1. Airtight Chamber

An acrylic chamber was designed to be the place where the mice receive drug vapor and receive other auditory or visual experiments. Because the opioid drug is hazardous to humans, the chamber has to be airtight in order to prevent potential drug leakage into the lab room.

Functional Requirements:

• Prevent any traceable amount of drug vapor from escaping the chamber

• Allow other components to be attached and remain airtight

• Friendly and safe to the mice

• Reliable and intuitive to use

• Easy to manufacture and assemble

Designs Considered

As shown in option 1, the initial design decision was to construct the chamber out of rectangular acrylic pieces and connect them with acrylic solvent. This method requires very little design effort and it is the commonly used method to build aquarium boxes, which are strong and watertight. However, when we consulted Chris Cassidy, our instructor responsible for laser cutting, he suggested we put slots and tabs along the edges of the rectangular walls. We thoroughly considered the pros and cons of the slot and tab design. The obvious benefits are that it would assembling much easier and the tabs could help to align the walls more precisely. However, our concern was that the slots and tabs would create extra gaps along the walls which may be potential leak points. Since the acrylic solvent is thin and only bonds flush surfaces (without filling gaps), we also purchased a silicone sealant which can fill any gaps created by the slots and tabs. We created the slots and tabs design as shown in option 2. All of the chamber walls were laser cut from 0.25” clear acrylic. We assembled them with the acrylic solvent by first aligning three walls together to make a corner, taping the walls together, applying acrylic solvent using a needle tip bottle, and holding the walls together for 3 minutes. We then added one wall at a time until all walls were fully assembled. After the chamber was constructed, silicone sealant was added to any gaps.

Performance

When assembling the chamber, we immediately saw the benefit of the slot and tab design. The tabs made it easy to hold the walls together when gluing them. Using the acrylic solvent to bond the walls made the chamber very sturdy. Also, the acrylic solvent and silicone sealant showed very impressive sealing results. Although we had several leakage points that were found in the first prototype through leak testing, they were all eliminated in our second prototype by applying the sealant more carefully and blocking any leakage points with extra silicone.

2. Automatic Mouse Door

A mechanical door mechanism was built to allow the mouse access to the operant chamber without the presence of lab staff and to seal the opening during the drug vapor session. This was an especially challenging component to design because the door has to provide enough force to effectively seal the entrance, but it cannot hurt or startle the mouse. Multiple designs were tried out in consideration of the space limitations, effectiveness of the sealing, power of the motor, and reliability.

Functional Requirements:

• Produce an airtight seal for the mouse entrance

• Does not pose any risk to the mouse

• Does not allow mouse to access the electronics and rubber seal

• Easy to access for maintenance

Designs Considered

During initial brainstorm, We came up with three types of door mechanisms shown in figures above: a sliding door, a rotating door, and a swing door. We discussed these options with our sponsor to evaluate the practicality of each mechanism, and found that they were especially fond of the sliding door design which is shown in option 1. The idea was to make the door slides side to side with a rack and pinion actuation method. The benefit is that the liner motion does not take much space inside the chamber. However, after more considerations within the team and advice from instructors, the idea was rejected because there would likely be too much friction when rubber sealing was installed, and the rack and pinon system is generally unreliable and complex to build.

We quickly decided that the swing door mechanism was the most viable option as it can create a pressing force on the seal and the rotating mechanism is straightforward to build. We initially put the swing door inside the chamber, as shown in option2 , because there would be a metal tube connecting to the door entrance from outside. However, several concerns were raised by the sponsors on this design. First, when the door is open, the mouse will likely climb on it and could chew on the electrical wires. Second, the door's high speed and torque poses a threat to the mouse's safety if it runs into the door while it is closing. Third, the door takes up a significant amount of space inside the chamber and the swing motion may interfere with other components.

We quickly redesigned the mouse door that cleverly solved all the mentioned concerns. As shown in option 3, the door was moved outside the main chamber and a transfer case was added to enclose the mouse door and connect to the metal tube. With this design, the mouse will not be able to climb onto the door, the swing motion does not take up any space inside the chamber, and the shutting motion will be very unlikely to hurt the mice because the mice cannot run to the door very quickly.

Performance

The outside mounting swing door design was implemented in the first prototype. After testing with servo motor, it was found that bottom of the mouse door did not get pressed enough against the wall to create an effective seal. After more testings and observations, several reasons were proposed to account for the failure of the mouse door: the motor is not strong enough; the door arm is not placed at the optimal position to exert force on the door; the rubber seal in the door groove may be too thick and dense.

To tackle this problem, we decided to buy a stronger motor and redesign the mouse door. Attempts were made to tweak the angle of the door arm so that it presses the lower part of the door. At the same time, an alternative door design as shown in option 4 was developed, which implements a simpler design and has less moving components.

Another problem that we found while testing the first prototype was that the servo motor made a loud noise while holding the door shut to compress the seals. The sponsors want to be able to control all the sound generated inside the chamber so it does not affect the results of their experiment, so they requested that the noise from the motor be reduced. We planned to increase the size of the housing for the motor and add in soundproofing foam to reduce the noise and vibrations.

Final Design

Ultimately, we kept the original design, modified the arm angle and used a stronger motor, as shown in figure above. The modified mouse door is able to create a strong seal. Additionally, the new motor is almost completely quiet while holding the door closed, so we did not need to add in the soundproofing foam.

The mouse door was constructed by 3D printing the base plate, which included the servo case and the extension tube, and the mouse door separately. Since we were concerned that the commonly used 3D printing material would not be airtight, the parts were printed on an Objet350 printer using a material called VeroClear that is solid and similar to acrylic. The motor arm was laser cut separately from 0.25" clear acrylic. The base plate was mounted onto the outside of the chamber using counterbored holes. The base plate and the mouse door are connected by two bolts and nuts through designed hinge pieces. The holes in the hinge pieces are close fit clearance holes in order to minimize wobble of the door during operation. Additionally, a rod was press fit into the motor arm and sticks out into the slider, which allows the motor arm to rotate the mouse door.

3. Removable Chamber Wall

One of the chamber's walls has to be removable so that researchers can access the interior of the chamber to clean and maintain it. However, it still has to maintain an airtight seal when it is closed. We carefully looked for latches and hinges that can meet our requirements.

Functional Requirements:

• Open and close conveniently

• Create an airtight seal

Designs Considered

We purchased latches that have long latching distance to accommodate the rubber gasket we put around the door. We initially considered a design with two hinges at the bottom for the door to open. However, we realized in practice that it was difficult to perfectly align the two hinges. When the hinges were misaligned, it created gaps at the bottom of the door. We then decided to only have one hinge at the bottom and have slots on one side to be able to adjust the hinge slightly to ensure there were no gaps.

Performance

In practice, the removable wall was one of the most difficult components to make airtight. Due to inconsistencies in the chamber walls because of laser cutting, the door did not sit perfectly flush against the chamber. Additionally, the latches compressed the seals and removed any gaps on the top and sides of the chamber, but we saw gaps at the bottom of the door, especially in the bottom corners. In order to fill the gaps, we had to increase the thickness of the gasket and also applied silicone sealant. We also had to adjust the hinge placement using the slots many times during assembly as we tried to decrease the gaps. This made it harder to fasten the latches, since we had moved the door much closed to the chamber. In the end, we eliminated all gaps; although, the latches require some force to close.

Final Design

As shown in figure above, the final design is a door that opens from the top with one hinge at the bottom. There are four latches to close and open the door, two at the top and one on each side. There is a 0.25" wide strip of rubber around the edge of the door that creates a seal when the door is closed. The latches compress the rubber to create a seal.

4. Speaker and Case

The sponsor wanted a removable speaker for sound cues for the mice.

Functional Requirements:

• Removable

• In-line with the chamber wall

• Sound isn't muffled by a wall

• Connectable to Med-PC

Final Design

To accommodate the speaker, we created a small case that is airtight. The speaker is held in place by two threaded rods that thread into the back of the chamber and are secured by nuts. There is an acrylic front cover also mounted on the rods and secured by nuts. The front cover prevents the mice from climbing behind the speaker, but it has a circular cutout so that the sound is not muffled. The speaker connects to Med-PC through an auxiliary connector that placed in a small hole in the back of the speaker and sealed with silicone.

5. Pellet Receptacle

The sponsors asked us to incorporate a pellet receptacle and attach the dispenser that they had previously purchased. The dispenser delivers food pellets into the receptacle as a reward when the mouse completes certain tasks. The process is automated by Med-PC. To mount the pellet receptacle, there is a cutout in the chamber wall to allow the mice to access it from inside. Around the edges of the cutout, we used adhesive silicone to attach the plate onto the outside of the chamber. The advantage of using silicone adhesive here rather than bolts is that we would not have to drill holes in the pellet receptacle. Additionally, the silicone adhesive works not only as a means of mounting the receptacle, but also as a sealer for the mounting seam. In order to remove the adhesive, heat from a heat gun or hair dryer may be applied to the mount seam.

6. Waste Pan and Grid Floor

Since mice are clean animals and do not like to walk in their waste, it is important for the chamber to have a floor that does not allow waste to collect. A waste pan and grid floor were integrated into the chamber for easy clean-up of the mice droppings.

Functional Requirements:

• Floor does not allow waste to collect

• Tray is removable for easy cleaning

• Compatible to the chamber

The waste pan and grid floor were purchased from Med Associates. The waste pan was placed on the bottom of the chamber and can be removed easily for cleaning. The grid floor was placed over the waste pan and needed to be secured onto the chamber walls.

Figure above shows our design considerations for the grid floor mounts. Our first design idea was to laser cut small rectangular mounts that would be fixed on the chamber walls using acrylic solvent. Holes cut in the acrylic mounts would allow the grid floor to be fastened to the mount. However, we thought that the acrylic pieces could snap off if the lab staff frequently removed the grid floor to clean the grid. We decided to use metal L brackets which will hold up better with long term use.

7. Nose Pokes

One of the most important components of the chamber are two nose pokes which were purchased

from Med Associates. There are photoelectric sensors inside the nose poke that can sense the

mouse’s nose. Since the nose poke is from Med Associates, it can be connected to the Med PC

software and can trigger the vaporizer to deliver the drug to the mice.

Functional Requirements:

• Mouse can easily put nose into the hole

• Vaporizer and exhaust vacuum can be connected

• Mouse nose can trigger vaporizer to deliver drug

Design Considerations

Since we planned to mount the nose pokes on the outside of the chamber, there needed to be a cutout in the chamber wall for the mouse to access the nose poke hole. The figure above shows the wall cutout options. We first designed a circular hole cut around the nose poke. The idea of the circular hole cut was that the mice would step onto the hole cut and put their nose into the nose poke. However, some of the mice have a head implant and the sponsor did not think that there was enough space around the circular hole cut for the implant. Therefore, we maximized the space around the nose poke so that the mice can have enough space to access the hole.

8. Pressure Gauge and Vacuum Relief Valve

Because the chamber is being vacuumed from the exhaust, we anticipated that the pressure inside the chamber may decrease because it is airtight. We needed to ensure that it would always be safe for the mice to be inside the chamber, so we added a vacuum relief valve that is mounted on the top of the chamber. The relief valve is a one-way mechanical valve that opens when the pressure drops below the threshold pressure and allows air to flow inside the chamber and restore the pressure. We intend to set the threshold pressure at slightly below atmospheric pressure. We also mounted a vacuum pressure gauge that displays the pressure inside the chamber. This will allow the lab staff to visually confirm that the pressure is at a safe level for the mice.

Functional Requirements:

• Visual way to check pressure in chamber to ensure mice’s safety

• Maintain the pressure inside the chamber at a safe level for the mice

Figure above shows the pressure gauge and relief valve that we chose. The pressure gauge can show negative pressure. Figure 15 shows how the pressure gauge and relief valve were connected to the chamber. A tee fitting was threaded into the top of the chamber. The relief valve and pressure gauge are threaded into the tee fitting.

9. Vaporizer

Since the drug is stored in a liquid form, a vaporizer is essential for this system to send drug vapor to the mice.

Function Requirements:

• Vapor must be sent to mice as quickly as possible.

• The device should be able to interface with Med PC in some way.

Figure above shows the chosen vaporizer, which is developed and distributed by Scientific Vapor. This device was chosen for two main reasons. First, it is capable of sending vapor just one second after being triggered. This is important because it is difficult to train mice to hold their noses in the nose poke holes. The mice are able to hold their nose in the hole for about 5-10 seconds. However, if they are not rewarded with vapor right away, they may not remain in the nose poke long enough to actually receive the vapor. The second reason this particular device was chosen is because it interfaces directly with Med-PC. This greatly streamlined the technical communication setup required to trigger the vaporizer with the nose poke, because the lab members can use the Med-PC user interface to directly control the vaporizer.

10. Pump

The vaporizer does not exert any pressure to drive the flow of vapor through the tubing to the nose poke, so we had planned for the vacuum exhaust to drive the vapor flow. However, the vacuum exhaust does not directly connect to the vaporizer; instead, it connects through the nose poke, which has a hole that opens to the rest of the chamber. Because of this, when we tested the second prototype in the lab, we found that the exhaust did not create enough suction to drive the flow of vapor. To solve this, we added a pump in line with the vaporizer to pull the vapor into the nose poke.

Functional Requirements:

• At least 3 liters per minute (LPM) flow rate.

• Does not leak vapor

• Quiet while running

Testing and Performance

Because of the challenging nature of making a chamber airtight while attaching so many different components, we built and tested two prototypes before making the final design. Improvement and adjustments were made based on the results of each leak test, in-lab vapor test, and compatibility test that were aimed to make the design more suitable for the lab and mice.

Leak Test

The main test for the chamber was a leak test that was performed by covering all possible leak locations (wall connections, and component-wall interfaces) with a 1:1 soap to water mixture. Subsequently, a bike pump was used to pressurize the chamber through the input port of the Nose Poke. Any bubbles that formed at the potential leak locations signified a leak that would need to be sealed with silicone. Even though the pressure is not be congruent to the vacuum that the chamber will endure in the lab setting, this home pressure test was the most accessible way to effectively determine leaks.

Test Method:

In order to conduct the leak test, a barbed Schrader valve was connected to a tube inlet on the Nose Poke, and a bike pump attached to the Schrader end of the valve. For the first prototype, rather than drilling holes in the Nose Poke and bolting it to the chamber as planned, the component was instead taped to the vapor port hole. The purpose of this was to avoid permanently altering the component in case anything aspect of the design changed. Additionally, because the mouse door was still in development, it was instead manually pressed into its rubber seal for the leak test. When leak testing the final design, all the attachments including the mouse door were properly installed.

Test Result:

The soapy water leak test successfully resulted in very large and long lasting bubbles at leak points. It must be assumed, however, that smaller leaks may not have been detected when larger leaks are present. This may be remedied by iteratively using the soapy water leak test to detect, repair, and detect more leaks until no more leaks are detected. Even though the pressure applied by the bike pump is not the same direction and magnitude of force that will be applied by the lab vacuum, it is sufficient for determining if the chamber is air tight as required.

Several leaks were detected on the prototypes, because of the gaps that were caused by the slots and tabs and slight misalignments of the bolt holes for the latches with their respective hooks ended up forcing the wall gaps wider when the chamber walls were latched together. For the final design, we improved the techniques of applying acrylic solvent and silicone paste which was very effective at covering all gaps. The mouse door which was a big area of concern was also tested to be completely sealed. However, there is one component that was inherently not airtight, which is the container that holds the food pellets. It is connected to the chamber by a tube, and its lid is not airtight. However, given the low operational air flow rate, the pellet system provides enough barrier to be a low priority leak point.

Vapor Test

One of the biggest unknowns going into the lab test of the second prototype was the flow rate of the house vacuum. However, this is also one factor that has a great impact on our system. If the vacuum is too fast, the vapor will flow too past the mouse, causing it discomfort and potentially making it uninterested in the drug. If the vacuum flows too slow, the vaporizer may not be able to supply vapor through the nose poke. The goal of this testing was to see whether the vapor would be able to reach the chamber. If it does a secondary vacuum pump will need to be integrated.

Test Method:

We used a laser pointer to be able to see the vapor in the chamber. Seeing the line of the laser pointer indicates vapor flow into the chamber. Figure above shows the laser pointer testing. The view is from the top of the chamber looking down. If the red line from the laser pointer is very visible, it means that a significant amount of vapor flowed into the chamber.

Test Result:

First, we kept the exhaust closed to see the strength of the laser pointer if there was no vacuum exhaust. Figure on the left shows the red line from the laser pointer was very visible, which meant that a significant amount of vapor flowed into the chamber. Then we opened the exhaust to have a vacuum. The goal of this test was to show that the vapor was correctly flowing into the nose poke and out to the vacuum exhaust. Figure on the right shows a very small hint of the red laser pointer. Since the red laser pointer was less visible than when the exhaust was closed, this shows that the exhaust was pulling the majority the vapor out of the chamber, which is a successful test. Although there were small flickers of the laser pointer, meaning that there was still a small amount of vapor flowing into the chamber, in operation, the mouse will have its nose in the hole which will minimize the vapor in the chamber.

Compatibility Test

After the test stages of our project, the last step was to enable all the components to talk to each other. The diagram above shows the basic control logic. The platform we’re using for this is a software called Med PC which is ideal because the lab is already familiar with it, and additionally the majority of our components (which are shown above on the right) were actually built to integrate with Med PC. So we just had to deal with the remaining components on the right side the diagram. To control the typical actuators that can be controlled by Arduino, a TTL box has to be integrated to received instruction from Med PC software and then transmit the signal to Arduino which carries out the preprogrammed actions.

Test Method:

All of the components controlled by Med-PC were plugged in at the back of the box. The vaporizer was placed next to the chamber and was connected by tubing to the nose poke. The exhaust vacuum was also connected to the nose poke through tubing. The vaporizer was tested with the solvent without any opioids in order to ensure safety during testing. The solvent is a 80/20 mixture of vegetable glycerol and propylene glycol. We prepared a Med PC script for our testing, which would activated all components once the nose poke is triggered.

Test Result:

Every components were able to be controlled as expected after adjusting some configurations and debugging the script. The mouse door can close quickly but can hold the seal strongly and produced loud noise. We replaced the motor with a stronger one and solved all the concerns. The vaporizer is able to deliver vapor but at a very slow rate. We had to install a vacuum pump between the vaporizer and nose poke to drive the vapor flow.