Ge ATR IR attachment user manual
The Germanium Attenuated Total Reflection Infrared attachment is used to study the reactions of
compounds (specifically thin film precursors) on a silicon wafer surface (SiOx). Here the IR beam is
directed through a hemispherical germanium crystal, where the beam is reflected inside the crystal off of
its flat surface, to the detector. An evanescent wave extends from the Ge surface at the area where the IR
beam reflects. The three dimensional electric field of the evanescent wave decays exponentially from the
sight of reflection. Absorbates inside this field will absorb the infrared light. The penetration depth of the
electric field of the evanescent wave formed from infrared light is around 100 nm. (Note: the penetration
depth is wavelength dependant.) Therefore this is a surface sensitive technique.
In general, the silicon wafer is brought into contact with the Ge crystal to measure the IR spectra of
the silicon surface. The SiOx surface is heated to a desired temperature and a compound is dosed inside
the attachment to react with the SiOx surface. The surface is cooled and brought in contact with the Ge
crystal to once again measure the IR spectra of the silicon wafer. Then the process is repeated.
1. Design
It is important to understand the design of the attachment to troubleshoot problems and to be able
to disassemble and reassemble the attachment. The attachment must be routinely taken apart to
decontaminate the internal surfaces, as the precursor compounds studied stick to the internal surfaces.
Additionally there are several places in the attachment where pieces of the attachment must be
properly aligned or oriented in order for the attachment to function properly.
1.1 General Design
The attachment has a translational control for the SiOx sample mount, shown as the black dial in
pictures above. The dial of the translational control has markings approximately 0.2 inches apart that
indicate a movement of the silicon wafer of 0.001 inches. It allows one to move the SiOx crystal in
contact with the Ge crystal, to collect IR spectra of the silicon surface, or to move SiOx crystal behind the
gate valve, to dose compounds onto the SiOx surface. The gate valve should only be closed when the
SiOx wafer is moved all the way back (-0.025 inches). The attachment is equipped with a heater and an
internal thermocouple that allows one to change and monitor the temperature of the SiOx surface. The
attachment is also equipped with a cooling rod, which allows the surface to be cooled more rapidly after
heating. The attachment has a pressure reader. This allows one to monitor the internal pressure of the
attachment and to indirectly monitor the pressure which the silicon crystal exerts on the germanium
crystal when they are placed in contact with each other. The later is important when obtaining
consistent IR spectra. In addition to the Swageloks and tubing used to allow Ar and precursors to be
separately introduced into the attachment, it has a heating wire wrapped around the outside and a
couple external thermocouples. This allows the precursor to be heated, to increase its vapor pressure,
along with the walls and for the temperature of each to the monitored. Typically, the walls are kept at
our above the temperature of the precursor.
The attachment is also equiped with a purge box that contains two mirrors. This box allows the IR
beam to be aligned to pass through the Ge crystal to the detector under nitrogen.
1.2 in-depth design, including clearances and alignments
The pictures above the silicon wafer mount and the wires that are attached to it. Here a Teflon
block is attached to the shaft through a barrel connector attached to the back of the Teflon block by two
screws (these two screws are not visible in the pictures). This barrel connector sits flush to another
barrel connector. This second barrel connector is used to attach a piece of stainless steel round stock
(not tubing) cut to a very specific length to extend the length of the shaft. Both barrel connectors use
1/8 inch 4/40 screws. The barrel connector that is screwed into the Teflon block may only have one
screw on top at the position furthest from the Teflon block the other two screws must be placed in the
bottom position. (The barrel connectors are threaded through on both sides.) This must be done or the
screws will hit the gate valve. Additionally, the Teflon block must be oriented so that the longest part
perpendicular to the shaft is pitched down 30 degrees counterclockwise from the horizontal when
looking at the tip (as shown in the pictures). This again is to provide clearance with the gate valve. Note:
there is a notch cut out of the top of the Teflon block. At the tip of the Teflon block is a sheet of nickel
that the silicon crystal will be placed. On each side of the nickel sheet is a sheet of copper used to attach
the wires used for each leg of the heater circuit. Two screws go through each copper sheet and the
nickel sheet to attach it to the Teflon block. Also attached to the nickel sheet is the thermocouple, which
is spot-welded to the surface. All of these wires are coiled around the shaft.
NOTE: THE ALIGNMENT OF THE NICKEL SHEET FLUSH TO THE FLAT TEFLON BLOCK IS ONE OF THE MOST
IMPORTANT ALIGNMENTS IN ORDER TO OBTAIN USEFUL SPECTRA. There is a separate section bellow
that indicates how to tell from the IR intensity and spectra if it is aligned properly.
The thermocouple and the first leg of the heater, the black wire in the picture, go from the
sample mount to barrel connecters connected to external leads that go through the mini-flange. The
thermocouple connections, through the mini-flange, are further away from the Teflon block that the
first leg of the heater. (You may want to note the orientation of the external leads in the first two
pictures in section 1.1.) The black wire, which is the wire for the first leg of the heater, is connected to
the external lead by a barrel connector that has sheathing over it (seen in the left picture). This
sheathing must be secure over the barrel connector or the circuit will be grounded to the walls of the
attachment prior to going through the nickel sheet. The black wire, which is the wire for the first leg of
the heater, is connected to the copper sheet (which is connected to the nickel sheet) by a barrel
connector that has its screws going through the copper sheet. This barrel connecter is secured under the
Teflon block. Unlike the first leg of the heating wire the second leg of the heating wire cannot be
secured to the copper sheet (which is secured to the nickel sheet) by a barrel connector. (The barrel
connector securing the wire from the first leg does not allow it to be mounted bellow the Teflon block,
the gate valve does not allow it to be mounted above the Teflon block, and the screws securing the
copper and nickel sheets do not allow it to be mounted on the sides.) Therefore, the 12 gauge wire is
secured directly by the copper sheet (see the picture on the left). The other end of this wire is attached
directly the cooling rod (which is in ohmic contact with the walls of the attachment). (Note: 12 gauge
copper wire was used as the largest diameter wire (that still has enough flexibility) to increase the
efficiency of thermal current between the copper cooling rod and the silicon sample.)(Note: the original
sheathing of the 12 gauge copper wire was replaced to add flexibility.) The copper cooling rod is custom
made from a piece of ? copper round stock cut to length and spun down to have sections that have ?
inch, 3/16 inch, and 1/8 inch diameters. The 1/8 inch diameter end is attached by a barrel connector to
the heating/cooling wire, while the ? inch diameter piece extends outside the chamber and is connected
to a funnel and tubing used to hold liquid nitrogen.
The inner walls of the attachment are comprised of five major parts. The first part consists of
the translational control and the shaft. The second part has the external connections for the heater and
the thermocouple as well as the cooling rod. These two parts are shown in the pictures above (at the
beginning of this sub section). The third part has a two inch extension and the ? inch tubing that is used
to attach the precursors and the argon line. The fourth part is the gate valve shown below. The final part
is the blank with a hole to which the Ge crystal is attached; it is shown in the bottom right of the top
right picture above (at the beginning of this sub section).
The first three parts are attached together with the same bolts; however, the bolts to attach the
third part and the gate valve should be inserted into the third part prior to attaching the first three parts
(do to clearance issues). These bolts are cut to length to allow them to fit. Note: the first three parts
must be attached so that the backing of the holder of the Ge crystal attached to the last part is vertical.
The gate valve must be attached with a preferential orientation. Here the side of the gate valve shown
in the left picture below must be closer to the translational control, while the side shown in the right
picture below must be closer to the Ge crystal.
2. RCA cleaning
Here RCA cleaning is used to remove the native oxide and to regrow a clean oxide. There are three
principle steps in the RCA cleaning method. There is the etching, which is done using hydrofluoric acid,
ammonium fluoride, or a mixture of the two. There is the SC1 cleaning, which involves exposing the
silicon wafer to a solution composed of water, hydrogen peroxide, and base (typically sodium hydroxide
or ammonium hydroxide). And there is the SC2 cleaning, which involves exposing the silicon wafer to a
solution composed of water, hydrogen peroxide, and acid (typically hydrochloric acid or sulfuric acid).
Read the hydrofluoric acid section of the chemical hygiene plan before performing any RCA cleaning.
Use Proper PPE. Do not use glass or metal glassware (or tweezers) with hydrofluoric acid solutions.
Although it may later be changed, the current RCA cleaning procedure is given below.
Prepare a water bath at 80 degrees Celsius. In a plastic beaker add 20 mL of milli-Q water. Then add
5 mL of 30% hydrogen peroxide solution to the beaker using a pasture pipette. (The pipettes full is
approximately 5 mL.) Then add 5 mL of sodium hydroxide to the beaker using a pasture pipette. (The
sodium hydroxide used is 29 weight percent solution prepared from sodium hydroxide pellets.) Add the
silicon wafer to this solution and place the beaker in the water bath for 10 minutes. Then rinse the wafer
in another beaker containing milli-Q water and place the wafer in an empty beaker. Place the SC1
solution and the first rinse of the beaker in a waste container. In the beaker containing the wafer add a
small amount of buffered HF (7:1 ammonium fluoride: hydrofluoric acid) and let it set for two minutes.
Then rinse the wafer in another beaker containing milli-Q water and place the wafer in an empty beaker.
Place the buffered HF solution and the first rinse of the beaker in a waste container. Add 20 mL of milliQ water to the beaker containing the wafer. Then add 5 mL of 30% hydrogen peroxide solution to the
beaker using a pasture pipette. Then add 5 mL of 12.1 N hydrochloric acid to the beaker using a pasture
pipette. Place the beaker in the water bath for 10 minutes. Then rinse the wafer in another beaker
containing milli-Q water and place the wafer in an empty beaker. Place the SC2 solution and the first
rinse of the beaker in a waste container. Add 20 mL of milli-Q water to the beaker containing the wafer.
Then add 5 mL of 30% hydrogen peroxide solution to the beaker using a pasture pipette. Then add 5 mL
of 29 weight percent sodium hydroxide solution to the beaker using a pasture pipette. Place the beaker
in the water bath for 5 minutes. Then rinse the wafer in another beaker containing milli-Q water and
place the wafer on a chem. Wipe to later be mounted in the attachment. Place the SC1 solution and the
first rinse of the beaker in a waste container.
3. Sample mounting
After RCA cleaning the silicon sample is mounted on the nickel sheet (shown in the picture at the
beginning section 1.2). A single drop of silver paste is placed at the center of the nickel sheet. The silicon
sample is placed on top of the silver paste with the polished side up and pushed against the nickel sheet.
(Note: Excess silver paste is undesirable.) The sample is then retracted using the translational control.
The gasket and Ge crystal are then cleaned. First they are rinsed with hexane and wiped with a chem.
wipe. Then they are rinsed with milli-Q water and wiped with a chem. wipe. This is followed by one
additional rinse with milli-Q water and wiping with a chem. Wipe. The gasket and Ge crystal is then
placed on the end of the attachment and secured with the Ge crystal holder. The holder is secured so
that the backing is vertical. The purge box is then attached. To align the purge box properly, a torpedo
level is used. The purge box is secured when the torpedo level indicates that the box is level. The
attachment is then placed in the beam path. A typical single beam intensity is around 1300. (The single
beam intensity can be viewed in the OPUS software by selecting the ”Signal” tab after clicking on the
“Advanced measure” button under the “Measure” table.)
4. SiOx alignment
As stated in section 1.2, the alignment of the SiOx sample with the Ge crystal is critical to proper
data collection. Fortunately, there are a couple of easily discernable indicators that will show how well it
is aligned once the attachment is assembled and placed in the IR beam path. The first is by examining
the change in the single beam intensity, as viewed under the “Signal” tab, between when the SiOx
sample is placed in contact with the Ge crystal and when it is not in contact with the Ge crystal. Here the
single beam intensity should drop by over 50% when the SiOx sample is in contact with the Ge crystal.
The second indicator is from IR spectra themselves. Below are two IR spectra collected with the SiOx
sample in contact with the Ge crystal (with the Ge crystal that is not in contact with the SiOx sample as
the background). The top spectrum below is a spectrum collected when the SiOx crystal was misaligned
while the bottom spectrum has the SiOx sample correctly aligned. The two most prevalent differences
between the two spectra are the differences in baseline absorbance and the visibility of vibrational
modes (such as CH and OH stretches).(Note: both indicators are checked after initially heating the SiOx
sample and allowing it to cool.)
If one finds that the SiOx sample is not aligned correctly, then one must remove the purge box
and the Ge crystal and realign the nickel sheet of the sample mount. The sample mount must be
extended to allow access to the screws holding the copper and nickel sheets to the Teflon block. (You
may want to refer to the picture at the beginning of section 1.2.) Take of the SiOx sample from the
mount. Loosen the screws on one side. Then press the nickel sheet flat against the Teflon block and
tighten these screws. Do this, separately, with the screws on the other side as well. If the nickel sheet
still does not feel flat on the Teflon block, repeat. After the nickel sheet is correctly secured to the Teflon
bock remount the SiOx sample and recheck the alignment with the IR beam/spectra. If the alignment is
still incorrect you may need to take the nickel and copper sheet of the Teflon block and file the end of
the Teflon block to make it flatter. (Checking the flatness of the end of the Teflon with a torpedo level
may help.)
5. Attachment cleaning
The organometallic precursors studied with this attachment can accumulated on the inner walls of
the attachment after repeated exposures. Eventually, the organometallic precursor concentration
becomes high enough that the organometallic precursors will react with the SiOx surface to its
maximum coverage once the sample is heated, without dosing any additional organometallic precursor
into the attachment. This makes it impossible to perform experiments that examine coverage versus
exposure to this organometallic precursor. It also makes the examination of the second half reaction
unlikely, as the surface will likely react simultaneously with the organometallic precursor already
present in the attachment. Therefore, at this point (or perhaps even before this point), the attachment
needs to be cleaned.
The way to determine if the attachment needs to be cleaned is to examine the IR spectrum of the
SiOx surface after it is heated (to the temperature that you are planning to perform that days set of
experiments) and allowed to cool without dosing any precursors. In particular, you need to examine the
C-H stretch region, for an indication of presence of reacted precursor on the surface, and the O-H
stretch region, for an indication of the presence of unreacted OH groups on the surface. (The Si-H
stretch region should be examined if a Si-H surface is used.) If there are no O-H stretches present, then
the attachment needs to be cleaned. If there O-H Stretches are present, then it must be confirmed that
an additional dose of precursor will result in a reduction of this peak (and possible increase of the C-H
stretch region). If not, the attachment must be cleaned. Otherwise use your own judgment.
To clean the attachment, it must be disassemble. Set the pressure reader aside, as it should not be
cleaned with any solvents. Also set the Ge crystal aside. Then the interior surfaces must be wiped down
with a chem. Wipe that has hexane or dichloromethane soaked into it. (A mixture of the two solvents
may also be used.) Be curtain to not only wipe the internal walls but also to wipe down the shaft. Be
curtain to expose the internal wires to the solvent as well. The Teflon surfaces must be examined for
discoloration. There are two internal pieces of Teflon: One inside the mini-flange and the other at the
end of the shaft. If the Teflon surfaces are discolored then they need to be filed to remove the
discoloration. The Teflon block at the end of the shaft must be removed from the shaft and have the
nickel and copper sheets removed to file all of its surfaces. The ? inch tubing must be cleaned with q-tip
that are soaked in hexane or dichloromethane (or some mixture of the two). While the elbows,
Swageloks, and the four-way connector can be cleaned easily with q-tips, to clean the ? inch tubing the
cotton from the q-tips must be removed (soaked in solvent) and forced through the tubing with the stick
of the q-tip. Once the internal surfaces have been wiped down, reassemble the attachment. Here, it is
recommended that you check that the assembled attachment can be pumped down to the same base
pressure that it could achieve prior to disassembly. Now take off any precursors and the pressure
reader. Also take off the Ge crystal. Place blanks on any openings of the ? tubing. Fill the attachment
(from the opening where the Ge crystal is normally attached) with hexane (not dichloromethane). Let it
sit overnight. Then drain the hexane. Remove the blank, to which the Ge crystal is normally attached, to
remove any hexane that has pooled inside the gate valve. Also, move the attachment around to drain all
the hexane from the ? inch tubing. Then reassemble the rest of the attachment (leaving a blank in place
of the pressure reader) and attach it to the mechanical pump. Start pumping the attachment. Be certain
the heating tap is raped around the attachment and turn it on (through a variac). The walls and the
sample mount should be held at a temperature of 62 degrees Celsius. After an hour reattach the
pressure reader. Continue pumping the attachment until it reaches base pressure. Then mount another
SiOx sample (see section 3) and check the extent of contamination.
NOTE: The walls of the attachment should not be heated to high above 100 degrees Celsius as its
maximum temperature, because of the gate valve. Most gate valves are rated to 150 degrees Celsius.
6. Data collection
IR spectra are inherently temperature dependent and ATR is also dependent on the amount of
absorbates and their proximity to the surface. The later dependence means that IR spectra are
dependent on the pressure that the Ge crystal and the silicon surface are pushed together. When
collecting IR spectra we must remember these fact and have protocols for collecting data that both
minimizes errors that may develop from these dependencies and takes advantage of these
dependencies.
To collect IR data, first the silicon sample needs to be mounted and the attachment needs to be
placed in the beam path (see section 3 for details). After the attachment is placed in the beam path
(with a single beam intensity of around 1300), a sample spectrum should be collected using a previous
background of the clean Ge crystal. This spectrum should be examined for signs that the crystal may
have contamination on the surface (such as C-H stretches). If this spectrum indicates contamination is
present, remove the Ge crystal and clean it with hexane and milli-Q water (refer to section 3 for details).
It may also help to wipe the mirrors in the purge box with a dry chem. Wipe. Then reassemble the
attachment and collect another sample spectrum to check the surface for contamination again. Once
the Ge crystal is clean collect a background spectrum.
If you have a background spectrum of a clean Ge crystal (or are unsure if you have one) there is
another way to tell is the Ge surface is clean that takes advantage of the temperature dependence of IR
spectra. First, collect a background spectrum of the Ge crystal. Next, heat the walls of the attachment.
The collect a sample spectrum (using the background spectrum you collected before heating the
attachment). Examine this spectrum for signs of contamination. These should appear as negative peaks
(such as negative C-H stretch peaks). Even if you did confirm cleanliness of the Ge crystal with an older
spectrum it is recommended that you collect this spectrum after heating the attachment.
After the cleanliness of the Ge crystal is confirmed, the contamination of the attachment must be
examined. After collection a background spectra of the Ge crystal, fully retract the silicon crystal (to a
position of -0.025 in as indicated on the translational control) and close the gate valve. Next, heat the
silicon sample to the temperature at which precursors will later be dosed. Then, cool the silicon sample,
open the gate valve, and move the silicon sample so it is in contact with the Ge crystal to collect a
spectrum (this is typically between 1.950 inches and 2.000 inches as indicated on the translational
control). Refer to section 4 to determine if the Silicon sample is aligned properly. Then refer to section 5
to determine if the attachment needs to be cleaned.
When you move the Silicon sample into position to collect a spectrum there are several key things to
consider. First, since IR spectra are temperature dependent, you want to collect all of your spectra at the
same temperature. Second, the design of the attachment allows us an indirect measurement of the
pressure between the Ge crystal and the silicon sample, which the IR spectra are also dependent. (Note:
currently a Teflon block is used to secure the Ge crystal to the attachment.) This is possible because the
Ge crystal sealed to end of the attachment with a gasket that covers an opening to the attachment.
When the silicon sample is moved in contact with the Ge crystal this seal can be partially broken. This
changes the pressure inside the chamber which can be easily measured. (Note: I typically collect spectra
when the pressure in the chamber is a steady 100 mTorr.)
When you have determined that there are no obvious problems with the attachment, you are ready
to start performing experiments. First, heat the attachment to the desired temperature. (This
temperature will depend on the precursor used. The precursor is heated to increase the vapor pressure
of the precursor and the walls are held at or above the temperature of the precursor. Typically, one sets
the variac to the voltage setting known to produce the desired temperature (with the heating wire) and
allows the attachment an hour to reach this temperature.) Next, collect a background spectrum of the
Ge crystal. Then collect a sample spectrum of the silicon sample before dosing any precursors.
When you are ready to dose your precursor fully retract the silicon sample and close the gate valve.
Then start doing argon gas. When the pressure is stable (around 100 mTorr), heat the silicon sample to
the desired temperature. Then open the Swagelok to the precursor for the desired amount of time. Stop
heating (and start cooling) the silicon sample and allow the argon gas to flow for a couple minutes to
purge the attachment. Then stop the flow of argon gas. (The pressure should now be the same as before
you started dosing.) (The sample can be more rapidly cooled by pouring liquid nitrogen down the tube
attached to the cooling rod.) Open the gate valve and bring the silicon sample in contact with the Ge
crystal. Repeat this process as many times as desired.
When collecting multiple spectra it is important to insure continuity and reproducibility of the
spectra obtained. First, insure that all spectra are collected at the same temperature and pressure. Then
be sure to compare the spectra (before any baseline correction) to insure that they are of the same
intensity and shape.