Manual

™
TFCalc
Thin Film Design Software for Windows
Version 3.4
© 1985-1999 Software Spectra, Inc. All rights reserved.
Software Spectra, Inc. • 14025 N.W. Harvest Lane • Portland, OR 97229 • USA
Telephone (503) 690-2099 • Fax (503) 690-8159
E-mail: [email protected]
Web: www.sspectra.com/support
blank page
Master Table of Contents
Windows Basics . . . . . . . . . . . . . . . . . . . . . .Section 1
Guide to TFCalc . . . . . . . . . . . . . . . . . . . . .Section 2
TFCalc Optimization Guide . . . . . . . . . . . .Section 3
TFCalc Examples . . . . . . . . . . . . . . . . . . . .Section 4
Information Sources . . . . . . . . . . . . . . . . . .Section 5
TFCalc Disk . . . . . . . . . . . . . . . . . . . . . . . . .Section 10

TFCalc, Copyright © 1985-1999 Software Spectra, Inc. All rights reserved.
Windows is a registered trademark of Microsoft Corporation.

Section 1 Windows Basics Page 1
Table of Contents
Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Using the Menus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Editing Data Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Editing Dialogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Configuring TFCalc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Page 2 Windows Basics Section 1
Getting Started
This manual assumes that you are somewhat familiar with the Windows operating system. We assume that TFCalc
already has been installed on your computer’s hard disk. The installer placed a shortcut to TFCalc on your desktop.
A screen similar to the one shown below should appear.
To start the program, move the cursor on top of the TFCalc icon and click the mouse button twice (with a very short
delay between the clicks). The program will start to load all the information it needs to operate.
Each copy of TFCalc has a unique serial number. If the serial number is less than 1521, then the following dialog
will appear the first time you run TFCalc:
The 9-digit access code is attached to the sleeve holding the TFCalc CD-ROM. Just enter the 9-digit code and click
OK. Unless you install TFCalc in multiple directories or on multiple computers, you should see this dialog only
once.
Using the Menus

Shown below is an example of TFCalc’s “menu bar”, which is the primary means of telling the
program what to do. Along the top of the screen are the names of the menu choices. To see the selections available
under each menu, move the cursor to one of the names in the menu bar, and click the mouse button. To select an
item from a menu, move the cursor to the item and click the mouse button again.
Note that all the menu names and each menu item have one letter underlined. These underlined letters are part of the
system for keyboard shortcuts: instead of using the mouse to select menu commands, you may use the keyboard. For
example, to select the Run menu, you would hold down the ALT key and press the R key; then to select the
“Analyze Only” command, you would press the A key.
Also note that some commands are followed by an “F” and a number. These commands can be selected by pressing
a function key.
Finally, note that you may obtain help by selecting the “Help” menu or by pressing the F1 function key. The TFCalc
Help file provides, except for the graphics, most of the same information as sections 1 and 2 of this manual. In the
TFCalc application, you may obtain help about any window or dialog by pressing the F1 function key when that
window or dialog is displayed.
Editing Data Windows

The figure below shows what we call a “data window.” These “windows” are the primary method of displaying and
editing data about a coating design.
This figure shows how the data about layers is displayed. Note that only four layers are shown here. However, it is
easy to move to the other layers:
• To move one layer at a time, move the cursor to one of the “scroll arrow” boxes in the bottom corners, and click
the mouse button.
• To move five layers at a time, click the mouse after moving the cursor into the area between the scroll arrows at
the bottom of the layer window.
• To quickly move to any section of five contiguous layers, move the cursor to the “scroll” box (somewhere
between the arrow boxes), press the mouse button, and, while continuing to hold down the button, move the
cursor to a new location. The scroll box will be “dragged” along. When you release the mouse button, the
position of the scroll box determines which layers will be displayed.
• To display fewer or more layers at one time, the size of the window can be changed by dragging the left or right
edges of the window.
To modify layer data, just move the cursor to the box you wish to change, and click the mouse button. The box will
reverse color. Type the new data. Numbers may be entered with an exponent: 3.14e-5 means 0.0000314. If you
make a mistake, the program will notify you. When you are done making a changes to a box, there are several ways
to tell TFCalc that you are done:
• Click the cursor in another box

• Press the Tab key, which selects the box to the right (or Shift-Tab to go to the left)
• Press the Return key, which selects the next box down (or Shift-Return to go up)
• Press the ESC key, which stops editing in the window
On your computer screen, you will note that the top-most (or “active”) window is the window that may be edited.
You can make a window active by moving the cursor to it and clicking the mouse. You can make a window
disappear by double-clicking the mouse inside the close box in the upper left corner.
Also note that when a window is active, the Options menu will display menu choices for that window. In particular,
this menu allows you to add, delete, or print the data in the window.
Editing Dialogs
The figure below shows an example of what is called a “dialog.” Dialogs are useful for entering small amounts of
data.
This dialog is for selecting how the electric field intensity should be computed. It contains four types of inputs:
• The boxes containing numbers are called “text boxes.” You may move from one to another by pressing the Tab
key or by clicking the mouse on the desired box.
• The circles in front of Low, Medium, and High are examples of “radio buttons.” Point and click the mouse at
the desired choice. Clicking on the text to the right of the circle also selects the radio button.
• The two marked squares are called “check boxes.” They are used for yes-no inputs. Point and click the mouse
on a check box to change its state. Clicking on the text to the right of the box also selects or deselects the check
box.
• The two rectangles at the bottom of the dialog are called “buttons.” After changing any of the data in the dialog,
you should click on the OK button to save the data, or on the Cancel button to disregard the data. Note that an
outline around the OK button indicates that this is the default button. On any dialog, you may select the default
button by pressing the Return or Enter keys. If a dialog has a Cancel button, a fast way to select it is to press the
ESC key.
Another type of input is illustrated in the Table Parameters dialog shown below.
Here, the box containing the down-arrow symbol is called a “pop-up menu.” By clicking the mouse in this box, a list
of choices will be displayed, as shown below.
To select an item from the list, move the mouse to the item and click the mouse button.
Configuring TFCalc
TFCalc allows coating designers to configure some aspects of the program to suit their needs. The configuration
dialog is shown below.
Specifically, the user may
• Change the units used for wavelengths. All wavelengths will be displayed using the selected units.
• Change the accuracy of the displayed results. An accuracy of “4” means there are 4 digits to the right of the
decimal point in all values displayed — a total of six digits.
• Indicate whether priority is given to optical or physical thickness. If the thickness priority is set to “Optical”,
then the optical thickness of layers remains constant when the reference wavelength or index of refraction is
changed. If the priority is “Physical,” then the physical thickness remains constant.
• Change the phase shift convention. There are two conventions for computing the P polarization of phase shift
on reflection. Using the Muller convention, the S and P polarizations always differ by 180° at normal incidence.
Using the Abelès convention, the S and P polarizations are the same at normal incidence.
• Change the field of view of the standard color observer. Select either 2° or 10°.
• Change how the QWOT (quarter-wave optical thickness) of absorbing materials is displayed in the Layers
windows. In the past, TFCalc displayed “n/a” for layers composed of an absorbing material. If the user selects
this option, the quantity 4 d n / λ0 is displayed, where d is the physical thickness of the layer and n is the real
part of the refractive index at the reference wavelength λ0.
• Indicate whether data tables should be extrapolated. If this is checked, then when a wavelength is outside of the
range of wavelengths in the data table, TFCalc will use the value at the closest endpoint. If this is not checked,
then TFCalc will warn the user when a wavelength is not within the range of data in a table.
• Indicate whether coating materials are allowed to have gain, which is the same as a negative extinction
coefficient (k < 0) in TFCalc.
These configuration settings are stored on the disk and read each time the user starts the program.
Section 2 Guide to TFCalc Page 1
Table of Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
File Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Edit Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Options Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Modify Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Environment Dialog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Stack Formula Dialog . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Layers Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Groups Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Targets (Discrete) Window . . . . . . . . . . . . . . . . . . . . . . . . 17
Targets (Continuous) Window . . . . . . . . . . . . . . . . . . . . . . 21
Comments Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Variable Materials Window . . . . . . . . . . . . . . . . . . . . . . . . 24
Environments Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Materials and Substrates Windows . . . . . . . . . . . . . . . . . 26
Illuminant Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Detector Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Distribution Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Run Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Results Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Miscellaneous Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Appendix A: Reading Data from a Text File . . . . . . . . . . . . . . . . 60
Appendix B: Needle/Tunneling Optimization . . . . . . . . . . . . . . . 61
Appendix C: Determining Refractive Index (N and K) . . . . . . . 63
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Page 2 Guide to TFCalc Section 2
Introduction
The flow diagram shown below expresses, in general, how to use the TFCalc program to design a thin film coating.
Let’s see how the diagram translates into actions taken by the designer.
open coating
file
modify coating display

design results
optimize print
coating results
analyze close coating

coating file
The process begins by opening a new or existing coating file (by selecting New or Open from the File menu). Next,
the designer modifies the coating design (by selecting from the Modify menu). Then the designer may optimize and/
or analyze the coating (by selecting from the Run menu). Now the results are viewed (by selecting from the Results
menu). The designer prints some results (by selecting from the Options menu while the results are displayed on the
screen). Finally, the coating design is saved for future reference (by selecting Save and then Close from the File
menu).
Although this is only one scenario, it shows how easy it is to use the program. Designers become familiar with the
program in just a matter of minutes, allowing them to focus their full attention on the design problem — instead of
the operation of the computer.
The Model
The diagram below details the physical system modeled by TFCalc.
Front Back
illum exit
inan medium
t
substrate
incident ...
medium
detector
layer 1
The illuminant, stored in the Illuminant database, is given as a table of spectral intensity versus wavelength. The
incident angle may vary from 0 to 89.999 degrees. The substrate and the incident and exit mediums are selected
from the Substrate database. The detector, given as a table of efficiency versus wavelength, is selected from the
Detector database. The thickness of the substrate, which is considered a massive layer, may be specified. Note that
the substrate could be, for example, air. Both the substrate and the exit mediums may be absorbing materials. The
materials that compose the front and back layers are selected from the Materials database (or from a name in the
Variable Materials window). Currently, there is a limit of 5000 layers. The optical properties of substrates and
materials are stored as tables or dispersion formulas of complex refractive index (n-ik) versus wavelength.
It is now possible to specify which surface the light encounters first: Front or Back. That is, light can come from
either side of the coating. See the Environment Dialog (page 10) for more details.
Note that layer 1 is next to the substrate.
Note: If the substrate and the exit media are different (or if there are back layers), reflections due to the back
surface of the substrate will be computed. If you wish to avoid this result, the exit medium must be the same
as the substrate and there must be no back layers.
Capabilities
The TFCalc program enables you to analyze and design multilayer thin film coatings. Some of its features are:
• Reflectance, transmittance, absorptance, optical density, loss, color, luminance, psi, phase shift on reflection or
transmission, and electric field intensity may be computed and plotted.
• The sensitivity of the coating to manufacturing errors may be analyzed. Optimization can be used to minimize
the sensitivity. Layer sensitivity can be computed and displayed.
• A coating can be analyzed for a cone of angles (as in a convergent beam of light) and a user-defined radiation
distribution.
• The coating materials, the substrate, and the exit medium may be dispersive and absorbing; the incident medium
may be dispersive. Dispersion formulas can be used to give the index of materials and substrates.
• The refractive index (n and k) of a substrate or layer can be determined from measured data.
• The illuminant may be specified. The reflectance or transmittance of a thin film may be stored as an illuminant,
so that the output of one filter can be made the input of another. Also, blackbody illuminants may be created.
• A detector function may be specified.
• The substrate may have a finite thickness; reflectance and transmittance calculations take into account the back
surface of the substrate and any attenuation within the substrate.
• A coating may have up to 5000 layers, composed of up to 150 different materials.
• “Stacks” of layers may be entered using a formula, such as H(LH)^5.
• Layers may be arranged in groups, and the groups may be optimized.
• The index (n and k) and thickness of a layer can be optimized.
• Layers on both sides of the substrate may be optimized simultaneously. During optimization, the layer
thicknesses can be constrained to be between a minimum and a maximum value.
• Global search may be used to locate the best design, rather than just the local minimum.
• Up to 5000 optimization targets may be specified. Also, a target can be an inequality, such as < 10% or > 90%.
• Multiple environments can be used to develop, for example, coatings for multiple substrates.
• Needle optimization can add layers to a design automatically, which is very useful if the design’s requirements
are unusual. The tunneling method can be used to automatically generate a sequence of optimal designs.
• Targets may be generated automatically. Also, targets may be read from files. Color and luminance targets can
be generated.
• First, second, and third derivatives (with respect to wavelength or wave number) may be used as targets.
• The equivalent index may be calculated. Also a layer may be replaced by an equivalent (HLH)^p or (LHL)^p
stack which matches the layer’s index.
• A choice of three local optimization methods are available: Gradient, Variable Metric, or Simplex.
• Data for an unlimited number of materials, substrates, illuminants, detectors, and distributions may be entered.
• The results of six thin film calculations may be compared by plotting them on the same graph. Different types
of plots may be overlaid: e.g., reflectance and transmittance.
• The minimum, maximum, and average values of a parameter (e.g., reflectance) can be computed for a range of
wavelengths.
• Optical monitoring curves can be computed and plotted.
• Results may be saved to text files for processing by other software.
File Menu
Use this menu to read, write, or print coating files. Below we briefly describe each menu item.
New Coating is for starting a coating file from scratch. If there is a coating file called DEFAULT in the same folder
as the TFCalc program, then that coating file will be read. Shortcut: Control-N.
Open Coating is for reading an existing coating file from the disk. The program will display the name of each
coating file and wait for you to select one. This is the standard Windows dialog for selecting a file. When TFCalc
begins, it displays this dialog. All coating files have the extension TFD. An example of the open coating dialog is
shown below. Shortcut: Control-O
Click the mouse on the name of the file you want to read, and then click the OK button.
Reopen Coating is used to open a file that has been used before. TFCalc keeps track of the last 20 coating designs
opened by the user. These files are displayed in a menu when the user selects this command.
Close Coating tells the computer that you are finished with the coating on which you are currently working. If you
have not saved the file, the program will ask if you wish to do so.
Save Coating is used to save the coating on which you are working. After a coating has been saved, you may
continue working on it. It is a good practice to occasionally save a coating while you are working on it; if you suffer
a loss of power, this practice will save your data. Shortcut: Control-S
Save Coating As is for saving a coating under a new name. You are asked for the new name. This command is
useful when you want to work with (or experiment on) several similar coatings. This is the standard Windows dialog
for saving a file. An example of this dialog is shown below.
Revert to Original is for going back to the previous copy of this coating: i.e., the coating you started with an Open
command, or the coating you lasted Saved. This command may be useful if you make changes to a coating, and the
new coating is not as good as the previous one.
Print Coating prints the current coating design on the printer. Shortcut: Control-P.
Printer Setup enables you to specify how text and graphics should be printed. You will see a dialog for configuring
your printer, such as shown below. Note that each type of printer has a different printer setup dialog.
Exit TFCalc stops the program. If the current coating has not been saved, you are asked whether you want to save
it.
About TFCalc gives some information about the program, including the serial number of your copy.
Edit Menu
The menu shows the standard Windows menu items for editing text. You may use these commands to edit text that
you have entered. Below we briefly describe each menu item.
Undo is not enabled in this program at this time. If you make an error, this menu item cannot help you!
Cut is for removing text you have selected. The text is put in the paste buffer.
Copy is for making a copy of the text you have selected. The text is put in the paste buffer.
Paste causes the contents of the paste buffer to be inserted at the current cursor position.
Clear causes the selected text to be erased.

Options Menu
This menu displays options available

when a particular window (e.g., Layers,
Targets, Plot) is the active window.
Below we briefly describe the first five
menu items. Because Option 1, Option
2, and Option 3 are customized for each
window, consult the appropriate section
of this manual for more details.
Add allows you to increase the number of entries in the active window. For example, to increase the number of front
layers in your coating, first select the Front Layers window and then select this menu item. The Add dialog is
illustrated below.
Delete allows you to decrease the number of entries in the active window. For example, to decrease the number of
front layers in your coating, first select the Front Layers window and then select this menu item. The Delete dialog
is shown below.
Print allows you to print the contents of the active window. For example, to print the layer data, first select the
Front Layers window and then select this menu item.
Copy and Append allow you to copy information from an existing coating file to the current coating. After this
command is selected, the standard dialog for selecting a file will appear.
Modify Menu
Use this menu to display and change the coating data. Below we briefly
describe each menu item. All of these items are covered in greater detail
on subsequent pages.
Environment causes the environment dialog to appear.
Stack Formula causes the stack formula dialog to appear.
Layers - Front displays the data about the layers on the front side of the
substrate.
Layers - Back displays the data about the layers on the back side of the
substrate. There may be a total of up to 5000 layers on both sides
(combined) of the substrate.
Groups displays the data on the groups. This option enables us to treat
groups of layers in a uniform way. For example, after putting the odd-
numbered layers in group 1 and the even layers in group 2, we could easily
increase the thickness of the layers in group 1 by 10%. There may be up to
5000 groups.
Targets - Discrete displays the discrete targets used to optimize a coating. A discrete target is an optimization target
at a single wavelength. The program aids you in designing a coating with the specified reflectance, transmittance,
absorptance, psi, density, color, luminance, or phase shift at up to 5000 targets.
Targets - Continuous displays the continuous targets used to optimize a coating. A continuous target is an
optimization target for a range of wavelengths. There may be up to 100 continuous targets.
Comments displays the comments about this coating.
Variable Materials displays the list of up to 150 materials whose index is allowed to vary.
Environments displays up to 10 environments that may be used during optimization.
Materials lets you display and modify the optical data for any number of coating materials. The names of materials
are displayed in alphabetical order.
Substrates lets you display and modify the optical data for any number of substrates. The names of substrates are
displayed in alphabetical order. The data for the incident and exit mediums (e.g., AIR and WATER) are stored with
the substrates.
Illuminants lets you display and modify data for any number of illuminants. The names of illuminants are displayed
in alphabetical order.
Detectors lets you display and modify data for any number of detectors. The names of detectors are displayed in
alphabetical order.
Distributions lets you display and modify data for any number of radiation distributions. The names of distributions
are displayed in alphabetical order.
Environment Dialog
When a new file is started or an existing file
is read from the disk, the program displays
the dialog shown here. This information is
called the coating environment. The
quantities in this dialog are briefly
described:
•Reference wavelength refers to the

wavelength used to specify the quarter-
wave optical thickness of the layers. If you
change the reference wavelength, and the
program is configured to give priority to
optical thickness, then the physical
thickness of all the layers will change.
Usually denoted by λ0.
• Incident medium is a name in the list of substrates. This name must be defined before it may be used. If it is an
absorbing material, TFCalc warns the user and assumes k=0.
• Illuminant is a name in the list of illuminants. This name must be defined before it may be used.
• As shown in the schematic, the incident angle is measured from a normal to the substrate.
• Substrate is a name in the list of substrates. This name must be defined before it may be used. The substrate is
treated as a massive (bulk) medium, which has no interference effect; only attenuation if it is an absorbing
material.
• Thickness is the physical thickness of the substrate, given in millimeters.
• Exit medium is a name in the list of substrates. This name must be defined before it may be used. If the
Substrate and the Exit mediums are the same (and there are no back layers), it means the substrate has infinite
thickness (with no reflections due to the back surface). If they are different, then multiple reflections are taken
into account. Also, when the Substrate and the Exit mediums are the same (and there are no back layers), the
computed transmittance is the transmittance into the substrate; if they are different, then the transmittance
through the substrate is computed.
• Detector is a name in the list of detectors. The name must be defined before it may be used. Specifying a
detector is useful when you are designing a filter for a detector, otherwise it is best to use IDEAL.
• First Surface is either Front or Back. “Front” indicates that light encounters the front layers first. “Back”
indicates that light comes from exit medium.
Select the OK button (or press the Return or Enter keys) when you are done entering the data. The Analysis
Parameters button leads to the Set Analysis Parameters dialog. If you select the Cancel button, any changes you
made will be ignored.
When more than one environment has been defined, then an environment can be selected using the Environment
pop-up menu. The selected environment is called the current environment. For more information, see the section
about the Environments Window (page 25).
Stack Formula Dialog
This dialog allows you to enter a stack using a formula. The formula, which may be up to 32000 characters long,
may contain up to 26 one-letter symbols, such as the L and H shown above. Use the “<<” and “>>” buttons for
additional symbols. Note that the substrate is next to the first layer. The syntax of a stack formula is very simple:
• Any sequence of symbols may be placed next to each other, e.g., ABCCCBA.
• Symbols may be grouped using parentheses, e.g., (ABC).
• The right parenthesis must always be followed by a caret (^) and an integer, e.g., (ABCBA)^10.
• A multiplier may precede any symbol or group of layers, e.g., 1.2 (0.5 A B 0.5 C)^2.
• Group factors may be used: G1 (HL)^5 G2 (HL)^5 means the first ten layers are in group 1 and the second ten
layers are in group 2.
• Spaces may be used anywhere for clarity, e.g., (1.5 A 2.3 B 0.7 C)^5 (ABC)^2.
After you have entered the formula, the meaning of the symbols should be entered. For each symbol used in the
formula, you need to specify the material it represents, its thickness, whether the layer should be optimized, and to
which group it belongs. If the thickness priority (set in the configure dialog) is optical, then the thickness value is
assumed to be QWOT, otherwise physical thickness in nanometers. You may also follow the thickness by “qw” or
“nm” to indicate the thickness units.
If you plan to use groups, be sure to define them in the Groups window before entering the formula.
It is best to use the Tab key to move between the boxes.
When you select the Generate Layers button, the program uses the stack formula to change the contents of the Front
Layers window. If you select the Cancel button, no changes are made. If you have substantially changed the design
in the Layers windows, you may want to use the Clear Formula button to erase the stack formula so that you do not
accidentally replace the Front Layers with the stack formula.
Layers Windows
There are two Layer windows: front layers and back layers.
These windows show information about the layers that make
up the coating. For each layer, you must specify:
• The material, which must either be in the material database or be defined in the Variable Material window.
• Either the quarter-wave optical thickness or the physical thickness. Note that only one thickness needs be
entered for each layer; if the optical thickness is entered, then the physical thickness will be computed, and vice
versa. If a layer consists of a material having a significant absorption at the reference wavelength, it is not
possible to compute the usual quarter-wave optical thickness. In this case, the program displays either “n/a” or
an approximate value in the QWOT row. See the section about configuring TFCalc (section 1, page 6) for more
details.
• Whether you want to optimize the layer.
• The group with which the layer is associated. If you plan to use groups, they should be defined first, using the
Groups window.
• Whether you want to constrain the thickness of the layer.
• The minimum and maximum thickness, given in either nanometers (nm) or quarter-wave optical thickness (qw).
• Whether needle layers are allowed to be inserted in the layer during needle optimization.
When the Layers window is the top-most window, you may use the Options
menu to perform various operations. In addition to allowing you to add, delete,
print, copy, and append layers, this menu has ten other options, which are
described next.
Reverse Layers is a command that reverses the order of the layers. Also note that there is a command on the Misc
menu to swap the front and back layers.
Adjust Layers for Angle is a command that changes the thickness of all non-absorbing layers. The thickness is
changed so that the effective thickness at the new angle remains the same. The dialog is below. Note that this dialog
always begins with 0.0 for both angles. This feature is also called angle matching.
Consolidate Layers is a command that removes layers having zero thickness, which may occur after a coating
design has been optimized. Also, if there are adjacent layers of the same material and the same group, this command
will combine the layers.
Compute Equivalent Index is a command that calculates the equivalent (Herpin) index of a range of layers. The
dialog for front layers is shown below.
The range of layers being analyzed must form a symmetric stack. That is, the thickness of the first layer must be the
same as the last layer, the thickness of the second layer must be the same as the next-to-the-last layer, etc. The
command computes the index and phase thickness (at normal incidence) of a layer that is equivalent to the range of
layers at the given wavelength and angle. Note that the equivalent index is undefined if the wavelength is inside a
“stop band” for the range of layers.
Compute Equivalent Stack is a command that calculates an HLH or LHL stack that is equivalent to a given layer.
The dialog for front layers is shown below.
If the incident angle is not 0, then the polarization must be selected; the equivalent stack will match only the selected
polarization. Select whether an HLH or LHL stack is desired. Enter the names of the materials to be used; they may
be in either the materials database or the Variable Materials window. When the Calculate button is selected, the
equivalent stacks are computed. There are actually an infinite number of solutions; we display the ones whose layers
have a phase thickness less than 360 degrees (4 QWOT at normal incidence). The arrow buttons may be used to
scroll through the solutions. The solutions are sorted by increasing thickness. (QWOTs are given at the reference
wavelength.) If the Replace button is selected, the displayed solution will replace the given layer in the Layers
window.
Show Thickness Totals is a command for displaying the physical thickness of each material used in the coating and
the total physical thickness of all the layers. The thinnest layer is displayed at the bottom of the dialog. The dialog
for the front layers is shown below.
Read Layer Data from File is a command that reads a text file containing two columns: material name and
physical thickness (in nm). A standard dialog will ask the user for a file name. This is a simple way of reading a
coating design into TFCalc.
Save Layer Data in File is a command that writes the material name and physical thickness (in nm) of each layer to
a text file. This command may be useful in creating a run sheet for the coating. A standard dialog will ask the user
for a file name.
Save Layer Data in Clipboard is a command that copies the material name and physical thickness (in nm) of each
layer to the clipboard. This command may be useful in creating a run sheet for the coating.
Substitute Layer Data is a command that enables the user to many data items quickly. The dialog, shown below,
works as follows: it finds all layers that match all the conditions in the first column, then it changes the fields of
those layers to the values given in the second column. If the first column is all blank, then it matches all layers. In
the first column, the QWOT and Thickness items can be inequalities, such as < 10.
Groups Window
The figure above shows how the group data is displayed. For each group, you must enter:
• The group “Factor” — a number that multiplies the thickness of each layer in the given group. This factor
always starts at 1.0. The true thickness of a layer can be computed by taking the thickness in the Layers window
and multiplying it by the factor of that layer’s group.
• Whether to allow the group factor to be changed during optimization.
• Whether to constrain the group factor during optimization.
• The minimum and maximum group factors, if constrain is “yes”.
When the Groups window is the top-most window, you may use the Options menu to perform various operations. In
addition to allowing you to add, delete, print, copy, and append groups, this menu has another option:
Normalize Groups is a command that multiplies the thickness of each layer by the group factor of the layer’s
group, stores the product as the new thickness of the layer, and then sets all the group factors to 1.0. So, for example,
to increase the thickness of each layer in group 1 by 10%, we would change the factor of group 1 to 1.1 and then use
this command.
Targets (Discrete) Window
The figure above shows how the data about the discrete optimization targets is displayed. Targets must be entered
before performing any optimization. For each discrete target, you must specify:
• Whether it is an intensity or a phase target by entering either I or P (there are also codes for color, luminance,
and derivative targets, but it is easier entering them using the Add Color Targets and Generate Derivative
Targets commands on the Option menu).
• The quantity the user wants to optimize. Commonly, it is reflection (R) or transmission (T). It can also be
absorption (A), psi (P), or density (D). The product of reflectance and transmittance can be entered also. Just
enter R*T, Rp*Ts, or Rs*Tp, where the p and s suffix indicates the polarization. If R*T is entered, then the
window’s polarization row can be used to select whether Rs*Ts, Rp*Tp, or Rave*Tave is optimized.
• The polarization of the target. Use P or S for the pure polarizations; A for average polarization; and D is for the
difference P-S. For example, if the target is reflectance, then average polarization means 0.5*(Rp + Rs) and the
difference is Abs(Rp - Rs).
• The target wavelength (in the units you have chosen with the Configure dialog)
• The incident angle (in degrees), measured in the medium in which the light originates
• The target value, whose units depend on the type of target. Use percentage for intensity targets such as
reflectance and transmittance; angle in degrees for phase and psi targets. The product R*T is allowed to be
between 0 and 100. Any target can be an “inequality” target; that is, the value may be preceded by “>” (or “<”)
to indicate that any result greater than (or less than) the target is acceptable.
• The tolerance. The inverse of the tolerance is usually called the weighting factor. See the merit function
(page 36) to determine how this tolerance is used. Although a tolerance of 1.0 generally works well, you may
decrease it to force the optimization method to reduce the difference between the target value and computed
value. Note that the tolerances are relative; if all the tolerances are the same, then the effect is the same as when
they are all 1.0.
• The environment. For typical coating designs, this value will be 1. However, if multiple environments have
been defined (see page 25), then you may enter the number of the environment here.
When the Discrete Targets window is the top-most window, you may use the Options menu to add, delete, print,
copy, and append targets. There are also options to generate targets and to read targets from a file.
Generate Targets is a command that enables you to create many similar targets. The dialog is shown below.
Most of the items in the dialog are the same as in the target window. The target can be an “inequality” target; that is,
the value may be preceded by “>” (or “<”) to indicate that any result greater than (or less than) the target is
acceptable. Note that you may specify a range of wavelengths: a beginning wavelength, an ending wavelength, and
an increment. If you change the “Number of targets” item, the increment will be adjusted so that the given number
of targets is generated. The generated targets may be evenly spaced by wavelength, wave number, or the logarithm
of wavelength. The last two options may be useful for optimizing broadband designs. If the “Replace current
targets” check box is checked, then the generated targets will replace all the targets in the Targets window;
otherwise the generated targets will be appended to the current list of targets. If you do not wish to generate any
targets, select the Cancel button; otherwise choose OK.
Read Wavelength Targets from File is a command that enables you to read targets from a file. The dialog is
shown below.
Again, most of the items in the dialog are the same as in the Target window. The text file containing the targets may
begin with any number of lines of nonnumeric data (comments, for example). Subsequent lines of text must consist
of a wavelength and a target value. If the file contains more data points than you want to use as targets, you can
command TFCalc to read only a subset of the data. (Some spectrophotometers create files with data points spaced at
0.1 nm. In that case, you may want only every 10th, 20th, or 100th data point.) Each data point may also be
multiplied by an adjustment factor as it is read. If the “Replace current targets” check box is checked, then the
targets from the file will replace all the targets in the Target window; otherwise the targets will be appended to the
current list of targets. If you do not wish to read any targets, select the Cancel button; otherwise choose Get File.
Read Angle Targets from File is another command that enables you to read targets from a file. The dialog is very
similar to the one shown above; the only difference is that it asks for the wavelength instead of the angle. Each line
of the text file containing the targets must consist of an angle and a target value.
Add Color Targets is a command for adding color and luminance targets to the Targets window. The dialog below
shows all the color parameters.
Most of the items in the dialog are the same as in the target window. Select either the CIE 1931 or the CIE 1976
standard. Color may be computed by selecting wavelengths spaced either every 10 nm or every 5 nm. Choosing the
latter may cause the optimization to take twice as long.
Note that if a design is optimized using color targets spaced every 10 nm, but the final design is analyzed with a 5
nm (or smaller) interval, then the color coordinates may differ slightly. If this result presents a problem, do a final
optimization using color targets spaced by 5 nm.
Generate Derivative Targets is a command for adding derivative targets to the Targets window. The dialog is
shown below.
Note that this dialog is exactly the same as the “Generate Targets” dialog, except for the derivative information in
the upper right corner.
Targets (Continuous) Window
The figure above shows the data about the continuous optimization targets. Either discrete or continuous targets
must be entered before performing any optimization. For each continuous target, you must specify:
• Whether it is an intensity or a phase target by entering either I or P.

• The quantity the user wants to optimize. Commonly, it is reflection (R) or transmission (T). It can also be
absorption (A), psi (P), or density (D). The product of reflectance and transmittance can be entered also. Just
enter R*T, Rp*Ts, or Rs*Tp, where the p and s suffix indicates the polarization. If R*T is entered, then the
window’s polarization row can be used to select whether Rs*Ts, Rp*Tp, or Rave*Tave is optimized.
• The polarization of the target. Use P or S for the pure polarizations; A for average polarization; and D is for the
difference P-S. For example, if the target is reflectance, then average polarization means 0.5*(Rp + Rs) and the
difference is Abs(Rp - Rs).
• The beginning and ending wavelengths (in the units you have chosen with the Configure dialog)
• The incident angle (in degrees), measured in the medium in which the light originates
• The beginning and ending target values, whose units depend on the type of target. Use percentage for intensity
targets such as reflectance and transmittance; angle in degrees for phase and psi targets. The product R*T is
allowed to be between 0 and 100. Any target can be an “inequality” target; that is, the value may be preceded by
“>” (or “<”) to indicate that any result greater than (or less than) the target is acceptable.
• The tolerance. The inverse of the tolerance is usually called the weighting factor. See the merit function
(page 36) to determine how this tolerance is used. Although a tolerance of 1.0 generally works well, you may
decrease it to force the optimization method to reduce the difference between the target value and computed
value. Note that the tolerances are relative; if all the tolerances are the same, then the effect is the same as when
they are all 1.0.
• The environment. For typical coating designs, this value will be 1. However, if multiple environments have
been defined (see page 25), then you may enter the number of the environment here.
When the Continuous Targets window is the top-most window, you may
use the Options menu to add, delete, print, copy, and append targets.
Note: Be careful mixing discrete and continuous targets; if the

tolerances are the same, then each discrete target has the same weight
as a continuous target.
Comments Window
You may edit the comments by clicking the mouse inside the Comments window. To begin adding text at a certain
location, click the mouse at that location. To end your editing, press the Enter key or click the mouse in another
window. These comments are saved in the coating file.
The text on the first line of the Comments window will be displayed as a “remark” on some windows and printouts.
When the Comments window is the top-most window, you may use the Options menu to copy or append comments
from another coating file or to print the comments.
Variable Materials Window
The figure above shows how the data about variable materials is displayed. Variable materials, unlike the materials
in the materials database, are part of the coating design; that is, you may have different variable materials for each
coating. You may use the name of a variable material wherever a material name can be used. For each variable
material, you must specify:
• The name of the variable material. Each variable material must have a unique name. Also, the name may not be
the same as a material in the material database. To use a variable material in a layer, just enter its name in the
Material row of a Layer window.
• The initial index (N and K) of the variable material.
• The minimum and maximum index (N) that the variable material may have. These limits are enforced only
when the variable material is being optimized.
• Whether the optimization method should vary the index (N) of the variable material.
• The minimum and maximum index (K) that the variable material may have. These limits are enforced only
when the variable material is being optimized.
• Whether the optimization method should vary the index (K) of the variable material.
Note: the index is not varied by the Simplex optimization method.
When the Variable Materials window is the top-most window, you may use the Options menu to add, delete, print,
copy, and append variable materials.
Variable materials are useful for experimenting with rugate or graded-index coatings — that is, coatings consisting
of two materials whose mixture is varied as the coating is produced. This type of coating may be simulated by a
number of thin layers, each having an index slightly different from the previous one. For example, to simulate a
coating whose index varies linearly (with thickness) from 1.5 to 2.5, you could enter 11 variable materials with
indices 1.5, 1.6, 1.7,..., 2.5 and then use those materials in 11 layers of equal thickness in the Front Layers window.
Variable materials can also be used to determine the index of unknown layers. In this case, the optimization targets
would be replaced by spectrophotometric or ellipsometric data. See Appendix C (page 63) for details.
Another use of variable materials is to find the ideal index of one or more layers; then the equivalent stack
computation could be used to replace the variable material layer with a stack of HLH or LHL layers (see page 14).
Environments Window
The figure here shows how the data about environments is
displayed. Note that the current environment, which is used
when a coating is analyzed, is marked with an asterisk in the
top row. Multiple environments are important for designing
coatings whose performance must be controlled in different
conditions (e.g., different illuminations, different incident or
exit mediums, different substrates, or different detectors). If a
coating contains layers whose refractive index changes due
to some outside influence (e.g., temperature, light intensity,
electric current, magnetic fields, moisture), then this
window’s active materials may to used. Multiple
environments also may be used to design coatings for
birefringent substrates — that is, substrates whose refractive
index depends on incident angle and polarization. As in the
Environment dialog (page 10), for each environment, you must specify:
• Incident medium is a name in the list of substrates. This name must be defined before it may be used. If it is an
absorbing material, TFCalc assumes k=0.
• Illuminant is a name in the list of illuminants. This name must be defined before it may be used.
• Substrate is a name in the list of substrates. This name must be defined before it may be used. It is treated as a
massive (bulk) medium, which has no interference effect; only attenuation if it is an absorbing material.
• Thickness is the physical thickness of the substrate, given in millimeters.
• Exit medium is a name in the list of substrates. This name must be defined before it may be used.
• Detector is a name in the list of detectors. The name must be defined before it may be used. Specifying a
detector is useful when you are designing a filter for a detector, otherwise it is best to use IDEAL.
• Up to three active materials -- materials whose index changes due to an outside influence.
• The first surface encountered by the light: Front or Back. Generally, this is Front, meaning that light originates
in the incident medium. However, as shown in the window above, it is possible to define environments for light
originating on both sides of a coating.
When active materials are used, each environment represents a different setting of the outside influence. For
example, if the outside influence is temperature, then each environment could represent a different temperature.
Then, for each environment, the name of the active material changes. Be sure that at least one of the layers in the
current environment uses the active materials listed under the current environment column (marked with an
asterisk). If you change the current environment, then TFCalc will automatically substitute the active material into
the layers windows, maintaining the physical thickness of each layer.This will be done automatically during
optimization if you use multiple environments in the optimization targets.
When the Environments window is the top-most window, you may use the Options menu to add, delete, print, copy,
and append environments.
Materials and Substrates Windows
The windows above show how the names of the available materials and substrates are displayed. To see the data for
a particular material or substrate, click on the name. When one of these windows is the top-most window, the
Options menu may be used to add, delete, and print the names of materials and substrates.
When the “Add Material” option is selected, the following dialog appears:
The new name must be entered. Refractive index data may be given as a table of data points or as a dispersion
formula. If the latter is selected, then the formula type must be selected for the n and k parts of the complex
refractive index. When OK is pressed, a blank table or formula will appear.
An example of a table of refractive index data for a material is shown below.
The optical data consists of up to 1001 points. When the program needs the value at a wavelength not in the table, it
uses linear interpolation to estimate the value; if the wavelength is outside the range of wavelengths, then the value
at the closest endpoint is used. The program keeps the data sorted by wavelength.
For all materials, the index N must be greater than zero. The extinction coefficient K is zero for non-absorbing
(dielectric) materials. For absorbing materials, K > 0. For material exhibiting optical gain, K < 0. Note that TFCalc
must be setup to handle gain materials in the Configure dialog.
The dialog below shows an example of how a dispersion formula is edited.
Dispersion formulas are valid only for a certain range of wavelengths. The range, along with the formula
parameters, is entered into the dialog. When the Print button is pressed, the dispersion formula parameters are
printed. When the Comment button is pressed, the user is able to edit the comment for this formula. When the
Convert button is pressed, the user may convert the dispersion formula to a table of data points. For materials (and
now substrates), the Fit Data button allows the user to find a dispersion formula that fits measured data. See
Appendix C (page 63) for more information.
When the user presses OK, extensive checks are applied to the dispersion formulas to insure that the formulas
compute a valid refractive index for the entire wavelength range.
For the dispersion formula n(λ), the following choices are available:
Name Dispersion Formula n(λ)
Sellmeier 1
Sellmeier 2
Sellmeier 2´
Sellmeier 3
Cauchy
Hartmann 1
Hartmann 2
Schott Glass
Drude
For the dispersion formula k(λ), there are four choices:
Name Dispersion Formula k(λ)
Zero
Sellmeier
Exponential
Drude
Note: the Schott Glass company now uses the Sellmeier 3 dispersion formula to characterize its glasses. However,
the Schott formula is still used by other glass manufacturers.
Note that the Drude formulas cannot be mixed with the other formulas; when you select Drude for n(λ), then TFCalc
automatically selects Drude for k(λ). Also note that n and k are given implicitly; TFCalc uses a procedure for
determining n and k separately. The Drude formulas are useful for modeling the dispersion of metals for
wavelengths in the infrared and far-infrared.
If the user changes the index of refraction of a material, the thickness of any layer using that material may be
affected. The program automatically adjusts the thickness; if you have given priority to optical thickness, then the
physical thickness will be recomputed, and vice versa. Note: you must close the window for the changes to take
effect in the current coating design.
When a window of material or substrate data is the top-most window, the Options menu may be used to add, delete,
and print the data. A short comment can be made about each material or substrate. There is also an option to read
material and substrate data (including internal transmittance data) from a text file created by the user. This option is
described in Appendix A (page 60). The user may also write the material data to a text file, making it easy to use this
data in other software.
The data about each material and substrate is stored in a separate file. For example, there is a file called MGF2 in the
MATERIAL directory and a file called GLASS in the SUBSTRAT directory. This feature makes it much easier to
make a copy of the data, to share the data with a colleague, or to move the data to the Macintosh version of this
software. NOTE: if you rename or create new files, be sure that the file name consists of at most 8 characters and
that the extension is correct; material and substrate data files use the MAT extension.
Illuminant Windows
The window above shows how the names of the available illuminants are displayed. When this window is the top-
most window, the Options menu may be used to add, delete, and print the names of illuminants.
To see the relative intensity of a particular illuminant, click on the name. The data for “WHITE” light is shown
below.
Intensity is given as a percentage. Linear interpolation determines the intensity at wavelengths not given in the table;
for a wavelength outside the range of wavelengths, the value at the closest endpoint is used. There may be up to
1001 data points for each illuminant. With the “Save as Illuminant” command (page 57) in the Results menu, it is
possible to store the reflectance or transmittance of a thin film coating as an illuminant. Note: you must close the
window for the changes to take effect in the current coating design.
When a window of illuminant data is the top-most window, the Options menu may be used to add, delete, and print
the data. A short comment can be made about each illuminant. There is also an option, described in Appendix A
(page 60), to read illuminant data from a text file. The user may also write the illuminant data to a text file, making it
easy to use this data in other software.
Normalize Illuminant is a command that scales the illuminant data so that the maximum intensity of the illuminant
is 100%.
Create Blackbody Illuminant is a command that enables the user to create an illuminant having a blackbody
spectrum of a given temperature. The illuminant is normalized so that the highest intensity is 100%. The self-
explanatory dialog is shown below.
The data about each illuminant is stored in a separate file. For example, there is a file called WHITE in the
ILLUMINA directory. This feature makes it much easier to make a copy of the data, to share the data with a
colleague, or to move the data to the Macintosh version of this software. NOTE: if you rename or create new files,
be sure that the file name consists of at most 8 characters and that the extension is correct; illuminant data files use
the ILL extension.
An illuminant may be displayed as a curve by creating a coating file which uses the IDEAL detector, AIR as the
incident and exit media, AIR as the substrate, and which does not have any front or back layers. You must enter the
name of the illuminant in the Environment dialog and then analyze the coating. The Plot window will display the
illuminant curve.
Detector Windows
The window above shows how the names of the available detectors are displayed. When the window is the top-most
window, the Options menu may be used to add, delete, and print the names of detectors.
To see the relative efficiency of a particular detector, click on the name. The data for the “IDEAL” detector is shown
below.
Note that the efficiency is given as a percentage. The program uses linear interpolation to determine the efficiency at
wavelengths not given in the table; if the wavelength is outside the range of wavelengths, then the value at the
closest endpoint is used. There may be up to 1001 data points for each detector. Note: you must close the window
for the changes to take effect in the current coating design.
When a window of detector data is the top-most window, the Options menu may be used to add, delete, and print the
data. A short comment can be made about each detector. There is also an option to read detector data from a text file
created by the user. This option is described in Appendix A (page 60). The user may also write the detector data to a
text file, making it easy to use this data in other software.
Normalize Detector is a command that scales the detector data so that the maximum efficiency of the detector is
100%.
The data about each detector is stored in a separate file. For example, there is a file called IDEAL in the
DETECTOR directory. This feature makes it much easier to make a copy of the data, to share the data with a
be sure that the file name consists of at most 8 characters and that the extension is correct; detector data files use the
DET extension.
A detector may be displayed as a curve by creating a coating file which uses the WHITE illuminant, AIR as the
incident and exit media, AIR as the substrate, and which does not have any front or back layers. You must enter the
name of the detector in the Environment dialog and then analyze the coating. The Plot window will display the
detector curve.
Distribution Windows
User-defined radiation distributions can be used in the Cone-Angle Average computation (see page 41 and page 47).
The window above shows how the names of the available radiation distributions are displayed. When the window is
the top-most window, the Options menu may be used to add, delete, and print the names of distributions.
To see the relative intensity of a particular distribution, click on the name. The data for the “EQUAL” distribution is
shown below.
Note that the intensity is given as a percentage. The program uses linear interpolation to determine the intensity at
angles not given in the table; if the angle is outside the range of angles in the table, then the value at the closest
endpoint is used. There may be up to 1001 data points for each distribution. Note: you must close the window for the
changes to take effect in the current coating design.
When a window of distribution data is the top-most window, the Options menu may be used to add, delete, and print
the data. A short comment can be made about each distribution. There is also an option to read distribution data
from a text file created by the user. This option is described in Appendix A (page 60). The user may also write the
distribution data to a text file, making it easy to use this data in other software.
Normalize distribution is a command that scales the distribution data so that the maximum intensity of the
distribution is 100%.
The data about each distribution is stored in a separate file. For example, there is a file called EQUAL in the
DISTRIBU directory. This feature makes it much easier to make a copy of the data, to share the data with a
be sure that the file name consists of at most 8 characters and that the extension is correct; distribution data files use
the DST extension.
Run Menu
Use this menu either to analyze or to optimize your coating. Note

that each of the seven commands in the top half of the menu has a
corresponding “set parameters” command in the bottom half of the
menu. Below we briefly describe each menu item.
Analyze Only commands the program to compute the reflectance, transmittance, absorptance, psi, density, loss, and
phase shift of the current coating; no layer’s thickness is changed. This command is useful when you modify some
parameters and want to see the effect. A small status dialog, as shown below, gives the progress of the analysis.
Note that you may quit the analysis at any time by selecting the Quit button.
The results may be displayed by selecting “Show Plot” or “Show Table” from the Results menu. Use the “Set
Analysis Parameters” command (page 42) on this menu to specify the range of wavelengths or angles for which the
coating is analyzed.
Optimize Design tells the program to vary the thickness of the layers or the group factor of groups (and the index of
variable materials) so that the following merit function is minimized:
m 1/k
1 I jD jC j – T j k 
F =  ----
m
∑ --------------------------- 
N j Tol j 
j=1
where m is the number of targets, k is the power of the method, I is the intensity of the illuminant, D is the efficiency
of the detector, T is the desired target value, C is the computed value (of reflectance, transmittance, etc. at the target
wavelength, angle, and polarization), Tol is the tolerance for a target, and N is the normalization factor for the target.
Although this formula looks complex, in most cases, I = D = Tol = 1.0.
For continuous targets, the summation above is replaced by a sum of integrals.

The quantity
I jD jC j – T j
---------------------------
Nj
is called the “deviation from target.” The purpose of the normalization factor N is to convert the units to a
compatible scale. The normalization factor for reflectance, transmittance, absorptance, and luminance targets is 1.0;
for phase targets it is 1.8; for density targets it is 0.09; and for color targets it is 0.01.
The number k, termed the power of the method, can have a significant effect on the optimization results. The
program allows the following values for k: 1, 2, 4, 8, 16, and Max. As the value of k is increased, the larger
deviations will be emphasized, forcing the optimization methods to equalize the deviations of the targets. Note: if
the gradient or variable metric optimization method is used, a power of 1 should be used only if the targets are at
their extreme values (e.g., reflectance 0 or 100%).
When k is set to Max, the merit function is defined to be the maximum of the deviations:
I jD jC j – T j
F = Max ---------------------------
N j Tol j
j
Because this function is not smooth, only the Simplex method works when k = Max.
Use the “Set Optimization Parameters” command (page 43) on this menu to select whether layers or groups are to be
optimized, the optimization method to be used, the number of iterations, etc.
When layers are being optimized, only the layers (and variable materials) you have selected to be optimized will be
changed. If there are layers on both sides of the substrate, then layers on both sides are optimized simultaneously.
When groups are being optimized, only the groups (and variable materials) you have selected to be optimized will
be changed.
The figure above shows the status dialog that indicates the progress of the optimization. The dialog shows two
graphs. The top graph displays the number of layers, the total physical thickness of the layers and the index profile
— the index and relative thickness of each layer. The incident medium, the substrate, and the exit medium are
shown as heavy horizontal lines. The bottom graph — the target deviations bar chart — shows the value of the
“deviation from target” for each discrete target. When continuous targets are used, then the bar chart becomes a
series of curves; one curve for each continuous target. Internally, TFCalc automatically converts each continuous
target into a series of discrete targets; the total number of discrete targets is displayed on the chart. This result is
illustrated below.
The bars and curves for average, P, and S polarization are displayed in different colors. Note that continuous targets
capture the fine detail in the behavior of the deviation. When both discrete and continuous targets are used, then you
will see a bar chart and the curves side-by-side. “Iteration” is the current iteration (or step) of the optimization
process. The number called “Deviation” is just the value of the merit function. “Change” indicates how much the
design has changed from the previous iteration; this quantity approaches zero as the optimization method converges
to the local minimum of the merit function.
Optimization continues until either the maximum number of iterations is attained, the solution is found, or the Quit
button is selected. After optimization stops, the Analyze and Continue buttons become highlighted. Selecting the
Continue button causes the optimization to continue. Selecting the Analyze button computes the reflectance,
transmittance, absorptance, psi, density, loss, and phase shift of the optimized coating. Note that you may quit the
optimization at any time by selecting the Quit button; the best design will be displayed. There is also a “delayed”
quit capability; if you are doing needle optimization, which involves a series of local optimizations, you may want to
quit after the current local optimization is complete. To do this, hold down the Control key while clicking the Quit
button. The button’s caption will change to “Quit*” to indicate that quitting will be delayed.
Global Search enables the designer to search for a global minimum rather than a local one. The “Set Global Search
Parameters” command (page 45) on this menu allows the user to control various aspects of the global search. While
global search is working, the following dialog is displayed:
Designs Generated counts all designs generated, including those that have been excluded from the global search.
Designs Accepted counts only those designs that have been included in the global search. Possible Minimums
counts the number of possible minimums found; this number can increase and decrease as the global search
progresses. The list of Best Deviations shows the deviations of the designs that are possible minimums. Under Status
there is short message indicating what function is being performed. The global search can be stopped at any time by
clicking on the Stop Searching button.
At the end of the search (or after the Stop Searching button has been selected), the other buttons become active.
Select the Done button to exit from the global search dialog. Warning: pressing the Done button does not save any
of the designs found by global search.
Select the Save Best Designs button to save the designs that are possible local minimums. The designs will be saved
with names {prefix}001, {prefix}002, etc. The dialog shown below will appear.
If the Optimize Designs button is selected, the following dialog appears:
This command gives the designer the capability of optimizing many designs simultaneously; either the gradient or
variable metric method can be used.
Compute EFI commands the program to compute the (squared) electric field intensity within the coating. Use the
“Set EFI Parameters” command (page 46) on this menu to specify the wavelength and the layers you want to
analyze. While the EFI is computed, the dialog shown below is displayed. The computation may be halted by
selecting the Quit button.
The results may be displayed by selecting “Show EFI Plot” from the Results menu.
Compute Sensitivity commands the program to determine how sensitive the coating design is to small
manufacturing errors (also called tolerancing). Use the “Set Sensitivity Parameters” command (page 46) described
below to set the expected error and the number of trials you wish to perform. The program will vary the thicknesses
of layers randomly, analyze the coating design, and record the results. The sensitivity computation selects
thicknesses that are uniformly or normally distributed within the error range. While the sensitivity is computed, the
dialog shown below is displayed. The computation may be halted by selecting the Quit button.
The results may be displayed by selecting “Show Plot” from the Results menu. Shown below is an example of a plot
produced with this command.
The heavy center line represents the original coating design; the two outer-most lines above it and below it represent
the minimum and maximum (that is, the worst-case) variation as the layer thicknesses are varied randomly. When
the “quartile” sensitivity analysis is selected in the “Set Sensitivity Parameters” dialog, then three additional curves
are computed and displayed: the first, second, and third quartiles. The second quartile is the median performance
(which is usually very close to the performance of the coating design). The first and third quartile curves are usually
slightly below and above the median performance; at each wavelength, the performance of half of the random
designs lie between the first and third quartile curves.
This command computes the sensitivity of all quantities (reflectance, transmittance, absorptance, psi, density, loss,
and phase shift) at S, P, and average polarization. To see another quantity’s sensitivity, use the “Set Plot Parameters”
command in the Results menu. If you find a coating design that is very sensitive to small errors, the design may be
improved by using the Minimize Sensitivity option in the “Set Optimization Parameters” dialog (page 43).
Compute Cone-Angle Average command is useful for analyzing a coating when the incident radiation is in the
form of a cone. Use the “Set Cone-Angle Parameters” dialog described on page 47 to specify the cone. The cone
axis may be normal or nonnormal to the coating. The coating is analyzed at a number of equally-spaced angles
specified by the user. The result (i.e., reflectance, transmittance, etc.) at each angle is weighted according to the
proportion of the radiation incident at that angle and the intensity of the radiation at each angle. While the cone-
angle average is computed, the dialog shown below is displayed. The computation may be halted by selecting the
Quit button.
The result of this calculation may be displayed by selecting “Show Plot” or “Show Table” from the Results menu.
Note: this command computes the proportion of transmission or reflection due to each polarization. However,
because we are computing the average of a 3-dimensional bundle of rays, there is no actual P or S
polarization. It is best to look at just the average polarization.
Compute Monitor Curve command enables the user to simulate the output of a light monitor as one or more layers
are deposited on a monitor chip. Use the “Set Monitoring Parameters” command described on page 47 to select the
layers and the monitoring wavelength. While the curve is being computed, the dialog shown below is displayed. The
computation may be halted by selecting the Quit button. The results may be displayed by selecting “Show Monitor
Chart” from the Results menu.
Compute Layer Sensitivity command lets the user determine which layers are most sensitive to manufacturing
errors. It is generally used only after a design has been completely optimized. It produces a bar chart like the one
shown here:
To understand this graph, there are two cases to consider: (1) When computing the layer sensitivity of a completely
optimized design, the first-order sensitivity is very close to zero. In this case, the second-order sensitivity indicates
the steepness of the “valley” in which this design lies. If a layer has a large second-order sensitivity, it means that a
small change in that layer’s thickness will lead to a large increase in the value of the merit function. (2) When this
computation is applied to a non-optimized design, then the second-order sensitivity is undefined because the design
is not at the bottom of a “valley.” In this case, the first-order sensitivity shows how the merit function increases or
decreases as thickness increases. The two buttons allow the user to save the numerical results to a file or to save the
chart to the clipboard.
Set Analysis Parameters enables you to enter the range of wavelengths and/or angles over which the coating
should be analyzed. The dialog is shown below.
Choose between studying a coating design for a range of wavelengths (at the given angle), a range of angles (at the
given wavelength), or for a range of wavelengths and angles; click the desired button.
For a range of wavelengths: enter the first wavelength, the last wavelength, and the increment between wavelengths.
The coating design may be analyzed at up to 32000 wavelengths. The incident angle is measured from a normal to
the substrate.
For a range of angles: enter the first angle, the last angle, and the increment between angles. The coating design may
be analyzed at up to 32000 angles.
For ranges of both wavelengths and angles, enter data about both ranges. The product of the number of wavelengths
and the number of angles may not exceed 32000.
Set Optimization Parameters enables you to select the number of iterations performed by the optimization method,
the accuracy to which it is to work, and the power in the merit function. The dialog is shown below.
There are three optimization methods available: Gradient, Variable Metric, and Simplex. The Gradient and Variable
Metric methods require the derivative of the merit function to be calculated; as the method approaches the
minimum, the derivative approaches zero and becomes less accurate. Hence the Gradient method will usually be
stopped by inaccuracies in the derivative. At that point, the Simplex method, which does not use derivatives, may be
used to improve the solution. These methods will find only a local minimum of the merit function.
Note that each iteration of the Gradient and Variable Metric methods will produce an improvement in the deviation;
this is not true for the Simplex method. Also, the Simplex method may need many more iterations to converge to a
solution. If N layers or groups are being refined, we suggest 5*N iterations for the Gradient and Variable Metric
methods and 10*N*N iterations for the Simplex method. Reducing the error tolerance will increase the number of
iterations needed to find a solution. In the above dialog, Max. Change is used only by the Simplex method, which
uses it to start its iterations. If you are near a minimum, it should be set to a small number (e.g., 1%).
The Variable Metric method usually locates a minimum faster than the Gradient method. However, the Gradient
method requires much less memory to operate, which may be an important consideration if many layers are being
optimized. Note that, if these methods are started far from a minimum, they may converge to different designs.
If the Minimize Sensitivity option is selected, the optimization method will use the parameters in the Set Sensitivity
Parameters dialog (see page 46) and attempt to minimize a merit function that allows random thickness variations.
For the best results, Number of Trials should be fairly large. However, a large number of trials also causes the
optimization to run more slowly.
Use the Stop Optimizing Zero-thickness Layers option if you want to eliminate a layer as soon as its thickness
becomes zero. If this option is not selected, then zero-thickness layers continue to be refined, possibly growing to a
useful thickness. However, it is often the case that very thin layers oscillate between zero and a small number, which
reduces the efficiency of the optimization method.
To watch the index profile during optimization, select the Display Index Profile While Optimizing check box.
The Needle Parameters button leads to the dialog shown below, which controls the behavior of the “needle”
optimization method, which is described in Appendix B (page 61).
If the Add “Needle” Layers box is checked, then the needle method will be used the next time the user optimizes the
design. If the Continue Adding Layers Until Optimal Design Is Found box is checked, then the needle method will
be used repeatedly until no more layers can be added. In this case, the design is said to be optimal. If the Add Many
Layers Simultaneously box is checked, then the needle method will attempt to add many zero-thickness layers
before the local optimization begins. This option works best when you know that a design (such as a bandpass or
edge filter) will require many layers. If you want to insure that the design is locally optimized before needles are
added, the Use Local Optimization Before Adding Needles option can be selected. This option is useful after you
have manually stopped an optimization and then want to resume optimizing.
When the Continue Adding Layers box is checked, then the Tunneling Parameters button becomes enabled. When
pressed, this button leads to the following dialog:
The tunneling method, described in more detail in Appendix B (page 62), searches for a better design after an
optimal solution has been found by the needle optimization method. The tunneling method is initiated only if (1) the
“Grow the coating” check box has been checked, (2) the total thickness of the coating is less than the amount given
in the dialog above, and (3) the deviation (i.e., the value of the merit function) is greater than the value given in the
dialog. Also note that there is an option to save each optimal design before it is changed by the tunneling method.
Set Global Search Parameters enables the user to control the global search process. The dialog shown below will
appear.
Global searching works by generating many designs (either randomly or systematically) and determining which of
those designs may be near a local minimum of the merit function. The best local minimum is called the global
minimum.
The designs are generated by varying the layer thicknesses (or group factors) of the current design; only those layers
(or groups) that have been selected for optimization will be varied. If a design uses variable materials and they have
been selected for optimization, then those indices will be varied during global search also.
In the dialog above, use the Number of Designs to Search to specify how many designs to search; the maximum
number is 32767.
What range of designs will be generated by global search? That is determined by (1) which layers (or groups or
variable materials) have been marked for optimization in the current design and (2) the percentage in the Maximum
Change from Current Design field. If the maximum change is 50%, then the thickness (group factor or index) is
allowed to vary ±50% from its current value. Thicknesses (and group factors and indices) are prevented from
becoming negative. If layer thickness (or group factor or index) constraints are being used, the global search will
stay within those constraints.
A systematic search is recommended only if a few layers (or groups or variable materials) are being optimized. The
systematic search generates designs that are uniformly spaced. For example, if the thicknesses of two layers are
being varied and 100 designs are being generated, then 10 equally-spaced thicknesses will be generated for each
layer; the result is 10 x 10 = 100 designs.
During a random search, designs are generated randomly; the thickness of each layer (or the group factor of each
group) being varied is picked from a uniform distribution. For example, if a layer being varied is 100 nm thick and
the design is being allowed to change by 50%, then random thicknesses will be picked to be between 50 and 150
nm.
Global search locates the best designs by creating a map of all the designs. For each new design the search
generates, it (1) uses the merit function to compute the deviation and (2) finds the other generated designs, called
neighbors, that are closest to it. If all of the design’s neighbors have larger deviations, then that new design is called
a possible local minimum. By changing the Number of Nearest Neighbors field in the dialog, the designer controls
how many of these local minimums will be found; if many neighbors are used, then fewer minimums will be found;
if only a few neighbors are used, then more (and possibly false) minimums will be found. In general, the number of
neighbors should be increased as more layers (or groups or variable materials) are being varied.
As designs are generated, it may be useful to exclude some of them from the global search. This exclusion is
accomplished by checking the Exclude Designs Having Thicknesses > or the Exclude Designs Having Deviation >
check boxes. The first excludes designs having a physical thickness greater than the specified number; the second
excludes designs that are far from a local minimum.
Set EFI Parameters enables you to control the EFI computation. You may select the wavelength of interest, the
layers to be analyzed, the resolution, and whether the EFI should be computed in the incident medium and substrate.
In high-resolution mode, the EFI is computed at 400 points. For medium and low resolution, 200 and 100 points are
computed, respectively. The dialog is shown below.
Set Sensitivity Parameters enables you to control the sensitivity computation. The dialog is shown below.
Note that errors in layer thickness and/or index may be simulated; check the “Vary Thickness by” box and/or the
“Vary Index by” box. The distribution of layer thicknesses and indices can be either uniform or normal (Gaussian).
The thickness of the layers may be varied either relatively or absolutely (by either physical or optical thickness);
click on the option you want and enter the desired error amount. The index error is always a relative amount. Select
the analysis type; worst-case analysis uses Monte Carlo simulation to estimate the minimum and maximum variation
of the coating’s performance; quartile analysis uses Monte Carlo simulation to estimate the quartiles of the
performance distribution (see page 40); the worst-possible analysis uses optimization to compute the worst possible
performance for a given manufacturing error. The number of trials is used only in the two Monte Carlo simulations.
In general, if a coating has more layers, it will take more trials for the sensitivity to converge to its minimum and
maximum values. If the Vary All Layers check box is checked, the thickness of every layer is varied; otherwise,
only those layers marked for optimization will be varied. If the Compute Color Sensitivity check box is selected,
then the color of each trial design will be plotted in the Color window.
Note: for the normal distribution, all generated thicknesses are within two standard deviations of the nominal
design. For the normal deviation, the numbers in the Relative and Absolute boxes are assumed to be the
standard deviation of the layer thicknesses.
Set Cone-Angle Parameters enables you to specify the cone of radiation for the cone-angle average computation.
The dialog is shown below.
The cone's central angle (the angle the cone's axis makes with the normal to the coating) is always the incident angle
entered in the Environment dialog. The angular size of the cone may be entered as either the half-angle or the f/
number. To calculate the cone-angle average, the coating is analyzed at the number of angles entered in the last box.
This number must be odd. The computation becomes more accurate as this number increases; doubling the number
of angles reduces the error by a factor of four. The user can select from either two built-in radiation distributions
(equal intensity or Lambertian) or a user-defined distribution (see page 34).
Set Monitoring Parameters command enables the user to specify the parameters used to compute the monitor
curve. The dialog shown below will be displayed.
Either reflectance, transmittance, delta, or psi may be monitored. When monitoring reflectance or transmittance,
then either P, S, or average polarization may be monitored. The curve can be computed at either low, medium, high,
or a specified (nm) resolution. Any number of layers can be monitored. The layers may be monitored at up to three
wavelengths. The chip substrate is a name that must be in the list of substrates. If the back of the monitoring chip
reflects light, be sure to check the box. The thickness of the chip is important only if the chip absorbs. The monitor
angle is the angle at which the light monitor source strikes the chip.
Each material may be a different monitor ratio, which may be changed by pressing the “Set Monitor Ratios” button.
The following dialog will appear:
The user may choose to give all the materials the same monitor ratio, or to give each material a different ratio. The
monitor ratio is proportional to the square of the ratio of the rack distance to the monitor distance.
Results Menu
Use this menu to display the results produced by the Run menu.
Note that each of the five commands in the first part of the menu
have a corresponding “set parameters” command in the second
part of the menu. Below we describe each item.
Show Table displays a table of the (1) reflectance, transmittance,

absorptance, psi, density, or loss or (2) phase shift on reflection or transmission. For example, see the figure below.
The table may be printed via the Options menu. The number of decimal places may be controlled with the
“Configure” command in the Misc menu.
Note that the quantity displayed here is actually the modified reflectance, which is the product of reflectance, the
illuminant intensity, and the detector efficiency. To obtain the actual reflectance, transmittance, absorptance, loss,
and optical density, use the WHITE illuminant and the IDEAL detector. Psi and the phase shift due to reflection and
transmission are not affected by the illuminant or detector.
When a coating design is analyzed for ranges of angles and wavelengths, then this window displays tables at one
angle (or wavelength) at a time. Press the left and right arrow keys to display other angles (or wavelengths).
This window can be resized to show more or fewer lines of results. As a shortcut, you may double-click in this
window to bring up the Set Table Parameters dialog (see page 54). Note that when this window is top-most, the
Options menu has commands to print the table, copy the table to the clipboard, save the table in a file, and to
calculate statistics.
When the Calculate Statistics option is selected, the dialog shown below appears.
When the user clicks in the Calculate button, it displays the minimum, maximum, and average of the parameter
being displayed in the Table window. The user may change the polarization and the wavelength range over which
the calculation is done. If the coating has been analyzed for ranges of wavelengths and angles, then this dialog will
display two additional text boxes for the range of angles.
Show Plot displays a plot of the (1) reflectance, transmittance, absorptance, psi, density, or loss or (2) phase shift on
reflection or transmission. It is possible to overlay the plot of one parameter on another. Here is a plot of the optical
density and phase shift on transmission.
When a coating design is analyzed for ranges of angles and wavelengths, then this window displays plots at one
angle (or wavelength) at a time. Press the left and right arrow keys to display other angles (or wavelengths). To
display the performance curves for all angles (or wavelengths), hold down the Shift key press the left or right arrow
key. This Shift-key technique also works for printing and for copying to the clipboard.
When the substrate has a finite thickness, then the reflectance and transmittance calculations take into account the
reflections due to the back surface; an incoherent light source is assumed. By selecting the “Set Plot Parameters”
menu item described on page 55, it is easy to:
• Display the reflectance, transmittance, absorptance, psi, optical density, or loss.

• Display the phase shift on reflection or transmission.
• Display the Average, S, and P polarizations.
• Change the vertical and horizontal scale of the plot.
• Choose the horizontal axis units from the following: wavelength, relative wave number, wave number, electron
volts, or frequency.
• Display the optimization targets (discrete targets as small circles and continuous targets as thin trapezoids).
Average, P, and S polarization targets are displayed in different colors.
• Display up to five previously computed results on the same plot.
• Overlay one type of plot on another.
The “remark” on a plot comes from the first line of the Comment window; if the first line is blank, then no remark is
displayed.
Note: The phase-change and psi computations always assume the substrate has infinite thickness; reflections
due to the back surface are ignored.
As in the Table window, to obtain a plot of the actual reflectance, transmittance, absorptance, loss, and optical
density, use the WHITE illuminant and the IDEAL detector. Psi and the phase shift due to reflection and
transmission are not affected by the illuminant or detector.
As the cursor is moved across the plot window, the location of the cursor is displayed in the plot header. If the
mouse button is held down while the cursor is moved, the distance moved is displayed. If the shift key is held down
when the mouse button is released, the statistics about the selected wavelength range will be displayed in a dialog. If
the Control key is held down while an area on the plot is selected by dragging the mouse, then that area will be
zoomed. You can unzoom by holding down the Control key and clicking in the plot. As a shortcut, you may double-
click in this window to bring up the Set Plot Parameters dialog (see page 55).
When the plot window is the top-most window on the screen, there are options to print the plot, copy it to the
clipboard, and calculate statistics. For information about the Calculate Statistics command, see the description of the
Show Table command on page 50.
Show EFI Plot displays the results of the electric field intensity computation. For example, see the plot below.
Note that the substrate, the medium, and the layers are identified. The relative physical thickness of the layers are
shown. The EFI computation always assumes the substrate has infinite thickness; reflections due to the back
surface are ignored. As a shortcut, you may double-click in this window to bring up the Set EFI Plot Parameters
dialog (see page 55).
As the cursor is moved across the EFI plot, the location of the cursor is displayed in the plot header. The x-
coordinate is the distance (in nm) from the top surface of the coating. If the mouse button is held down while the
cursor is moved, the distance moved is displayed. Using the Options menu, the plot may be printed or copied to the
clipboard, or the numerical results may be saved to a text file.
Show Color Chart displays the color coordinates of light reflected or transmitted by the coating. It uses the results
of the last analysis and the spectral tristimulus values for the 1931 CIE standard observer (two or ten degree field,
depending on which value is selected in the Configure dialog) to calculate the color coordinates. The results of this
calculation are displayed as a 1931 CIE or 1976 CIE UCS chromaticity diagram or as LAB coordinates in a window
as shown below.
For best results, the coating should be analyzed from 380 nm to 780 nm (or wider) in 5 nm increments. The
excitation purity and the dominant and complementary wavelengths are also displayed. If the chart is displayed on a
color monitor, the approximate color of the reflected and/or transmitted light appears in two small squares. As a
shortcut, you may double-click in this window to bring up the Set Color Chart Parameters dialog (see page 56).
When a coating design is analyzed for ranges of angles, then this window displays the color at all angles (as shown
above). Press the left and right arrow keys to display the numerical values at each angle.
As the cursor is moved across the color chart, the location of the cursor is displayed. If the mouse button is held
down while the cursor is moved, the distance moved is displayed. If the Control key is held down while an area on
the chart is selected by dragging the mouse, then that area will be zoomed. You can unzoom by holding down the
Control key and clicking in the chart.
Using the Options menu, the chart may be printed or copied to the clipboard, or the color sensitivity results may be
saved to a text file. In the text file, the chromaticity coordinates (X,Y,Z) are written, tab-delimited, for reflectance
and/or transmittance (depending on what is displayed in the chart). The first line of the file contains the color of the
nominal design.
Show Monitor Chart displays the monitoring curve produced by the Compute Monitor Curve command. The chart
displays either the reflectance, transmittance, delta, or psi as one or more layers of the design is deposited on the
monitoring chip. The top of the graph displays the layer number.
As a shortcut, you may double-click in this window to bring up the Set Monitor Chart Parameters dialog (see
page 56).
As the cursor is moved across this chart, the location of the cursor is displayed in the chart header. If the mouse
button is held down while the cursor is moved, the distance moved is displayed. If the Control key is held down
while an area on the chart is selected by dragging the mouse, then that area will be zoomed. You can unzoom by
holding down the Control key and clicking in the chart.
Using the Options menu, the chart may be printed or copied to the clipboard, or the numerical results may be saved
to a text file.
Set Table Parameters controls what is displayed in the Table window. The dialog is shown below.
If dB is selected, then the parameter will be displayed in dB units; that is, 10 Log(x), where x is reflectance,
transmittance, etc. If previous results were saved, then those results may be displayed in the Table window by
clicking in the Show Saved Results check box. Note that this method may be used to display the results of the
sensitivity calculation (1 = worst-case minimum, 2 = worst-case maximum, 3 = first quartile, 4 = third quartile, and
5 = median performance).
Set Plot Parameters controls what is displayed in the Plot window. The dialog is shown below.
Most of the inputs are self-explanatory. To overlay two types of displays, click on the Display Overlay box. This
will enable the Overlay Parameters button. When the Overlay Parameters button is clicked, the parameters for the
overlay will be displayed and the name of the button will change to Primary Parameters. When the overlay
parameters are displayed, note that the plot type (wavelength or wave number) may not be changed. If the plot range
of the overlay is different from the primary plot, then the Plot window will have a second scale to the right of the
plot (as shown in the description of the Show Plot command). Minor Ticks refers to the number of extra grid lines
between the numbered grid lines.
Set EFI Plot Parameters controls how the EFI results are displayed. The dialog is shown below.
Set Color Chart Parameters controls how the Color Chart is displayed. The dialog is shown below.
Most of the inputs are self-explanatory. The Reference White, which must be a name in the illuminant list, is used to
compute the excitation purity and the dominant and complementary wavelengths. The Color Sensitivity check box
may be checked if color sensitivity was selected in the Sensitivity Parameters dialog.
Set Monitor Chart Parameters controls how the monitor chart is displayed. The dialog shown below will appear.
Most of the inputs are self-explanatory. Minor Ticks refers to the number of extra grid lines between the numbered
grid lines.
Save Results #1 and Save Results #2 and Save Results #3 and Save Results #4 and Save Results #5 save the
results of the last analysis of the coating. These commands enable you compare these results with subsequent
calculations; just check the “Display saved results” box shown in the “Set Plot Parameters” dialog. When results are
saved, a check mark will appear in front of the selected command.
Remove Saved Results eliminates the saved results, which can occupy a large amount of memory, from the
computer’s memory.
Save as Illuminant enables you to save the reflectance or transmittance of a coating as an illuminant. The following
dialog is displayed:
Choose the result to be saved and type the name you wish to give this new “illuminant.” If previous results were
saved, then those results may be saved by clicking in the Show Saved Results check box. Note that this method may
be used to save the results of the sensitivity calculation as an illuminant (1 = worst-case minimum, 2 = worst-case
maximum, 3 = first quartile, 4 = third quartile, and 5 = median performance).
Miscellaneous Menu
Use this menu for some miscellaneous commands. Below we briefly describe each menu item.
Swap Front-Back Layers switches the front layers to the back and the back layers to the front. This command is
useful when you want to see how a filter behaves when light goes through it in the opposite direction. It is also
useful when you want to copy front layers from an existing coating to the back layers of the current design. Just
swap the front and back layers, copy the front layers from a file, and swap the front and back layers again. The same
procedure can be followed for using the stack formula to put layers on the back side of the substrate.
Compute Color from Data File command allows the user to compute the color of reflectance or transmittance data
read from a text file. The following dialog collects information about the data:
When the Get File button is pressed, the user will be asked for the name of the text file. The data file should consist
of two columns: wavelength and percent reflectance (or transmittance). The wavelengths must be equally-spaced
with an increment not exceeding 10 nm.
This command is available only when no coating design file is open. Be sure that the wavelength units in the text
file are the same as those in the Configure dialog.
Configure allows the user to (1) change the units used for wavelengths, (2) change the accuracy of the displayed
results, (3) indicate whether priority is given to optical or physical thickness, (4) change the phase shift convention,
(5) change the field of view of the color observer, (6) indicate how QWOT should be displayed in the Layers
windows, (7) indicate whether data tables should be extrapolated, and (8) indicate whether gain materials are
allowed. See the section about configuring TFCalc (section 1, page 6) for more details.
Window Options command allows the user to indicate (1) that the size and position of windows should be saved by
TFCalc and (2) which windows TFCalc should open automatically when it opens a coating design file. The dialog is
shown here:
If the Plot or Table window is selected, then the coating design is analyzed automatically before the window is
displayed.
Appendix A: Reading Data from a Text File

It is possible to read material, substrate, illuminant, detector, and distribution data from text files produced in other
programs. This feature makes it easier to import data from other computers and to import data created from
formulas.
The file type of the file being read must be TEXT. Most text editors and spreadsheet programs are capable of
producing such files. Text editors usually call this a “text-only” file. Spreadsheet programs may call them “tab-
delimited” files.
Materials and Substrates. Each line of a file of index data must have a format such as
500.0 1.38 0.001
where the first number is the wavelength, the second is the index of refraction (n), and the third is the extinction
coefficient (k) or the internal transmittance. If k is 0, then it does not have to appear on the line. There may be up to
1001 lines of index data; each wavelength should appear only once in the file. The wavelengths do not have to be
sorted. Sometimes you may have data for n at some wavelengths and data for k at other wavelengths. Just use -1 to
indicate the missing data and TFCalc will interpolate to find the n or k at the missing wavelengths. As example, for
the data
500.0 1.4 0.1

550.0 -1 0.2
600.0 1.6 -1
650.0 1.7 0.4
TFCalc would use interpolation to determine n at 550 and k at 600.
Illuminant and Detectors. Each line of illuminant or detector data must have a format such as
500.0 45.56
where the first number is the wavelength and the second is the intensity or efficiency (a number between 0 and 100).
There may be up to 1001 lines of illuminant or detector data; each wavelength should appear only once in the file.
The wavelengths do not have to be sorted.
Distributions. Each line of distribution data must have a format such as
10.0 45.56
where the first number is the angle in degrees and the second is the intensity (a number between 0 and 100). There
may be up to 1001 lines of distribution data; each angle should appear only once in the file. The angles do not have
to be sorted.
When a material, substrate, illuminant, detector, or distribution data window is the front-most window, the Option
menu displays an option to read the data from a file. The data in the file will replace all the data in the open window.
Appendix B: Needle/Tunneling Optimization

The Needle Optimization method1 works by adding new zero-thickness layers into a coating design. After the new
zero-thickness layer has been added, one of the local optimization methods (gradient or variable metric) is used to
improve the new design. If the zero-thickness layer has been placed in the correct position, the local optimization
method will force the new layer to grow.
The needle optimization method, as implemented in TFCalc, works with up to eight materials. There must be at least
two materials. The coating design can contain layers of other materials, but only the eight materials are involved in
the needle optimization. The needle method examines the coating design and decides where to position the new
layer. It is also possible for the needle method to add no new layers.
Note that the needle method: (1) does not work with the Simplex method, (2) does not work with group
optimization, and (3) does not work with the sensitivity optimization.
Suggestions for Use:
• This method may improve existing designs, but it seems to work best when it either (1) starts with a single layer
or (2) is used in conjunction with the tunneling method.
• If starting with a single layer, the method seems to work best when the contrast between the substrate and layer
is greatest. For example, if SiO2 and TiO2 are coated on a glass substrate, make the starting design a single
layer of TiO2 (which has an index much higher than SiO2 and glass).
• The initial thickness of the single layer is very important. A thin single layer will lead to designs consisting of
only a few layers; a thick single layer can become a design with many layers. A good strategy is to try the
needle method on several single-layer designs of increasing thickness. Alternatively, the tunneling method (see
the next page) can be used to generate many optimal designs of increasing thickness.
• It is a good idea to check the “Stop Optimizing Zero-Thickness Layers” box in the “Set Optimization
Parameters” dialog when the needle method is used.
• Use a liberal number of optimization targets. It slows down the optimization, but you are more likely to get the
design you want.
• After optimization, the needle method may have left some very thin layers in the design. To remove them from
the design, set their thickness to zero and use the “Consolidate Layers” command on the Layer Window’s
Option menu. Then re-optimize the design (with needle optimization turned off).
For more suggestions, see section 3: TFCalc Optimization Guide.
1. For details see A.V. Tikhonravov, “Some theoretical aspects of thin-film optics and their applications,”
Applied Optics 32, 5417-5426 (1993).
The Tunneling method seeks to solve two difficulties in using the needle optimization method: (1) deciding what
the initial design should be and (2) deciding what to do if the design found by needle optimization is not good
enough. By solving these two problems, the tunneling method greatly simplifies the optimization process: now the
designer need only enter the optimization targets and a single thin layer; needle optimization and tunneling will
generate a sequence of designs whose performance approach the optimization targets more and more closely. In fact,
if inequality targets (see page 17) are used, then the needle/tunneling method will stop when all of the inequalities
have been attained.
Why is it called the “tunneling” method? You can imagine the optimization problem as finding the “valleys” of a
multi-dimensional function. This method seeks to find a tunnel from the current valley (i.e., the current optimal
design) to a better valley (where a better design may be found). The contour plot below shows this process for a
two-layer AR design.
The arrow shows that the tunneling method has decided to tunnel from one valley to another by growing the H layer.
After the new valley has been located, then local optimization is used to find the bottom of that valley. Then needle
optimization is used to add layers to the design. When needle optimization has found an optimal design, tunneling is
employed again.
Note that when an optimal design contains many layers, there are many possible tunnels to other valleys. The
tunneling method considers all the tunnels and selects (in some sense) the best one. There is no general rule that
describes which layer will be grown.
Appendix C: Determining Refractive Index (N and K)

The problem of determining the refractive index of a thin film is difficult for several reasons:
• A thin film layer may be inhomogeneous.

• Thin film properties may change with time.
• Thin film properties may depend on how the layer was deposited.
• Reflectance and transmittance measurements are typically accurate to only ±0.5%.
• If the thickness of the thin film layer is known imprecisely, there may be a number of different index-thickness
combinations that have similar reflectance/transmittance measurements.
In TFCalc, there are two methods of determining the reflective index of a layer:
• If ellipsometric measurements are available, or if you assume that the index is constant over a wavelength
range, then you can use a variable-index material (see page 24).
• Alternatively, you can fit a dispersion formula (see page 27) to measured data.
Using variable-index materials. In this case, TFCalc is used to find the design that matches the measured data.
Here is the procedure:
1. Create a new coating design file.

2. Set the incident and exit mediums and the substrate in the Environment dialog (see page 10).
3. Create a variable-index material (see page 24), and enter your best guess as to what N and K are. Also indicate
whether N and/or K is being optimized.
4. Add a layer to the Layers-Front window (see page 12) and change the material name to the name of the
variable-index material you created in step 3. Set the layer’s thickness to the estimated thickness of the thin film
that was measured. If the thin film’s thickness is not known accurately, set Optimize to Yes. It is also a good
idea to use the constraints so that the thickness stays within some limits.
5. In the Targets window (see page 17), enter the measurements taken from the thin film. If the measurements are
in a text file, you may use the “Read Wavelength Targets from File” option (page 19) or “Read Angle Targets
from File” option (page 19) to read that data into the Targets window. Note that measurements at any
combination of angle, wavelength, and polarization may be entered.
6. Before you optimize this design, save it to disk. Now you are ready to optimize the design.
7. If you think there is only one possible solution, then use the “Optimize Design” command (page 36) on the Run
menu. The local optimization method should converge to a solution fairly fast. You know that a valid solution
has been reached if all the deviations are small — on the same order as the error in the measured data.
8. If you believe that there may be multiple solutions, then use the “Global Search” command (page 39) on the
Run menu. You may want to change the global search parameters (page 45) to increase the number of designs
to search and to do a systematic search. After some possible global minimums have been found, use the
“Optimize Designs” button to converge to accurate solutions. Use the “Save Best Designs” button to save the
results of the global search.
Note that you can use this method with more than one layer also. However, for this method to converge to a valid
solution, the measured data must be quite accurate.
Fitting a dispersion formula. In this case, the first step it to create a new material or substrate (see page 26) and to
select a type of formula that is appropriate. In most cases, the low-quality of the measured data dictates that the
simplest formulas — those with the fewest parameters — be used. Use a more complex dispersion formula if your
data is more accurate. When the dispersion formula dialog (see page 27) appears, enter some approximate values for
the parameters and then press the “Fit Data” button. The following dialog will appear for materials:
(For fitting substrate data, the dialog is very similar; only the substrate name and the estimated layer thickness are
not used.) This is one of the most complex dialogs in TFCalc. Using this dialog, you enter information about the
measured data, which will be used to fit the dispersion formulas. Note that one, two, or three data files may be
specified. The top part of the dialog allows you to enter information about each data file:
Measurement (T, R, B). This is the kind of measurement that the data file contains. T stands for transmission, R for
reflectance, and B for reflectance from the back side of the substrate. If this input is blank, then the data file will not
be used. In the data files, all measurements must be given in percentages. Also, the same wavelengths must be used
in every data file.
Incident angle (deg). This is the angle of incidence at which the measurement was taken.
Polarization (Ave, P, S). This is the polarization of the measured data.
Back surface reflects? Sometimes the back surface of a substrate is coated to improve the measurement. For
example, for R measurements, the back surface is sometimes coated with a black paint so that there is no reflection
from the back surface.
Name of data file. You may select a data file by clicking the mouse in the box that says “click here.” After the file
has been selected, “click here” is replaced by the name of the file.
The bottom part of the dialog contains information common to the data files:
Wavelength unit. This is the wavelength unit that is used in the data files. Very important.
Substrate thickness. This is important only if the substrate absorbs.
Substrate name. This name must be in TFCalc’s Substrate list. The index data for the substrate must be accurate; an
inaccurate substrate index will lead to inaccurate results. (Not used in fitting substrate data.)
Est. layer thickness (nm). The estimated thickness of the thin film layer should be entered here. Because there may
be many solutions to this data fitting problem, the accuracy of the thickness is important. (Not used in fitting
substrate data.)
Use every _th data point. Quite often there are too many data points in a data file created by a spectrophotometer.
This entry tells TFCalc to use a subset of data. Up to 1001 data points may be used from each data file.
Constraints. Press this button if you want to control which parameters should be optimized and to constrain those
parameters to a particular range of values. By default, the layer thickness is not varied; use this button to allow the
layer thickness to vary. When this button is pressed, a dialog similar to the one below will be shown; there will be a
line for each parameter in the dispersion formulas.
For each parameter, select the “Opt?” check box if you want to optimize that parameter; select the “Limit?” check
box to constrain the parameter value between two limits; then enter the minimum and maximum values. For most of
the dispersion formulas, there are “natural” constraints, which TFCalc enforces so that division by zero does not
occur. Note that some of these natural constraints are determined by the current parameter values and by the range
over which the dispersion formula is being fitted. Press the Restore Natural Constraints button to replace the current
constraints with the natural ones.
Pressing the OK button returns to the previous dialog. After entering all this information it is a good idea to use the
“Save Info” button to save this information. Then, if you want to use this information again, use the “Get Saved
Info” button. Select the “Compute” button to start the data fitting calculation.
As the data fitting calculation proceeds, the dialog shown below displays the progress. For each iteration of the
process, the dialog will change. The top part of the dialog contains the latest dispersion parameters.
The middle of the dialog shows a bar chart for each data file. Each bar shows the difference between the
measurement (T, R, or B) and the computed value using the latest dispersion parameters. The vertical axis has units
of percent. The numbers at the bottom of the bar chart are used to number the measurements.
The bottom of the dialog shows the current iteration, the deviation (in percent), the computed thickness of the thin
film layer, and a status message. As the fitting process continues, the deviation will decrease. At any time, you may
press the “Stop” button to stop the iterations; it will stop after the current iteration ends.
After the iterations are done (or if you have pressed the “Stop” button), the “Accept” and “Reject” buttons will
become active. If the “Reject” button is pressed, then the results will not be saved.
How does one decide whether the iterations have converged to a dispersion formula that accurately fits the measured
data? Let us suppose that the measurements are known to have an accuracy of ±X%. Although X will depend on the
equipment used, values of 0.5% are probably typical. There are two indications that the iterations have converged to
an accurate dispersion formula: (1) the final deviation is less than X and (2) all the bars on the bar chart are in the
±X range. If the iterations do not converge to a dispersion formula that accurately fits the measured data, there may
be several reasons for this:
• The measurements are not accurate enough.

• The type of dispersion formula does not model the actual dispersion very well.
• The range of wavelengths is too wide.
• The initial dispersion formula parameters are not close enough to the actual parameters. It usually helps if the
parameter values are constrained to a range of values.
• The estimated thickness is not correct.
• The layer being measured is too thin. Experience indicates that it helps if the measured data has at least one
peak or valley; if the data is only increasing or only decreasing, then there are usually many parameter values
that fit the data.
• There is not enough information. For some materials, it helps if measurements can be made at different angles
and polarizations.
Index Page 67
A Edit Illuminant Comment ..............................2-31

Edit Material Comment .................................2-29
Edit Substrate Comment ................................2-29
Accuracy ....................................................... 1-6, 2-58 Environment ....................................................2-9
Active materials ................................................... 2-25 Environments...................................................2-9
Angle Matching.................................................... 2-13 Exit TFCalc......................................................2-6
Angles, Range of.................................................. 2-42 Front Layers...................................................2-12
Generate Derivative Targets ..........................2-20
B Generate Targets............................................2-18
Global Search.................................................2-39
Groups..............................................................2-9
Birefringent substrate ........................................... 2-25 Illuminants .......................................................2-9
Blackbody Illuminant........................................... 2-31 Layers - Back...................................................2-9
Layers - Front ..................................................2-9
C Materials ..........................................................2-9
New Coating ....................................................2-5
Normalize Detector........................................2-33
Capabilities............................................................. 2-4 Normalize Distribution ..................................2-35
Cauchy dispersion formula .................................. 2-28 Normalize Groups..........................................2-16
Chromaticity coordinates ..................................... 2-53 Normalize Illuminant.....................................2-31
CIE Color ............................................................. 2-53 Open Coating ...................................................2-5
Coating Environment ........................................... 2-10 Optimize Design ............................................2-36
Color coordinates ........................................ 2-53, 2-58 Paste.................................................................2-7
Color observer......................................1-6, 2-53, 2-58 Print..................................................................2-8
Commands Print Coating....................................................2-6
About TFCalc.................................................. 2-6 Printer Setup ....................................................2-6
Add.................................................................. 2-8 Read Angle Targets from File .......................2-19
Add Color Targets......................................... 2-19 Read Detector from File ................................2-33
Adjust Layers for Angle................................ 2-13 Read Distribution from File...........................2-35
Analyze Only ................................................ 2-36 Read Illuminant from File .............................2-31
Append from ................................................... 2-8 Read Layer Data from File ............................2-14
Back Layers .................................................. 2-12 Read Material from File.................................2-29
Calculate Statistics ............................... 2-50, 2-52 Read Substrate from File ...............................2-29
Close Coating.................................................. 2-5 Read Wavelength Targets from File..............2-19
Comments .............................................. 2-9, 2-23 Remove Saved Results ..................................2-56
Compute Color from Data File ..................... 2-58 Reopen Coating ...............................................2-5
Compute Cone-Angle Average..................... 2-41 Reverse Layers...............................................2-13
Compute EFI................................................. 2-40 Revert to Original ............................................2-6
Compute Equivalent Index ........................... 2-13 Save as Illuminant..........................................2-57
Compute Equivalent Stack............................ 2-14 Save Coating....................................................2-5
Compute Layer Sensitivity ........................... 2-41 Save Coating As ..............................................2-6
Compute Monitor Curve ............................... 2-41 Save Layer Data in Clipboard........................2-15
Compute Sensitivity...................................... 2-40 Save Layer Data in File .................................2-14
Configure ............................................... 1-6, 2-58 Save Results...................................................2-56
Consolidate Layers........................................ 2-13 Set Analysis Parameters ...................... 2-36, 2-42
Copy................................................................ 2-7 Set Color Chart Parameters ...........................2-56
Copy from ....................................................... 2-8 Set Cone-Angle Parameters...........................2-47
Create Blackbody Illuminant ........................ 2-31 Set EFI Parameters .............................. 2-40, 2-46
Cut................................................................... 2-7 Set EFI Plot Parameters .................................2-55
Delete .............................................................. 2-8 Set Global Search Parameters.............. 2-39, 2-45
Detectors ......................................................... 2-9 Set Monitor Chart Parameters .......................2-56
Distributions.................................................... 2-9 Set Monitoring Parameters ............................2-47
Edit Detector Comment ................................ 2-33 Set Optimization Parameters ............... 2-37, 2-43
Edit Distribution Comment........................... 2-35 Set Plot Parameters ........................................2-55
Page 68 Index
Set Sensitivity Parameters.................... 2-40, 2-46

Set Table Parameters..................................... 2-54 F
Show Color Chart ......................................... 2-53
Show EFI Plot ............................................... 2-52 First Surface ................................................ 2-10, 2-25
Show Monitor Chart ..................................... 2-54 Fitting a dispersion formula ..................................2-64
Show Plot ...................................................... 2-51
Show Table ................................................... 2-49
Show Thickness Totals ................................. 2-14 G
Stack Formula ........................................ 2-9, 2-11
Substitute Layer Data.................................... 2-15 Gain material................................................. 1-6, 2-27
Substrates ........................................................ 2-9 Gradient method........................2-37, 2-39, 2-43, 2-61
Swap Front-Back Layers............................... 2-58 Group
Targets - Continuous....................................... 2-9 Factor ................................................... 2-16, 2-36
Targets - Discrete............................................ 2-9 Options...........................................................2-16
Undo................................................................ 2-7 Window..........................................................2-16
Variable Materials........................................... 2-9
Window Options ........................................... 2-59
Cone-Angle Average computation.............. 2-34, 2-41 H
Constraints
Group factor .................................................. 2-16 Hartmann dispersion formula ...............................2-28
Layer thickness ............................................. 2-12 Herpin index..........................................................2-13
Variable material........................................... 2-24
Constraints, Dispersion formula parameter.......... 2-65
I
D Illuminant
Intensity .........................................................2-30
Data Name.................................................... 2-10, 2-25
Extrapolated ........................................... 1-6, 2-58 Plot.................................................................2-31
Interpolation.................................................. 2-27 Reading from file................................. 2-31, 2-60
dB units ................................................................ 2-54 Incident angle.............................................. 2-10, 2-42
Detector Incident medium ......................................... 2-10, 2-25
Efficiency...................................................... 2-32 Index profile................................................ 2-37, 2-43
Name .................................................... 2-10, 2-25 Internal transmittance
Plot ................................................................ 2-33 Reading ................................................ 2-29, 2-60
Reading from file ................................. 2-33, 2-60 Iterations ...............................................................2-43
Deviation .............................................................. 2-38
Deviation from target ........................................... 2-37
Dialogs, How to use ............................................... 1-5 L
Dispersion formula...................................... 2-27, 2-64
Distribution LAB Color ............................................................2-53
Reading from file ................................. 2-35, 2-60 Lambertian radiation source .................................2-47
Drude dispersion formula..................................... 2-28 Layer
Adjust for angle .............................................2-13
E Options...........................................................2-12
Reading from a file ........................................2-14
Saving to a file ...............................................2-14
EFI (Electric field intensity)...............2-40, 2-46, 2-52 Sensitivity ......................................................2-41
Environment......................................................... 2-10 Swap front and back ......................................2-58
Environments ....................................................... 2-25 Thickness constraints........................... 2-12, 2-45
Equivalent index .................................................. 2-13 Thickness totals .............................................2-14
Equivalent stack ................................................... 2-14 Thinnest .........................................................2-14
Exit medium ................................................ 2-10, 2-25 Windows ........................................................2-12
Zero-thickness...................................... 2-43, 2-44
Index Page 69
Light-monitor simulation ...................2-41, 2-47, 2-54 Monitor Chart ................................................2-54

Plot.................................................................2-52
M Substrate .............................................. 2-26, 2-29

Table ..............................................................2-50
Target................................................... 2-18, 2-22
Material Variable Material ................................. 2-24, 2-25
Gain........................................................ 1-6, 2-27
Variable......................................................... 2-24
Menus P
Edit.................................................................. 2-7
File .................................................................. 2-5 Phase shift conventions................................. 1-6, 2-58
How to use ...................................................... 1-3 Plot
Miscellaneous ............................................... 2-58 Overlay ..........................................................2-55
Modify............................................................. 2-9 Print................................................................2-52
Options............................................................ 2-8 Remark...........................................................2-51
Results........................................................... 2-49 Zoomed ..........................................................2-52
Run................................................................ 2-36 Power of the method .............................................2-37
Merit function
Derivative of ................................................. 2-43
Maximum...................................................... 2-37 Q
Powers 1, 2, 4, 8, 16...................................... 2-36
Monitor ratio ........................................................ 2-48 Quartiles....................................2-40, 2-46, 2-54, 2-57
Multiple environments ......................................... 2-25 QWOT........................................................... 1-6, 2-12
N R
N and K, Determining .......................................... 2-63 Radiation distribution.................................. 2-34, 2-47
Needle Optimization ................................... 2-44, 2-61 Reference wavelength...........................................2-10
Normalization factor ............................................ 2-36 Reference White....................................................2-56
Normalize Groups ................................................ 2-16 Refractive index
Data................................................................2-26
O Determining ...................................................2-63
Reading from file................................. 2-29, 2-60
Variable..........................................................2-24
Optical monitoring curve ...................2-41, 2-47, 2-54 Remark in Plot window ........................................2-51
Optimization Rugate coatings, How to simulate ........................2-24
Accuracy ....................................................... 2-43
Delayed quit .................................................. 2-38
Global................................................... 2-39, 2-45 S
Local ............................................................. 2-43
Needle .................................................. 2-44, 2-61 Schott dispersion formula .....................................2-28
Parameters..................................................... 2-43 Sellmeier dispersion formula ................................2-28
Progress......................................................... 2-37 Sensitivity
Options Color .................................................... 2-46, 2-56
Color Chart.................................................... 2-53 Layer ..............................................................2-41
Comment....................................................... 2-23 Minimizing .......................................... 2-41, 2-43
Detector......................................................... 2-33 Quartiles.............................2-40, 2-46, 2-54, 2-57
Distributions.................................................. 2-35 To random errors ................................. 2-40, 2-46
EFI Plot ......................................................... 2-52 Worst-case .........................2-40, 2-46, 2-54, 2-57
Group ............................................................ 2-16 Worst-possible ...............................................2-46
Illuminant...................................................... 2-31 Simplex method ........................2-24, 2-37, 2-43, 2-61
Layer ............................................................. 2-12 Stack formula .............................................. 2-11, 2-58
Material ................................................ 2-26, 2-29 Statistics, Calculating.................................. 2-50, 2-52
Page 70 Index
Substrate Targets - Continuous......................................2-21

Birefringent ................................................... 2-25 Targets - Discrete...........................................2-17
Name .................................................... 2-10, 2-25 Variable Materials .........................................2-24
Thickness ............................................. 2-10, 2-25 Worst-case sensitivity analysis .2-40, 2-46, 2-54, 2-57
Worst-possible sensitivity analysis .......................2-46
T
Targets
Color and luminance ............................ 2-17, 2-19
Continuous ........................ 2-21, 2-36, 2-38, 2-51
Density ................................................. 2-17, 2-21
Derivative............................................. 2-17, 2-20
Discrete .......................................2-17, 2-38, 2-51
Inequality ....................................2-17, 2-18, 2-21
Options................................................. 2-18, 2-22
Phase shift ............................................ 2-17, 2-21
Reading from a file ....................................... 2-19
Tolerance.....................................2-17, 2-21, 2-36
Thickness constraints ........................................... 2-12
Thickness priority ............. 1-6, 2-10, 2-11, 2-29, 2-58
Tolerance, Target ...............................2-17, 2-21, 2-36
Tolerancing .......................................................... 2-40
Tunneling method ....................................... 2-44, 2-62
V
Variable materials ....................................... 2-24, 2-63
Variable Metric method ........... 2-37, 2-39, 2-43, 2-61
W
Wavelengths
Range of ........................................................ 2-42
Units....................................................... 1-6, 2-58
Weighting factor ......................................... 2-17, 2-21
Windows
Color Chart.................................................... 2-53
Comments ..................................................... 2-23
Detector......................................................... 2-32
Editing data ..................................................... 1-4
EFI Plot ......................................................... 2-52
Environments ................................................ 2-25
Groups........................................................... 2-16
Illuminant...................................................... 2-30
Layers............................................................ 2-12
Materials ....................................................... 2-26
Monitor Chart................................................ 2-54
Plot ................................................................ 2-51
Radiation Distribution................................... 2-34
Substrates ...................................................... 2-26
Table ............................................................. 2-49
Section 3 TFCalc Optimization Guide Page 1
Table of Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Optimization Targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Broadband Designs
High-Reflectance (Low-Transmittance) Designs
Controlling Endpoint Behavior
Multi-Band Designs
Better: More Targets
Inequality Targets
Multiple Targets
Color Targets
Derivative Targets
Global Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Starting Designs
Best Design
Other Local Minimums
Needle Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Tunneling Method
Local Optimization Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Optimizing Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Page 2 TFCalc Optimization Guide Section 3
Introduction
The purpose of this section is to help the thin film designer create the best possible coatings. Because the experience
of TFCalc users varies from novice to expert, this section contains information that ranges from basic to advanced.
In almost all cases, real thin film coatings are designed using optimization -- not “canned” formulas. That is because
the few explicit designs are for idealized conditions: non-dispersive non-absorbing materials, normal incidence,
average polarization, etc.
A good way of understanding the optimization process is to see it as a control problem; the optimization method
attempts to control the output of a thin film coating by (1) gathering information at the various optimization targets
and (2) using that information to adjust the thickness and/or index of the coating’s layers.
All designs start with a specification or list of requirements. Sometimes the specification is very explicit, other times
it can be very vague. It is the designer’s job to translate the specification into inputs that TFCalc can use. Here are
the inputs that describe the physical model of the thin film coating:
• optimization targets
• substrate, incident, and exit mediums
• materials to be used
• number of layers
• illuminant
• detector
Then there are the optimization parameters that the user must specify:
• global optimization
• needle/tunneling optimization
• local optimization method
• number of iterations
• merit function “power”
• error tolerance
All of these inputs will be discussed in this section of the TFCalc manual.
Optimization Targets
TFCalc accepts a wide variety of target types:
• reflectance
• transmittance
• absorption
• optical density
• phase change on reflection
• phase change on transmission
• psi
• first, second, and third derivatives (with respect to wavelength or wave number) of the above quantities
• color
• luminance
For color and luminance targets, the wavelength range is 380 to 780 nm. For all other target types, the designer must
give the wavelength at which the target applies. In addition to the wavelength, the designer must specify the
following for each target:
• angle of incidence
• polarization
• target value
• tolerance
These are explained in detail in the Guide to TFCalc (see Targets Window). Below we list recommendations for
various designs.
Broadband Designs, which must operate over a wide range of wavelengths, should be designed with optimization
targets spaced by wave number or log of wavelength. This is because (1) the features (i.e., maximums and
minimums) of spectral curves are spaced by wave number and (2) the optimization method needs accurate
information across the entire spectral range over which the thin film coating must operate. This is just an instance of
a general rule:
The optimization method controls the thin film

coating at only those wavelengths, angles, and
polarizations for which there are targets.
The graphs below illustrate how the features are distributed by wave number. In the top graph, the horizontal unit is
wavelength (in nm); the bottom graph shows the same reflectance curve, but with wave number units (1/cm).
High-Reflectance (Low-Transmittance) Designs, using dielectric materials, can be difficult if the high-reflectance
must be maintained over a wide spectral range. There are two alternatives: (1) use a sequence of quarter-wave stacks
(see page RF-7 in the TFCalc Examples section) or (2) use optimization to find a non-quarter wave design.
It is well known that this type of coating requires a large number of layers. It is also easy to see that more layers
allow a coating to have a more complex behavior. So for high-reflectance coatings, it is important to have many
targets. This is an example of the rule:
As the number of layers increases, the number of

targets will have to be increased to control the more
complex behavior of the coating.
To see the behavior of a high-reflectance or low-transmittance design, it is best to view the optical density or to use
a dB scale, both of which are logarithmic. In fact, if the optical density is very important in a design, then using
density targets would make sense. For example, use an inequality target such as >3 to indicate that the transmittance
should be less than 0.1%.
Controlling Endpoint Behavior. Quite often, the behavior of a coating at the endpoints of the wavelength range
needs improvement. Consider the plot of an antireflection coating shown below.
Note that the reflectance increases significantly at the endpoints. This problem is easily fixed by decreasing the
tolerance of the endpoint targets and re-optimizing the design; that is, if all the target tolerances are 1.0, make the
tolerances at the endpoints 0.5. The plot below shows the improved behavior at the endpoints.
Note, however, that the reflectance has increased slightly in other parts of the spectrum. These are examples of the
rules:
If the behavior of the coating at an optimization

target needs to be improved, decrease the tolerance
of the target.
For an optimized design, decreasing the tolerance

of a target causes the coating’s performance to
suffer at other targets.
Multi-Band Designs have requirements for two or more (perhaps non-adjacent) spectral ranges. Many traditional
designs fall into this category: bandpass, edge, and notch filters. It also includes many advanced designs: multi-band
antireflection, multi-band bandpass, and multi-band rejection filters. As might be expected, these designs are much
more difficult than designs that have requirements for only one spectral range.
A common problem in multi-band designs is that the performance in one or more bands will fail to meet the required
targets. Often this problem can be traced to unbalanced targets; if one band has a 100 targets and another has 10
targets (and the tolerances are the same), then the optimization method will do a better job in the first band. Another
factor is the size of each band (measured in wave numbers); larger bands need more targets. If we let Wi denote the
reciprocal of the tolerance of the i-th target, and if we let Nj denote the width (in wave numbers) of the j-th band,
then the targets are balanced if
∑ W
--------------i
Nj
is the same for each band (where the sum is taken over all targets belonging to the j-th band).
This analysis does not take into account the physics of thin films; there may be physical reasons why one band is
more difficult to control than another band.
We summarize this as a rule:
For optimal performance of multi-band designs,

balance the targets in each band.
Better: More Targets. When in doubt, it is better to use more targets. There are several reasons for this: (1) it gives
the optimization method more information, (2) the merit function becomes smoother, which should speedup the
optimization process, (3) it is less likely that the optimization method will get stuck in a local minimum, and (4) it is
more likely that a good design will be found automatically. The main drawbacks of more targets are that the process
will take more time and more computer memory (RAM).
Inequality Targets. There are many situations in which inequality targets are useful. Here are some examples: (1)
an AR design requiring that reflectance be less than 1% from 380 to 780 nm, (2) a color filter design requiring that
(x,y) color be (2±0.1, 3±0.1) could be handled by four inequalities, and (3) a beamsplitter design may require that
reflectance be 50±1%.
Inequality targets are advantageous because they tell the optimization method what is important. For example, if the
goal is to create a design so that the reflectance is less than 1% from 380 to 780 nm, then as the optimization gets
close to the solution, it will concentrate only on those wavelengths at which the reflectance is greater than 1%.
The plot below shows a mirror coating designed with inequality targets R > 97.5%.
Suggestion: Because variations in coating manufacturing processes generality result in somewhat degraded
performance compared to the theoretical design, make the inequality “tighter” than necessary. That is, if the design
calls for < 1%, then make the inequality, say, < 0.9%. This ensures that the design will be well within the required
value.
When a design is nearly complete, an easy technique for creating an equal-ripple design is to use inequality targets.
This technique works because it forces the optimization method to concentrate on reducing the peak reflectances.
Multiple Targets at the same wavelength and angle are computed just once by TFCalc. This means that if the
designer uses multiple targets (say, reflectance and phase change on reflection) at the same wavelength and angle, it
will not slow down the optimization process.
Color Targets enable the designer to control the color of light reflected or transmitted by a coating, which is
important in many applications using visible light. In fact, in many visible applications there may be an implicit
requirement that the reflected or transmitted light have no color.
When there is a color target, TFCalc automatically creates “hidden” targets for wavelengths 380 to 780 nm with
either a 5 or 10 nm spacing. Remember that if the 10 nm spacing is selected, the color coordinates will be calculated
faster, but with less accuracy. A reasonable compromise is to start with the 10 nm spacing and then to switch to the
5 nm spacing for the final design.
Derivative Targets enable the designer to create designs which control the shape of the performance curves.
Note: if the design uses index, illuminant, or detector data from a table containing more than one data point, then the
derivative will be discontinuous at the wavelengths occurring in the tables. If the tabular data is fairly smooth, this
discontinuity will not be a major problem. For smooth derivatives, use dispersion formulas or constant-index
materials.
Global Optimization
TFCalc’s global search capability has several uses:
• finding good starting designs

• finding the best design
• finding other local minimums
Starting Designs. By entering a simple design such as (HL)^5, the global search procedure can quickly find several
designs which can be optimized by using one of the local optimization methods.
Best Design. In general, it is very difficult to find the global minimum -- the best design. If a minimum is found, it is
always necessary to qualify it, such as “using 7 layers comprised of TiO2 and SiO2, this design is probably the
best.” Having said that, here is a good technique for finding the best design:
• Start with a non-optimized design; using an optimized design would bias the global search.
• Use the “Exclude Designs Having a Deviation >” option in the Global Search Parameters dialog. When this
option is used, the global search does not waste time on designs that are far from optimal; the global search
finds the designs that are near the bottom of “valleys”. Be sure not to use a deviation that is not too small,
otherwise global search will have difficulty finding acceptable designs.
•
Other Local Minimums. When a local minimum has been found, the global search procedure can be used to find
nearby local minimums. Just set the “Maximum Change from Current Design” value to a small number, such as
10%.
Needle Optimization
Needle optimization is probably the most important recent development in the design of thin film coatings. Its
operation is described briefly in Appendix B in the Guide to TFCalc. Here are more observations about this
powerful method.
1. Keep in mind that when the needle method is used, a sequence of local optimizations are done. Hence, the
computational time required by the needle method will be longer than that of a simple local optimization.
2. The designs that the needle method finds are called optimal, which means that (using the given targets and
materials) the design cannot be further optimized by the needle method. It does not mean that the design is the best
possible.
3. If the needle method is started with a single layer, the thickness of that layer will determine the final design
found by the needle method. For example, in designing an AR coating on glass (index 1.52) for the range 380-780
nm using indices L=1.38 and H=2.3, it was found that ranges of initial thicknesses of L led to the same designs.
Here are the results:
Initial Thickness of L Layer (nm) Final Thickness of All Layers (nm)
92-120 118
130-260 145
280-540 523
550-840 805
The lesson: choose an initial thickness that is close to the final thickness you expect. For example, if you expect a
blocking filter to be about 5000 nm thick, use an initial layer 5000 nm thick. If you are not sure about how thick the
design needs to be, consider using the Needle/Tunneling method to generate a series of designs.
4. Appendix B states that it is best if contrast between the initial layer and the substrate is great. That is, if the
substrate has a low index, use the H material as the initial layer. Although this technique works well, it has been
found that there are some designs which can not be reached using this rule.
5. If more than two materials are used in a design, try making one of the materials the same as the substrate. If,
say, L is the same as the substrate, try using L as the initial layer. This approach makes it very easy to increase the
thickness of the design -- just add some additional thickness to the L layer next to the substrate.
6. Use the “Add Many Layers Simultaneously” option if the required design must be very reflective. This option
may also be good advice for any design that requires many layers.
7. For some types of coatings (such as normal incidence ARs) it can be shown that two materials are enough to
find an optimal design. Other coatings (such as broadband nonpolarizing beamsplitters) require at least three
materials.
8. If a design contains some very thin H or L layers, it may be useful to replace them with third material (call it M)
whose index lies between H and L, and then use local optimization to return to a minimum deviation. The M layers
will be thicker than the layers they replace and the performance of the coating will be maintained.
9. Because (a) designs containing thin layers are difficult to manufacture and (b) the needle method tends to create
thin layers as it approaches the optimal design, it may be advantageous to stop the needle optimization before it
reaches the optimal design (and, instead, use local optimization to complete the design). Even the non-optimal
designs can be very good.
10. For some designs, particularly those that start with a very thick initial layer, the layers added by the needle
method will stay thin until a large number of layers have been added. This is another case in which using the “Add
Many Layers Simultaneously” option may speedup the optimization.
Tunneling Method. This extension of the needle optimization method, described in Appendix B of the Guide to
TFCalc, is capable of improving any coating design by repeatedly making a layer thicker and using needle
optimization to add more layers. Here are a few suggestion for using the tunneling method:
1. Be sure that enough optimization targets have been entered. Tunneling will increase the thickness and number
of layers in a design, which will lead to more complex behavior, which must be controlled by using more targets.
2. Consider using inequality targets and setting maximum allowed deviation to 0.0 in the Tunneling Parameters
dialog. Then the tunneling method will continue working until all of the inequality targets have been met.
3. If the tunneling method stops because it has passed the maximum thickness set in the Tunneling Parameters
dialog, just increase that maximum thickness and restart the optimization.
4. Allow the tunneling method to generate designs that are somewhat thicker than you actually desire. In many
cases, this enables the tunneling method to find other designs that are somewhat thicker, but which may have fewer
thin layers.
5. After the tunneling method has generated a sequence of designs, it is possible to examine them and determine
how complexity, manufacturability, and performance are related.
Local Optimization Parameters

As discussed in the Set Optimization Parameters section in the Guide to TFCalc, there are three local optimization
methods: Gradient, Simplex, and Variable Metric. For almost all designs, the following settings are recommended:
Method: Variable Metric. It is usually faster than the Gradient method. If a very large number of layers are being
optimized, use the Gradient method to reduce the amount of memory required. These methods may converge to
different designs.
Max. Iterations: 1000. Use a large number so that the optimization method finds the local minimum.
Error Tolerance: 1.0e-6. Using a larger value will cause the optimization to stop before it reaches the local
minimum. To obtain a design closest to the true minimum, use an error tolerance of 1.0e-8.
Max. Change (%): 10.0. This value is used only by the Simplex method; it has no meaning for the other methods.
Power: 2, This value works well in most cases. If the final design has quite a bit of variation in the target deviations,
power = 16 will tend to equalize the target deviations (as much as is physically possible). Power = 16 is good for
producing equal-ripple designs.
Stop Optimizing Zero-Thickness Layers: Yes. If a layer thickness is forced to zero, it is not likely that the layer will
play an important part in the design; it is best that the optimization method ignore the layer in the remainder of the
optimization.
Display Index Profile While Optimizing: Yes. Although this option is not required, the index profile will often show
patterns from which the designer can learn.
Optimizing Sensitivity
TFCalc has the capability of optimizing the sensitivity of a coating design. That is, it is capable of finding designs
which are less sensitive to manufacturing errors. Because many random designs are computed, this optimization is
very slow. However, it can give the designer some insight into how less sensitive designs may be created.
For example, shown below is the sensitivity envelope (assuming 5% error) of a 4-layer AR coating designed for 450
to 650 nm:
After optimizing the sensitivity, the worst-case performance of the coating has been improved:
It is also possible to reduce the sensitivity of color. For more complex designs, the improvement may be more
prominent.
TFCalc Examples Page 1
__________________________________________________________________________________
Table of Contents
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Antireflection Filters
Single-Layer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AR- 1
Two-Layer (V coating) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AR- 2
Two-Layer (W coating) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AR- 3
Three-Layer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AR- 4
Four-Layer Two-Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . AR- 5
Two-Zero . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AR- 6
Graded-Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AR- 7
Graded-Index, Equal Ripple . . . . . . . . . . . . . . . . . . . . . . . . . . AR- 8
Wide Band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AR- 9
Detector-Relative. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AR-10
Bandpass Filters
Fabry-Perot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BP- 1
Phase Conjugate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BP- 2
Narrow-Band. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BP- 3
Wide-Band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BP- 4
Beam Splitters
Neutral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BS- 1
Cold Mirror . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BS- 2
Polarizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BS- 3
Nonpolarizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BS- 4
Edge Filters
Long Wave Pass 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ED- 1
Long Wave Pass 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ED- 2
Long Wave Pass 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ED- 3
Extended Rejection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ED- 4
Extended Rejection, Two-Sided. . . . . . . . . . . . . . . . . . . . . . . . ED- 5
Short Wave Pass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ED- 6
Notch (or Minus) Filters

Boudot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . NF- 1
Rugate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . NF- 2
Phase Retarders
Achromatic, 180°. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PR- 1
Achromatic, 270°. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PR- 2
Reflectors
Metallic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RF- 1
Enhanced Metallic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RF- 2
Dielectric 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RF- 3
Wide-Band Dielectric 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RF- 4
Dielectric 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RF- 5
Dielectric 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RF- 6
__________________________________________________________________________________
Page 2 TFCalc Examples
__________________________________________________________________________________
Introduction When these examples are modified, it is best to

give the new design a different name. Use the
Save Coating As command on the File menu to
The following pages show a wide variety of thin do this.
film coating designs. They have two purposes:
(1) to illustrate the features of TFCalc and (2) to Note: some of these designs are very sensitive to
give TFCalc users designs that can be modified changes in the index of refraction. Hence, if
to solve their own coating design problems. your material data differs from the data used to
design these coatings, you may have to optimize
Each example has a description of the general these designs using your indices.
type of coating, a plot of the example, a
comment about the plot, the name of the file Note: some of these designs have layer
containing the design, and up to three references thicknesses or layer configurations which would
from information sources listed in section 5 of make them difficult or impossible to
this manual. manufacture. The examples illustrate the
features and power of TFCalc; they are not
In general, there are three ways of modifying meant to be final products.
these designs to create similar coatings:
1. The reference wavelength may be changed. If

the thickness priority is set to Optical (in the
Configure dialog), then the features of the
coating will be shifted. For example, in the
single-layer antireflection filter, the minimum
reflectance can be shifted to 700 nm by setting
the reference wavelength to 700 nm.
2. The materials, substrate, exit medium, and

illuminant may be modified. Many of the
examples are for coatings in the visible
region. If these are shifted to the infrared,
materials and substrates that are transparent
for infrared wavelengths will need to be used.
With a few modifications, the basic design
will probably still work. If the indices are
very different, it is best to use optimization to
change the design.
3. Many of the designs are based on stack

formulas such as (HL)^9 1.2(HL)^9. New
coatings may be created by changing the
powers and factors occurring in these
formulas. This is also the best place to change
the coating materials.
__________________________________________________________________________________
Category Antireflection (AR) Filters Page AR- 1
Subcategory Single-Layer
Description This is the oldest AR design. Assuming the incident medium is air and the
substrate has index Ns, then a quarter-wave layer of any material with an index
less than Ns will minimize the reflectance at the reference wavelength. If the
index of the layer is Sqrt(Ns), then the reflectance will be zero at the reference
wavelength. However, for Ns < 1.9, there are no durable materials that allow
the reflectance to be reduced to zero. In this case, two-layer coatings are
required.
Illuminant: WHITE Angle: 0.0 (deg)

Medium: AIR Reference: 550.0 (nm)
Substrate: GLASS Polarization: Ave
Exit: GLASS Remark: Single Layer of MGF2
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
350 400 450 500 550 600 650 700 750
Reflectance (%) vs Wavelength (nm)
Plot The reflectance of the bare substrate is about 4.2%. A quarter wave of MgF2
Comment reduces the reflectance to about 1.3% at 550 nm.
Design File, PC ar\single.tfd Mac Antireflection:Single-Layer Pages

References A. Thelen, Design of Optical Interference Coatings 86-88
J.D. Rancourt, Optical Thin Films Users’ Handbook 82-83
H.A. Macleod, Thin Film Optical Filters 94
© 1993, Software Spectra, Inc.

Subcategory Two-Layer (V coating)
Description A two-layer AR coating allows more control of reflectance than a single layer.
One capability of a two-layer filter is to provide zero reflectance at a single
wavelength. This is also called a V-coating. For a given pair of materials, there
are usually two designs that result in zero reflectance. A disadvantage of this
type of coating is that the zero-reflectance wavelength is very sensitive to
manufacturing errors — which can be demonstrated by using TFCalc to display
the sensitivity of this design.

Exit: GLASS Remark: ZrO2-MgF2 "V" Coating
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
350 400 450 500 550 600 650 700 750
Plot The plot shows the reflectance of two designs. A fairly thick layer of ZrO2 is
Comment between the glass substrate and the MgF2 layer. To reduce the reflection to zero,
optimization was used to determine the thicknesses of the layers.
Design File, PC ar\2layer-v.tfd Mac Antireflection:Two-Layer V Pages

J.A. Dobrowolski, “Coatings and Filters”, Handbook of Optics 46-49

Subcategory Two-Layer (W coating)
Description A two-layer AR coating allows more control of reflectance than a single layer.
Another capability of a two-layer filter is to provide a range of low (but not zero)
reflectance. This is also called a W-coating.

Exit: GLASS Remark: ZrO2-MgF2 "W" Coating
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
350 400 450 500 550 600 650 700 750
Plot This is the classic H/2 L/4 design.

Comment
Design File, PC ar\2layer-w.tfd Mac Antireflection:Two-Layer W Pages

H.A. Macleod, Thin Film Optical Filters 95-100

Subcategory Three-Layer
Description Three-layer AR coatings allow more control of reflectance than two-layer
designs. The success of these designs depends heavily on finding three materials
whose indices are compatible with the substrate and incident medium.

Exit: GLASS Remark: Three-Layer Coating
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
350 400 450 500 550 600 650 700 750
Plot This is the classic M/4 H/2 L/4 design. It has the advantage of being fairly
Comment achromatic.
Design File, PC ar\3layer.tfd Mac Antireflection:Three-Layer Pages


Subcategory Four-Layer Two-Material
Description Three-layer AR coatings have the disadvantage of requiring 3 materials which —
for the best performance — may be difficult to find. This difficulty is overcome
by replacing the layer nearest the substrate by an equivalent 3-layer stack
composed of the other two materials. Thelen gives a table of designs (due to W.
Geffcken) for various substrate indices. However, using optimization, it is very
easy to design 4-layer coatings for specific substrates and wavelength ranges.

Exit: GLASS Remark: Four-Layer Coating
0.5
0.4
0.3
0.2
0.1
0.0
350 400 450 500 550 600 650 700 750
Plot The thin line gives the reflectance of the 4-layer design created by replacing layer
Comment 1 with an equivalent HLH stack. After optimizing the design for 450 to 650 nm,
the reduced reflectance is shown by the heavy line. Note that the reflectance
range was changed to 0-0.5% to show the performance more clearly.
Design File, PC ar\4layer.tfd Mac Antireflection:Four-Layer Pages


Subcategory Two-Zero
Description A three-layer design may also be the basis of a coating which must be
antireflecting at two wavelengths. Such a coating is common in frequency
doubling, in which one wavelength is twice the other. A common design is to
use three quarter-wave layers, and to use optimization to adjust the indices of two
layers so that the reflectance is exactly zero at the two desired wavelengths. If
the resulting indices are not close to those of real materials, one or both may be
replaced by equivalent LHL stacks. In this case, the design should be optimized
again to correct for the slight amount of dispersion contributed by the LHL
stacks.

Exit: GLASS Remark: Two-Zero Coating
5.0
4.0
3.0
2.0
1.0
0.0
600 700 800 900 1000 1100 1200
Plot This is not a frequency-doubling example; the wavelengths are 600 and 1100
Comment nm. In this case, we allowed the optimization process to adjust both the
thickness and indices of the layers. The variable indices were replaced by LHL
stacks composed of MgF2 and ZrO2 The same technique works on a wide
variety of substrates.
Design File, PC ar\two-zero.tfd Mac Antireflection:Two-Zero Pages

References H.A. Macleod, Thin Film Optical Filters 122-127

Subcategory Graded-Index
Description As noted by Dobrowolski, inhomogeneous coatings are appropriate for
high-index substrates, such as GE. In general, the index of the inhomogeneous
layer decreases monotonically from the substrate to the incident medium. Linear,
exponential, and hyperbolic index profiles all have similar performances.
An inhomogeneous coating can be simulated by a sequence of homogeneous

layers of various indices. Optimization can then be used to improve the
performance of such discretized filters.

Medium: AIR Reference: 4.0 ( m)
Substrate: GE Polarization: Ave
Exit: GE Remark: Graded-Index Coating
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
3 4 5 6 7 8 9
Reflectance (%) vs Wavelength ( m)
Plot A broad-band AR coating on GE. The upper curve gives the reflectance of the
Comment initial 10-layer design, which had quarter-wave layers whose index varied
linearly from 1.35 to 3.735. Using optimization, the reflectance was
substantially reduced, as shown in the lower curve.
Design File, PC ar\graded.tfd Mac Antireflection:Graded-Index Pages

References J.A. Dobrowolski, “Coatings and Filters”, Handbook of Optics 46-49
P. Yeh, Optical Waves in Layered Media 168-186

Subcategory Graded-Index, Equal Ripple
Description Thelen presents several theories for the production of AR coatings having the
equal-ripple property. However, a fundamental problem with these synthesis
methods is the reliance on indices that do not occur in nature. An alternate
approach is to use optimization to create equal-ripple coatings. For example, in
the graded-index design on a previous page, the reflectance increases at higher
wavelengths, which is not pleasing to the eye. This can be remedied by
changing the optimization targets to inequalities.

Substrate: GE Polarization: Ave
Exit: GE Remark: Graded-Index Coating
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0.6 0.8 1 1.2 1.4 1.6
Reflectance (%) vs Relative Wavenumber
Plot The appearance of this broad-band AR coating on GE was improved by changing

Comment the optimization targets to R < 0.5. Then optimization tends to produce
equal-ripple designs, as shown above. The reflectance was plotted has a
function of wavenumber to highlight the symmetry of the reflectance curve.
Design File, PC ar\graded2.tfd Mac Antireflection:Graded-Index Equal-Ripple Pages

References A. Thelen, Design of Optical Interference Coatings 55-72,

Subcategory Wide Band
Description By increasing the number of layers and using optimization, it is possible to
design broadband AR coatings.

Medium: AIR Reference: 1.0 ( m)
Substrate: G Polarization: Ave
Exit: G Remark: Baumeister’s contest
2.0
1.5
1.0
0.5
0.0
8 8.5 9 9.5 10 10.5 11 11.5 12
Reflectance (%) vs Wavelength ( m)
Plot This plot shows the reflectance of a 21-layer design composed of Ge and ZnS on
Comment a Ge substrate. The value of RMS merit function for this design is 0.6%.
Design File, PC ar\contest.tfd Mac Antireflection:Baumeister Contest Pages

References

Category Antireflection (AR) Filters Page AR-10
Subcategory Detector-Relative
Description If the characteristics of a detector are known, it is possible to design AR coatings
that are optimized for that detector. Consider the human eye, whose varying
sensitivity to visible wavelengths is well known. We can design an AR coating
for this detector by minimizing the luminous reflection of a coating.

Exit: GLASS Remark: Detector-Relative
0.5
0.4
0.3
0.2
0.1
0.0
400 450 500 550 600 650 700 750
Plot The "Four-Layer" example was optimized to minimize the luminous reflection.
Comment The thin line was the original design; the thick line, the optimized design. The
luminous reflection was reduced from 0.08% to 0.04%.
Design File, PC ar\detector.tfd Mac Antireflection:Detector-Relative Pages

References

Category Bandpass Filters Page BP- 1
Subcategory Fabry-Perot
Description This filter is essentially: a cavity between two mirrors. The width of the peak
decreases as the size of the cavity increases or as the number of layers in the
mirrors increases. The cavity must be a whole number of half-waves.
If manufacturing errors cause the peak to occur at a wavelength higher than

desired, the coating can be tilted to move the peak to the desired wavelength.

Exit: GLASS Remark: Fabry-Perot
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
500 510 520 530 540 550 560 570 580 590 600
Transmittance (%) vs Wavelength (nm)
Plot
Comment
Design File, PC bp\fp.tfd Mac Bandpass:Fabry-Perot Pages


Subcategory Phase Conjugate
Description Less sensitive to manufacturing errors than a Fabry-Perot filter. Patent by R.
Austin. Note that the peak is shifted slightly away from the reference
wavelength.

Exit: GLASS Remark: Phase Conjugate
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
500 510 520 530 540 550 560 570 580 590 600
Plot
Comment
Design File, PC bp\phasecon.tfd Mac Bandpass:Phase Conjugate Pages

References J.D. Rancourt, Optical Thin Films Users’ Handbook 112-114

Subcategory Narrow-Band
Description A narrow bandpass filter may be created by using multiple cavities. The slope
from transition may be increased by increasing the size of the cavity. However,
this also decreases the width of the pass region somewhat.

Exit: GLASS Remark: Square Bandpass
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
500 510 520 530 540 550 560 570 580 590 600
Plot This is a three cavity filter.

Comment
Design File, PC bp\narrow.tfd Mac Bandpass:Narrow-Band Pages


Subcategory Wide-Band
Description One technique for creating very wide bandpass filters is to coat two edge filters;
either a short wave pass filter on top of a long wave pass filter or a filter on each
side of a substrate. The advantage of this technique is that it is fairly easy to
position the two edge filters, making it easy to design filters of almost any width.

Exit: GLASS Remark: Wide Bandpass Filter
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
400 500 600 700 800 900 1000
Plot This was created from the "Long Wave Pass 3" and "Short Wave Pass" filters in
Comment this set of examples.
Design File, PC bp\wideband.tfd Mac Bandpass:Wide-Band Pages


Category Beam Splitters Page BS- 1
Subcategory Neutral
Description The goal for this type of filter is to split an incident beam into two parts for a
given wavelength range. Although a 50%-50% split is probably the most
common, other ratios are required for some applications. The incident angle may
be normal or nonnormal. At nonnormal incidence, polarization effects may be an
important consideration.
For nonnormal incidence, this type of coating may be placed between two
prisms, in which case the incident and exit mediums are the same.
Like antireflection coatings, optimization provides the best way of designing

these beam splitters. If dielectric materials are used, it is possible to design for
normal incidence and then adjust the layer thicknesses for another incident angle.

Exit: GLASS Remark: Neutral Beam Splitter
51.0
50.5
50.0
49.5
49.0
400 450 500 550 600 650 700
Plot This 8-layer TiO2-SiO2 coating was optimized to split the beam 50% from 400 to
Comment 700 nm.
Design File, PC bs\neutral.tfd Mac Beam Splitters:Neutral Pages

A. Thelen, Design of Optical Interference Coatings 110-122

Subcategory Cold Mirror
Description This is an example of a dichroic filter, in which one range of wavelengths is
transmitted and another range reflected. For a cold mirror, infrared wavelengths
are transmitted and visible light is reflected. An edge filter could be used to
accomplish this, but, in general, the cutoff does not have to be as abrupt as in a
typical edge filter. The color of the reflected light is important in some
applications. If the illuminant is known, then the coating can be optimized for
that illuminant.
Another coating of this type — a hot mirror — reflects infrared wavelengths and
transmits visible wavelengths.

Exit: GLASS Remark: Cold Mirror
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
400 600 800 1000 1200 1400 1600 1800 2000
Plot This 12-layer TiO2-SiO2 coating was optimized to reflect from 400 to 700 nm
Comment and transmit from 800 to 2000 nm.
Design File, PC bs\cold.tfd Mac Beam Splitters:Cold Mirror Pages


Subcategory Polarizing
Description A dielectric polarizing beam splitter can be created by tilting an edge filter and
optimizing it to improve the transmission of p polarized light. Although the
widest polarizing range is possible at the Brewster angle for the two thin film
materials, it is not difficult to design these filter for other angles.

Substrate: GLASS Polarization: S P
Exit: GLASS Remark: Polarizer
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
500 510 520 530 540 550 560 570 580 590 600
Plot An edge filter was optimized for 525-565 nm.

Comment
Design File, PC bs\polarize.tfd Mac Beam Splitters:Polarizing Pages


Subcategory Nonpolarizing
Description The design of nonpolarizing beam splitters for nonnormal incidence is a difficult
problem. Thelen explains several theoretical approaches — which are valuable
for providing a good starting design for optimization.

Substrate: GLASS Polarization: S P
Exit: GLASS Remark: Nonpolarizing Beam Splitter
23.0
22.0
21.0
20.0
19.0
18.0
17.0
500 510 520 530 540 550 560 570 580 590 600
Plot This is a 17-layer MgF2-AnS coating that reflects 20% at 45° over a fairly wide
Comment band. The s and p components are within 1% of the target reflectance.
Design File, PC bs\nonpolar.tfd Mac Beam Splitters:Nonpolarizing Pages


Category Edge Filters Page ED- 1
Subcategory Long Wave Pass 1
Description In this type of filter, the goal is to minimize transmission below a given
wavelength and maximize transmission above it. These designs are usually
based on a quarter-wave stack (HL)^p, which has a well-defined transition from
reflecting to transmitting.

Exit: GLASS Remark: Long Wave Pass
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
350 400 450 500 550 600 650 700 750
Plot This was created from the design A (0.5L H 0.5L)^9 B, where A and B are
Comment materials whose index and layer thickness were allowed to vary. Optimization
was used to find a design that minimizes transmittance in the rejection band and
maximizes transmission in the pass band. Layers A and B could be synthesized
using equivalent stacks.
Design File, PC ed\long1.tfd Mac Edge:Long Wave Pass 1 Pages


Description Optimization can be used to improve this type of filter.

100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
350 400 450 500 550 600 650 700 750
Plot This was created from the design (0.5L H 0.5L)^12. The thickness of the first
Comment three and last three layers was allowed to vary. Optimization was used to find a
design that minimizes transmittance in the rejection band and maximizes
transmission in the pass band. Note that the result is every close to the previous
design, but without the problem of synthesizing layers A and B.

References

Description Optimization can be used to improve this type of filter.

100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
350 400 450 500 550 600 650 700 750
Plot This design was created from the previous one by using optimization and
Comment allowing the thickness of four additional layers to vary. Note that transmittance
is greater than 99% in the pass band.

References

Subcategory Extended Rejection
Description The rejection region of an edge filter may be extended by coating one edge filter
on top of another; shifting the reference wavelength slightly for the second
coating.

Exit: GLASS Remark: Extended Rejection Region
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
300 400 500 600 700 800 900 1000
Plot This was created from the design (0.5L H 0.5L)^12 0.75(0.5L H 0.5L)^12.
Comment The thickness of the first five and last five layers was allowed to vary.
Optimization was used to find a design that minimizes transmittance in the
rejection band and maximizes transmission in the pass band.
Design File, PC ed\long4.tfd Mac Edge:Extended Rejection Pages


Subcategory Extended Rejection, Two-Sided
Description The rejection region of an edge filter may be extended by coating an edge filter on
each side of the substrate; shifting the reference wavelength slightly for the
second coating.

Exit: AIR Remark: Extended Rejection, Two Sided
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
300 400 500 600 700 800 900 1000
Plot This was created by putting an edge filter on each side of the substrate. The
Comment thickness of the first five and last five layers was allowed to vary. The two edge
filters were optimized simultaneously.
Design File, PC ed\long5.tfd Mac Edge:Extended Rejection, Two-Sided Pages


Subcategory Short Wave Pass
Description In this type of filter, the goal is to maximize transmission below a given
wavelength and minimize transmission above it. These designs are usually based
on a quarter-wave stack (HL)^p, which has a well-defined transition from
reflecting to transmitting.

Exit: GLASS Remark: Short Wave Pass
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
400 500 600 700 800 900 1000
Plot Example based on a 37-layer design in Thelen's figure 6.12. Optimization can
Comment be used to improve the performance in the pass band.
Design File, PC ed\short.tfd Mac Edge:Short Wave Pass Pages


Category Notch (or Minus) Filters Page NF- 1
Subcategory Boudot
Description A notch, or minus, filter requires that the transmittance be minimized in a
rejection band and maximized outside of that band. This type of filter is one of
the most difficult to design.

Exit: GLASS Remark: Minus Filter
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
400 500 600 700 800 900 1000 1100 1200
Plot This design comes from a thin film design "challenge" proposed by FTG
Comment Software.
Design File, PC nf\boudot.tfd Mac Minus:Boudot Pages

References J.D. Rancourt, Optical Thin Films Users’ Handbook 130

Category Notch (or Minus) Filters Page NF- 2
Subcategory Rugate
Description The classical rugate is a coating in which the index varies sinusoidally about an
average index Na. To reject a wavelength w, the physical thickness of each
period of the rugate must be 0.5 w / Na. Increasing the number of periods
improves the rejection.
The beauty of a rugate lies in (1) its ability to reject in just one narrow band and
(2) the capability of "adding" rugates so that multiple narrow bands may be
rejected.

Exit: GLASS Remark: 24.5 period Rugate coating
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
700 800 900 1000 1100 1200 1300 1400
Plot A rugate coating was simulated by dividing each period of the rugate into 20 thin
Comment layers. Note that is coating is similar to a H(LH)^24 design.
Design File, PC nf\rugate.tfd Mac Minus:Rugate Pages

References P. Yeh, Optical Waves in Layered Media 186-192

Category Phase Retarders Page PR- 1
Subcategory Achromatic, 180°
Description A dielectric phase retarder may be created by optimization.

Medium: GLASS Reference: 550.0 (nm)
Substrate: AIR Polarization: Dif S P
Exit: AIR Remark: Phase Retarder
190.0
185.0
180.0
175.0
170.0
400 450 500 550 600 650 700
Phase Change on Reflection (deg) vs Wavelength (nm)
Plot In this design, total internal reflection is used to reflect 100% of the beam. It is
Comment assumed that the mirror is coated on the hypotenuse of a right prism, so that the
incident medium is glass and the exit medium is air.
Note: The Abeles phase convention is used.
Design File, PC pr\180deg.tfd Mac Phase Retarders:180 Degrees Pages

References H.A. Macleod, Thin Film Optical Filters, 2nd Edition 348-350

Category Phase Retarders Page PR- 2
Subcategory Achromatic, 270°
Description A dielectric phase retarder may be created by optimization.

Medium: GLASS Reference: 550.0 (nm)
Substrate: AIR Polarization: Dif S P
Exit: AIR Remark: Phase Retarder
280.0
275.0
270.0
265.0
260.0
400 450 500 550 600 650 700
Phase Change on Reflection (deg) vs Wavelength (nm)
Plot In this design, total internal reflection is used to reflect 100% of the beam. It is
Comment assumed that the mirror is coated on the hypotenuse of a right prism, so that the
incident medium is glass and the exit medium is air.
Note: The Abeles phase convention is used.
Design File, PC pr\270deg.tfd Mac Phase Retarders:270 Degrees Pages

References H.A. Macleod, Thin Film Optical Filters, 2nd Edition 350-354

Category Reflectors Page RF- 1
Subcategory Metallic
Description Advantages: similar color appearance at all angles of incidence; broadband
reflection; thin metallic layers may be used as broadband partial reflectors.
Disadvantages: all but gold react with air; relatively soft coating; not good for
laser mirrors because metals absorb, leading to heat damage.

Exit: GLASS Remark: Three Metals
100.0
90.0
80.0
70.0
60.0
50.0
40.0
400 500 600 700 800 900 1000
Plot A comparison of Ag, Al, and Au reflectors. At 500 nm, gold has the lowest
Comment reflectance; silver the highest.
Design File, PC rf\metallic.tfd Mac Reflectors:Metallic Pages

J.A. Dobrowolski, “Coatings and Filters”, Handbook of Optics 91

Subcategory Enhanced Metallic
Description Advantages: metal is protected from oxidation; hard coating; similar color
appearance at all angles of incidence; fairly broadband reflection; thin metallic
layers may be used as broadband partial reflectors; may be used with lasers
because absorption is reduced.
Disadvantages: the phase shift on reflection is much greater than with bare metal.

Exit: GLASS Remark: Enhanced Aluminum
100.0
95.0
90.0
85.0
80.0
400 450 500 550 600 650 700
Plot The thin line gives the reflectance of bare Aluminum. The thick line gives the
Comment reflectance of Al covered with six layers of dielectric materials.
Design File, PC rf\enhanced.tfd Mac Reflectors:Enhanced Aluminum Pages


Subcategory Dielectric 1
Description The width of the reflector increases as the refractive index ratio nH/nL increases.
The reflectance increases as the number of layers increases.
Advantages: Low absorption because we use dielectric materials.
Disadvantages: limited reflection band and variable phase change on reflection.

Exit: GLASS Remark: Dielectric Reflector
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
350 400 450 500 550 600 650 700 750
Plot The plot shows a H(LH)^9 reflector .

Comment
Design File, PC rf\dielect1.tfd Mac Reflectors:Dielectric 1 Pages


Subcategory Wide-Band Dielectric 1
Description A wide-band dielectric reflector can be created changing the thickness of layers
so that adjacent layers have a common ratio. Let the ratio r be less than 1. Then,
starting with the outer layer, the QWOT of each layer would be 1, r, r^2, r^3,
etc. The resulting reflector will not be as efficient as a quarter-wave stack, but it
can have a much wider bandwidth. This "geometric" design may be improved
further by using optimization.

Exit: GLASS Remark: Geometric Stack
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
400 500 600 700 800 900 1000 1100
Plot This plot shows the reflectance of a 35-layer geometric stack composed of MgF2
Comment and ZnS. The common ratio is 0.97.
Design File, PC rf\geometrc.tfd Mac Reflectors:Geometric Stack Pages

H.M. Liddell, Computer-aided Techniques for the Design of 46-49

Description Slight modifications of the quarter-wave stack can improve the out-of-band
transmission.

100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
350 400 450 500 550 600 650 700 750
Plot The plot shows a (0.5L H 0.5L)^9 reflector, which improves the
Comment short-wavelength transmission.


Description Slight modifications of the quarter-wave stack can improve the out-of-band
transmission.

100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
350 400 450 500 550 600 650 700 750
Plot The plot shows a (0.5H L 0.5H)^9 reflector, which improves the
Comment long-wavelength transmission.


Description Another method of creating wide-band reflectors is to superimpose two or more
reflectors. The stack formula for such a reflector is
(0.5L H 0.5L)^p f (0.5L H 0.5L)^p, where p is chosen to obtain the desired
efficiency and the factor f can be found by trial an error or by optimization. An
advantage of this design is that most layers have the same optical thickness.

Exit: GLASS Remark: Superposition Reflector
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
400 500 600 700 800 900 1000 1100
Plot This plot shows the reflectance of a 21-layer design composed of SiO2 and
Comment TiO2. The shift factor was optimized using group optimization. By using
materials with a greater refractive index ratio, the reflection band could be made
even wider.
Design File, PC rf\superp1.tfd Mac Reflectors:Superposition 1 Pages

H.M. Liddell, Computer-aided Techniques for the Design of 46-49

Description The superposition idea can be generalized using the group optimization capability
of TFCalc. Consider the stack formula a(0.5L H 0.5L)^p b(0.5L H 0.5L)^p
c(0.5L H 0.5L)^p d(0.5L H 0.5L)^p e(0.5L H 0.5L)^p f(0.5L H 0.5L)^p
g(0.5L H 0.5L)^p h(0.5L H 0.5L)^p i(0.5L H 0.5L)^p j(0.5L H 0.5L)^p,
where p is a small integer and the factors a, b, c, d, e, f, g, h, i, and j must be
specified. Note that if all the factors are 1, then the result is a good reflector
centered at the reference wavelength. By setting optimization targets over a wide
wavelength range and then using group optimization, a wide-band reflector is
created. This is the case even if p=1, in which case the 0.5L H 0.5L components
can hardly be called reflectors. This method is advantageous when the coating
materials are dispersive. It is also better than optimizing all layers, which tends
to produce some very thin layers.

Exit: GLASS Remark: Superposition Reflector
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
400 500 600 700 800 900 1000 1100
Plot This plot shows the reflectance of a 21-layer design composed of SiO2 and
Comment TiO2. The ten shift factors were optimized using group optimization. In the
initial design, the first five group factors were set to 1 and the last five were set to
1.41.
Design File, PC rf\superp2.tfd Mac Reflectors:Superposition 2 Pages

References

Section 5 Information Sources Page 1
The following may be useful for learning more about thin film coatings:
Books
H. Bach and D. Krause, Editors, Thin Films on Glass, Springer, 1997.
M. Bass, Editor, Handbook of Optics, Second Edition, McGraw-Hill, New York, 1995.
P.W. Baumeister, Optical Interference Coating Technology, Lecture Notes, UCLA.
F.W. Billmeyer and M. Saltzman, Principles of Color Technology, Second Edition, Wiley, 1981.
J.A. Dobrowolski, “Coatings and Filters”, Section 8, Handbook of Optics, McGraw-Hill, New York, 1978.
J.A. Dobrowolski, “Optical Properties of Films and Coatings”, Chapter 42, Handbook of Optics, Second Edition,
McGraw-Hill, New York, 1995.
W.G. Driscoll, Editor, Handbook of Optics, McGraw-Hill, New York, 1978.
Sh.A. Furman and A.V. Tikhonravov, Optics of Multilayer Systems, Editions Frontiers, 1992.
R.W.G. Hunt, Measuring Colour, Second Edition, John Wiley & Sons, New York, 1991.
P. Klocek, Editor, Handbook of Infrared Optical Materials, Marcel Dekker, New York, 1991.
Z. Knittl, Optics of Thin Films, John Wiley & Sons, New York, 1976.
H.M. Liddell, Computer-aided Techniques for the Design of Multilayer Filters, Adam Hilger, 1981.
H.A. Macleod, Thin Film Optical Filters, Second Edition, Macmillan Publishing, New York, 1986.
E.D. Palik, Editor, Handbook of Optical Constants of Solids, Academic Press, New York, 1985.
E.D. Palik, Editor, Handbook of Optical Constants of Solids II, Academic Pr., New York, 1991.
J.D. Rancourt, Optical Thin Films Users’ Handbook, Macmillan Publishing, New York, 1996.
W.S. Stiles and G. Wyszecki, Color Science, Second Edition, Wiley, New York, 1982.
A. Thelen, Design of Optical Interference Coatings, McGraw-Hill, New York, 1989.
L. Ward, The Optical Constants of Bulk Materials and Films, Adam Hilger, 1995.
R.R. Willey, Practical Design and Production of Optical Thin Films, Marcel Dekker, New York, 1996.
P. Yeh, Optical Waves in Layered Media, John Wiley & Sons, New York, 1988.
Page 2 Information Sources Section 5
Conference Proceedings
Thin Film Technologies I, Proc. SPIE 401, 1983.
Thin Film Technologies II, J. Roland Jacobson, Editor, Proc. SPIE 652, 1986.
Thin Films for Optical Systems, Karl H. Guenther, Editor, Proc. SPIE 1782, 1993.
Optical Thin Films V: New Developments, Randolph L. Hall, Editor, Proc. SPIE 3133, 1997.
Optical Interference Coatings, 1988 Technical Digest Series, Vol. 6 (OSA, 1988).
Journals
Applied Optics, Optical Society of America, New York.
Journal of the Optical Society of America, Optical Society of America, New York.
Photonics Spectra, Laurin Publishing Co., Pittsfield, MA.
Laser Focus World, PennWell Publishing Co., Tulsa, OK.
Courses
Institute of Optics, Rochester, NY, Telephone (716) 275-4800.
Optical Interference Coating Technology, Los Angeles, CA. Call UCLA at (310) 825-1047.
Thin Film Center, Tucson, AZ. Telephone (520) 322-6171.
Internet Newsgroups
sci.optics
sci.engr.color

Manual

Uploaded by

Copyright:

Available Formats

Manual

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Manual

Uploaded by

Copyright:

Available Formats

™

© 1985-1999 Software Spectra, Inc. All rights reserved.

Telephone (503) 690-2099 • Fax (503) 690-8159

Windows Basics . . . . . . . . . . . . . . . . . . . . . .Section 1

Guide to TFCalc . . . . . . . . . . . . . . . . . . . . .Section 2

TFCalc Optimization Guide . . . . . . . . . . . .Section 3

TFCalc Examples . . . . . . . . . . . . . . . . . . . .Section 4

Information Sources . . . . . . . . . . . . . . . . . .Section 5

TFCalc Disk . . . . . . . . . . . . . . . . . . . . . . . . .Section 10

Windows is a registered trademark of Microsoft Corporation.

Using the Menus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Editing Data Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Using the Menus

Editing Data Windows

• Click the cursor in another box

Specifically, the user may

modify coating display

analyze close coating

Note that layer 1 is next to the substrate.

Clear causes the selected text to be erased.

This menu displays options available

Environment causes the environment dialog to appear.

Stack Formula causes the stack formula dialog to appear.

Comments displays the comments about this coating.

Environments displays up to 10 environments that may be used during optimization.

•Reference wavelength refers to the

Stack Formula Dialog

It is best to use the Tab key to move between the boxes.

Targets (Discrete) Window

Targets (Continuous) Window

• Whether it is an intensity or a phase target by entering either I or P.

Note: Be careful mixing discrete and continuous targets; if the

Variable Materials Window

Note: the index is not varied by the Simplex optimization method.

Materials and Substrates Windows

An example of a table of refractive index data for a material is shown below.

The dialog below shows an example of how a dispersion formula is edited.

Name Dispersion Formula n(λ)

For the dispersion formula k(λ), there are four choices:

Name Dispersion Formula k(λ)

Use this menu either to analyze or to optimize your coating. Note

For continuous targets, the summation above is replaced by a sum of integrals.

If the Optimize Designs button is selected, the following dialog appears:

Show Table displays a table of the (1) reflectance, transmittance,

• Display the reflectance, transmittance, absorptance, psi, optical density, or loss.

Appendix A: Reading Data from a Text File

500.0 1.38 0.001

500.0 1.4 0.1

TFCalc would use interpolation to determine n at 550 and k at 600.

Distributions. Each line of distribution data must have a format such as

Appendix B: Needle/Tunneling Optimization

Suggestions for Use:

Appendix C: Determining Refractive Index (N and K)

• A thin film layer may be inhomogeneous.

1. Create a new coating design file.

Polarization (Ave, P, S). This is the polarization of the measured data.

Substrate thickness. This is important only if the substrate absorbs.

• The measurements are not accurate enough.

A Edit Illuminant Comment ..............................2-31

Set Sensitivity Parameters.................... 2-40, 2-46