|
Temperature controlled fluid stage
The design of this stage can be found in R. K. Workman, S. Manne,
"Variable temperature fluid stage for atomic force microscopy", Rev. Sci. Instrum.,
71 (2), 2000.
It is often useful to study the behavior of a system as a function of temperature. For our research on surfactant aggregation
at the solid/liquid interface, we wanted to study temperature induced phase changes of these aggregates. There are often
interesting phenomena both above and below room temperature for the same surfactant, so we wanted to be able to heat and cool
the system which consisted of the sample, the surfactant solution and the AFM tip.
Surfactant aggregates are imaged with AFM by using the repulsive double-layer force for ionic surfactants, or a stearic
repulsive force for non-ionic surfactants. For a review of this imaging technique, click
here.
Design criteria
AFM imaging of surfactant aggregates using these repulsive stabilization forces is sensitive business. Because the
surfactant structures are very delicate, the imaging force must be maintained within a few 100s of piconewtons in order
to avoid “breaking through” the structures. The cantilevers used for imaging surfactant aggregates must have a moderately
low spring constant on the order of ~0.05 to 0.5 N/m to have adequate sensitivity, however the typically used silicon nitride
levers have a thermally induced bending of about 100 nm / K. Therefore, to be able to image the surfactant aggregates, a
temperature stability of ~10 mK must be maintained over the time it takes to capture one image (~1 minute). The mechanical
stability of the temperature controlled stage also had to be high enough to allow the necessary molecular resolution over
the entire desired temperature range.
The AFM we designed the temperature controlled stage for was the Digital Instruments Dimension 3100 AFM. This AFM scans the
tip from above a stationary sample. This type of design allows for scanning large samples, and in our case allows room for
a compact temperature controlled stage. The height for the stage had to be less than the 38 mm distance from the fully
retracted AFM tip to the granite slab the Dimension 3100 is built on (See figures 1 and 2).
In order to achieve heating and cooling in the same instrument led to the choice of a thermoelectric module. We
wanted the ability to heat and cool over the entire aqueous temperature range of 0 – 100 C. Typical thermoelectric coolers
have a maximum temperature limit of ~100 - 120 C, which was a little too close to our design maximum temperature.
Fortunately, Melcor makes a line of high temperature thermoelectric modules
that have a high temperature limit of 200 C (now available up to 225 C). We have also used modules from
Tellurex for improved low temperature performance, when we don't also need to go to
temperatures above 100 C.
Construction
The difficulty in using a thermoelectric module for cooling the sample is removing the heat from the hot side without
introducing any mechanical vibrations. The first approach taken was to immerse a heat sink in a pool of water or antifreeze
which was in turn cooled by heat exchange tubing.
Because of the large operational temperature range, the thermal expansion and contraction of the stage is important. A design
was created with low thermal expansion legs which extended slightly below the heat sink, so that the heat sink was free to
expand and contract independent of the rest of the stage. Figure 3 shows a schematic drawing of the heat stage. A
thermoelectric cooler is sandwiched between a stainless steel or titanium fluid cell and an aluminum heat sink. Magnets glued
in the bottom of the ceramic legs hold the fluid stage to a custom Invar base chuck.
The final design ended up being about 23 mm tall, which was too tall to use with the existing Dimension base chuck, which gave
only 21 mm clearance (see figure 2). A thinner base chuck was made from Invar (the same material the existing chuck was made
from) to give the needed clearance. It would be possible to make the heat sink and legs 2 mm shorter so the existing base
chuck could be used. This should not effect performance significantly.
The cooling performance of the stage was limited by the ability of the heat sink to remove heat from the
thermoelectric module. Improved performance could be realized by an actively cooled heat sink.
The next generation
To test this idea, the next iteration consisted of a 1/4" thick aluminum plate that had water passages drilled out
under the thermoelectric cooler. This time, the fluid cell and thermoelectric module were screwed directly into
the heat sink (aluminum plate). The thermal expansion of the heat sink should not be a serious problem if the water
flow is able to keep the heat sink at a constant temperature. The only concern with this design was whether the water
flow would disturb the delicate imaging of surfactant aggregates. The centrifugal pump we used provided very smooth
water flow at ~1/3 L/minute. No difference could be noticed in the images whether the pump was on or off. A larger version
of the heat sink was made (5 1/2" diameter), which had a center pin to interface with the dimension base chuck similar to the
top (vacuum) chuck. The dimension base chuck vacuum holds this aluminum heat sink tightly to the base chuck.
Using this type of heat sink allowed lower temperatures with room temperature water, and actually provided more
stable temperature control because the heat sink temperature stayed more constant than with the passive heat sink.
A temperature difference of 30 C was now possible between the hot and cold side of the thermoelectric cooler. With
ice water as the cooling fluid, the cold side temperature could now reach –30 C with excellent thermal and mechanical
stability.
Contact info:
Srin Manne
Physics Dept
PAS 575
520-626-5305
smanne@physics.arizona.edu
|
Figure 1. Dimension 3100 scanner and base chuck. The vacuum chuck has been removed.
 |
Figure 2. Schematic drawing of Dimension 3100 sample stage dimensions
 |
Figure 3. Schematic drawing of the heat/cool fluid stage.
 |
Figure 4. Water cooled heat sink. Water cooling lines are seen coming out of the left side of
the heat sink, the wires for the thermoelectric cooler and the platinum resistance thermometer (platinum RTD) are going off
to the right side of the image. Ceramic screws (difficult to see in image) are used to hold the fluid cell to the heat sink.
 |
|
