Caution: Use all appropriate safety practices when performing an experiment under high temperature and pressure conditions, including the use of engineering controls (H₂ flow limiter, pressure regulators, and rupture disc assembly) and personal protective equipment (safety glasses, temperature-resistant gloves, lab coat, full-length pants, and closed-toe shoes). Consult all relevant material safety data sheets (MSDS) before use. Carry out micro-reactor loading and clean-up in a fume hood, as these steps involve the use of harmful volatile organic solvents (toluene and dichloromethane).

NOTE: Setup description (see supplemental file).

1. Micro-reactor Loading

1. Clamp the micro-reactor vertically and upside-down, with the bottom face-seal (thus positioned at the top) open.
   NOTE: At this stage, the sapphire window, the custom-machined magnet, the 1/16" ferrule, the brass pad, and the bottom nut should not be assembled yet.
   1. Ensure that the fittings that are used to connect the micro-reactor to the gas lines are closed.
2. Load around 0.6 g of sample into the reactor through the open face-seal using a thin spatula.
   1. If the sample is originally kept in a large container, make a sub-sample before loading the micro-reactor.
   2. To estimate the amount of sample loaded inside the reactor, weigh the container and the spatula before and after loading, and calculate the mass difference.
3. Slide the custom-machined magnet onto the thermocouple.
4. Slide the 1/16" front ferrule so that the larger circle is facing up.
5. Ensure that the sealing surface (i.e., the fitting groove where the sealing ring sits) of the bottom face-seal fitting is clean and dry.
   NOTE: Given the highly viscous nature of most heavy oil samples, it is very likely that the sealing surface got accidently smeared by the sample during the loading process.
   1. Dip the tip of a cotton swab in toluene and apply it to the sealing surfaces to clean them. Be careful not to drip toluene inside the reactor cavity, which would contaminate the sample.
   2. If cleaning with toluene is required, make sure the sealing surfaces are dry before proceeding to the next step.
6. Ensure that the sapphire window is clean and dry.
   1. If the sapphire window is dirty, use a cotton swab soaked in a suitable solvent, and then perform a final wash using acetone to clean the window surfaces; let it air dry.
7. Place the sealing ring on the sealing surface, then the sapphire window on the top of the sealing ring, and then the brass pad on the top of the sapphire window; it is preferable to apply tiny, pinhead-sized drops of lubricant on the brass pad.
8. Thread the bottom nut on the bottom face-seal fitting while encapsulating the brass pad and the sapphire window. Adjust the bottom nut until it reaches finger-tight position.
9. While holding the reactor upside-down, transfer it to a vice. Use a wrench to tighten the bottom nut by 90° from finger-tight position.
   NOTE: After this step, the reactor does not need to be held upside-down any longer.
10. Check the micro-reactor for potential defects in the seal.
    NOTE: The window may show some chips or cracks, or a faulty seal can be identified if the compressed surface of the seal on the window does not make a continuous circle.
    1. In case of a defect, open the micro-reactor for inspection.
    2. After taking remedying action, use a brand-new sealing ring when attempting to reseal the reactor.

2. Micro-reactor Installation

1. Once the micro-reactor is loaded and sealed, connect the reactor to the gas lines and perform tests for leaks.
   1. Always begin leak testing by using N₂ at a maximum pressure of 5 MPa.
      NOTE: The preferred method for leak testing is the pressure-decay test, where the setup is pressurized and then isolated from the cylinder (closing valves V2 and V3). If the pressure remains stable for an extended amount of time (more than 30 min), no leak is observed.
   2. Carry out additional leak tests at higher pressures if the target pressure for the upcoming experiment is higher than 5 MPa.
      NOTE: These additional leak tests can be performed with maximum pressure increments of 6 MPa until the desired pressure condition for the experiment is matched. Consider 16 MPa as the upper limit of pressure for both leak testing and setup operation.
      NOTE: If the gas used to pressurize the setup in the upcoming experiment is not inert (such as flammable gases), carry out another series of leak tests using the target gas contingent on a successful series of leak tests with N₂.
2. After successful leak tests, depressurize the setup before undertaking the next installation steps.
3. Place the micro-reactor in the stainless steel heating block, which is itself inserted in the coil heater. Position the assembly on the platform located above the microscope objective.
4. Encase the reactor, the heater, and the heating block with the two halves of a casing filled with ceramic wool. Clamp the two halves of the casing together using a hose clip.
5. Fine-tune the position of the reactor over the microscope objective.
   1. Turn the microscope on using cross-polarized light. Adjust the vertical position of the objective using the lowest magnification so as to focus on the inside surface of the sapphire window.
   2. Position the reactor so that the field of view at the lowest magnification (typically 50X) covers a radial portion of the window surface where the inner boundary comprises the edge of the 1/16" front ferrule, as described in Figure 1.
      NOTE: Actual micrographs acquired by the software should be centered subsets of this field of view, which would avoid showing the ferrule itself.
6. Connect the thermocouple of the micro-reactor (TT1) to the temperature controller (TIC1).
7. Turn on the motor driving the external magnet to a speed of 120 rpm.
8. Pressurize the setup to the desired set-point.
   NOTE: Atmospheric pressure runs are performed by opening all outlet valves to the vent. Batch experiments can be performed by closing valve V4. Experiments under a constant head of pressure (preferable for high-pressure conditions) can be carried out by using the back-pressure regulator PV2.

3. Regular Procedure for the Visualization of Cracking Reactions

1. Throughout the entire experiment, place the microscope objective under the reactor only when visualizing the sample or taking a snapshot. Avoid leaving the microscope objective under the reactor when it is not needed.
   NOTE: Leaving the microscope objective under the reactor at high temperatures will cause an artificial brightening of the images, resulting in poor data, and may lead to a deterioration of the objective.
2. Turn the temperature controller on and apply a temperature set-point of 200 °C. Once the sample temperature reaches 200 °C, perform a round of verifications.
   NOTE: A round of verifications entails verifying the pressure, temperature, reactor position, focal distance of the microscope objectives, and stirring. As the temperature changes, the platform supporting the reactor and the heating assembly deforms slightly, so the vertical position of the microscope objective must be adjusted for the sapphire/sample interface to remain in focus. Stirring can be detected by the motion of the 1/16" ferrule or of small heterogeneity in the sample (such as small mineral solids).
3. If everything is in order as the sample reaches 200 °C, perform a set-point change to 300 °C. Once the sample temperature reaches 300 °C, perform another round of verifications.
4. Repeat the previous step, with 350 °C as the new temperature set-point.
   NOTE: 350 °C can generally be considered as the higher temperature limit where cracking reactions are not significant (in the time scale of min).
5. Change the set-point temperature to the desired reaction temperature, generally in the 400-450 °C range.
6. After making the final temperature set-point change, begin monitoring the reaction and recording data at regular time intervals, preferably every min.
   1. Carry out each step of data recording as follows: rotate the nosepiece of the microscope to place the objective under the reactor. Adjust the focus. Take a snapshot. Rotate the nosepiece to move the objective away from underneath the reactor. Note the temperature.
      NOTE: For future quantitative image analyses, snapshots should be taken with consistent settings throughout the experiment, namely in terms of magnification, lighting conditions, and camera acquisition settings (photosensitivity response and exposure time). As guidelines, the micrographs presented in this manuscript were taken with 100X magnification, maximum lighting conditions (using a halogen lightbulb), linear sensitivity response of the camera, and exposure times ranging from 200-400 ms.
   2. Perform the data recording steps repeatedly for as long as needed.
      NOTE: Generally, the duration of the observation is guided by visual changes in the sample (color, brightness, and heterogeneity) or by an estimate of reaction conversion.
      NOTE: Preferably, avoid continuing the experiment after the formation of large amounts of mesophase coke (which makes the reactor more difficult to clean).

4. Shutdown and Clean-up

1. Terminate the experiment by turning the temperature controller and the stirrer off and depressurizing the setup. Let the reactor cool.
   NOTE: Cooling the reactor can be facilitated by removing the micro-reactor from the casing and out of the heating assembly. Applying a cool air flow to the micro-reactor can also make this process faster and easier.
   1. Once the micro-reactor is cooled to room temperature, disconnect it from the gas lines of the setup, place it in a vice to loosen the bottom nut, and unseal the micro-reactor.
2. In a fume hood, take the micro-reactor apart by removing the bottom nut, the brass pad, the sapphire window, the 1/16" ferrule, and the magnet. Remove the sealing ring.
   NOTE: Coke may have formed during the experiment, which may cause the 1/16" ferrule and the magnet to be stuck to the thermocouple.
   1. Use tweezers to pull the 1/16" ferrule and the magnets out. Use a spatula to lever the sealing ring out of the sealing groove. However, take care not to scratch the sealing groove in the process.
3. To remove the bulk of the material stuck to the micro-reactor walls, scrub the inside cavity of the micro-reactor with solvent-soaked (toluene or dichloromethane) pieces of paper towels. Repeat the process with pieces of emery cloth, preferably coarse grit (#100).
   NOTE: During this process, do not scratch the sealing surfaces. At the end of this step, the metallic shine of stainless steel inside the micro-reactor cavity should be apparent.
4. Remove the material stuck to the flat surfaces of the custom-machined magnet using a piece of emery cloth, preferably coarse grit (#100).
   1. Use a solvent-soaked 1/16" wire to remove the material stuck inside the hole of the custom-machined magnet.
5. Use solvent-soaked (toluene, dichloromethane, or acetone) cotton swabs to remove the material stuck to the sapphire window.
6. To remove the remainder of the material stuck to the reactor walls, including the sealing surfaces, use solvent-soaked (toluene or dichloromethane) cotton swabs.
   NOTE: The cleanup process is finished when, after scrubbing with a solvent-soaked cotton swab, the cotton swab comes out with negligible traces on it.
   NOTE: However tedious this process might be, this step is important to avoid cross-contamination between experiments.
7. Let the micro-reactor air-dry.

5. Image Analysis⁹

1. Extract information from micrographs pertaining to the mean values of the red, green, and blue (RGB) channels, as well as the corresponding information in the hue, saturation, and intensity (HSI) color space.
   NOTE: The HSI color space is described by cylindrical coordinates, where hue, saturation and intensity correspond to the angular, radial, and vertical coordinates, respectively. The relationships between the RGB values of a pixel and the corresponding HSI values are given by the following equations^12,13, where m is the minimum of RGB values, while α and β are the pair of chromaticity coordinates:

  Eq. 1

  Eq. 2

  Eq. 3

  Eq. 4

  Eq. 5

  Eq. 6