associated process program

and a Boolean variable to enable and disable the

override feature. In addition, the supplier should

document for each parameter: (1), the associated

Equipment Constant, and (2), any restrictions on the

state (active or not) or the recipe in which the parameter

may be modified. Also, since override of the process

setpoints may be provided by a Host controller

application element, it is recommended that the

equipment provide an event report whenever the

associated recipe has been modified.

A8.3 Example

A8.3.1 In the example of Figure A.8.1, a Chemical-

Mechanical Planarizer (CMP) single wafer “polishing”

system includes a GEM compliant planarizer

(equipment), a thickness metrology unit and a (host)

controller. The tool polishes a wafer to a target

thickness. The post-process thickness is measured by

the metrology unit and reported to the host controller.

The controller utilizes a feedback control algorithm to

determine the appropriate polish “time” recipe

parameter for the next wafer. This time should be

reported to the CMP equipment utilizing the

equipment’s GEM interface so that the information can

be utilized for the next wafer processing event.

A8.3.2 The mechanism described in this application

note could be utilized to implement process control as

follows. A settable equipment constant is associated or

“linked” with the “time” parameter on the equipment.

Equipment system documentation indicates the linkage

and the conditions under which the linkage is valid.

When the controller determines an appropriate “time”

parameter value for the next wafer to be polished, it sets

the equipment constant to this value (see Section 4.5).

The equipment is configured to accept this equipment

constant change and, if the equipment is in (or possibly

when the equipment reaches) the appropriate state, the

recipe parameter is modified.