Quantitative analyses of products and rates in polyethylene depolymerization and upcycling

Summary Depolymerization and upcycling are promising approaches to managing plastic waste. However, quantitative measurements of reaction rates and analyses of complex product mixtures arising from depolymerization of polyolefins constitute significant challenges in this emerging field. Here, we detail techniques for recovery and analysis of products arising from batch depolymerization of polyethylene. We also describe quantitative analyses of reaction rates and products selectivity. This protocol can be extended to depolymerization of other plastics and characterization of other product mixtures including long-chain olefins. For complete details on the use and execution of this protocol, please refer to Sun et al.1


STEP-BY-STEP METHOD DETAILS
In this section, we describe in detail how to accomplish polyethylene depolymerization and the steps for recovery and quantification of hydrocarbon reaction products.

Polyethylene depolymerization
Timing: Variable -24 h for our study Depolymerization was carried out in a batch reactor at a set temperature under an inert atmosphere.
1. Prior to each reaction, place a stainless-steel autoclave (90 mL), its lid and a Pyrex-encapsulated stir bar into the antechamber of an Ar-filled glovebox and evacuate for 12 h, then bring the autoclave into the glovebox. 2. Inside the glovebox, load the stir bar, the desired masses of polyethylene (typically, 0.120 g), and a freshly reduced Pt catalyst (0.200 g, Pt/F-Al 2 O 3 as an example) into the autoclave. a. Mix the solids with a spatula. b. Seal and remove the reactor from the glovebox.
Note: The solid catalyst can be pre-activated under H 2 (please refer to Sun et al.) 1 and stored in the glovebox prior to use.
Note: To accelerate the initial mass transport of PE and promote heat transfer to the polymer, grind and sieve the polymer into a fine powder (250-425 mm) before loading it into the reactor.
Cryogenic grinding: Under air, place the polymer beads (5-10 g) into a clean Dewar flask (500 mL). Next, add enough liquid N 2 to cover the beads. Allow the majority of the liquid N 2 to boil off, then promptly pour the beads into a coffee grinder and grind for a few seconds to obtain polymer powder.
3. Place the sealed reactor inside an external vessel heater, insert a thermocouple through a feedthrough into the reactor. a. Set the desired temperature, stirring rate, and reaction time (typically, 250 C, 675 rpm, 8 h, Figure 1A). b. Start the reaction timer when the reactor reaches the set temperature ($25 min).
Note: If the internal thermocouple does not touch the polymer/catalyst mixture, it may measure the internal gas temperature rather than the temperature of the solid/liquid phase. We also used a thermocouple to measure the external temperature of the autoclave wall and found that, typically, the temperature of the reactor wall is 280 G 5 C, i.e., 30 C higher than that of the gas inside the reactor).

Instruments/Techniques Equipment
Gel Permeation Chromatography GPC, equipped with a refractive index (RI) detector, calibrated with Varian monomodal, linear PE standards. Columns: PL-Gel Mixed B Guard column, three PL-Gel Mixed B columns.
Gel Permeation Chromatography GPC, equipped with Waters 2410 RI detector and Waters 2998 photodiode array detector (PDA), calibrated with polystyrene standards (Agilent EasiVial kit, molecular weights in the range of 200-400,000 g mol À1 ). Columns: two PL-Gel MiniMIX-D column, a guard column.
For detailed information regarding instrument parameters, please refer to Sun et al. 1 ll

OPEN ACCESS
Note: The temperature at which the depolymerization reaction occurs relies on the measurement of the external temperature of the reactor. For autoclaves that do not contain an internal thermocouple, the desired temperature can be set through the external heating system and the temperature can be measured externally. 4. After the desired reaction time, remove the reactor from the vessel heater and quench the reaction by immersing the reactor vessel in a water bath at 20 C-25 C for ca. 30 min.

Recovery and quantification of hydrocarbon reaction products
Timing: $12 h for workup and analyses These steps describe the recovery and quantification of each class of hydrocarbon products (gas, liquid, and solid) generated by polyethylene depolymerization.
5. Connect the autoclave and a Schlenk flask (typically, 100 mL) capped with rubber septum #2 to a Schlenk line equipped with a vacuum gauge and rubber septum #1, as shown in Figure 1B. 6. Evacuate the Schlenk flask and line (% 10 À2 Torr), then isolate both from the vacuum pump. 7. Expand the gases from the autoclave headspace into the line and Schlenk flask. Isolate the Schlenk flask.
Note: Wait an additional minute or so after the pressure reading stabilizes in the Schlenk line before closing the Schlenk flask, to ensure all gas components are well-mixed.
CRITICAL: Ensure the pressure of the line and the Schlenk flask will be lower than 1 atm after gas expansion, to avoid ejecting the rubber septum. The pressure in the reactor, post-reaction, can be measured using a pressure transducer, and this value can be used with the ideal gas law to estimate the pressure that will be generated in the line and the Schlenk flask after gas expansion. Adjust the size of the Schlenk flask so that the pressure after expansion does not exceed 1 atm.
8. Remove an aliquot of gas (400 mL) via rubber septum #1 ( Figure 1B) using a Luer lock gas-tight syringe, a. Inject the aliquot into the GC-FID to record the initial chromatogram. b. Perform this step twice to ensure consistency of the result. 9. Inject propene (400 mL, 200-400 mbar) as internal standard into the Schlenk flask via rubber septum #2 ( Figure 1B).
Note: Pull and push the syringe plunger several times to ensure good mixing between the internal standard and the gases in the Schlenk flask.
10. Remove an aliquot (200 mL) of gas from the Schlenk flask using a Luer lock gas-tight syringe. a. Inject into the GC-FID for light hydrocarbon analysis (C 1 -C 8 ).
Note: The difference in areas for the propene peaks in both chromatograms can be used to calculate the amount of propene present as a reaction product, relative to the amount added as an internal standard.
b. Perform this step twice to ensure well-mixing of the gaseous products and the internal standard. 11. Disconnect the reactor vent hose from the Schlenk line. Prior to opening the reactor, add 5 mL of methylene chloride (in 53 1 mL aliquots) via the vent hose ( Figure 1B).
Note: This step ensures that any liquid hydrocarbons that have condensed in the upper part of the reactor including the reactor's lid will dissolve and combine with the liquids present at the bottom of the reactor vessel.
12. Open the reactor, allow the CH 2 Cl 2 to enter, and transfer the solution and solid using a glass Pasteur pipette onto a Buchner filter funnel equipped with a fine glass frit (4.0-5.5 mm). 13. Add another 5 mL CH 2 Cl 2 to wash the solid remaining on the frit to increase the recovery of liquid hydrocarbons adsorbed on the solid residue in the filtrate. 14. Transfer the filtrate to a volumetric flask and dilute to 10.00 mL volume using CH 2 Cl 2 .
a. Filter $1 mL of the solution through a 0.2 mm PTFE filter attached to a 1 mL plastic syringe for analysis by GC-FID. 15. Evaporate the majority of the CH 2 Cl 2 solvent using rotary evaporation (30 C, 350 Torr, 15 min), then remove residual solvent under reduced pressure on a vacuum line (0.1 Torr, 1 h) at 20 C-25 C.
Note: These steps also remove volatile liquid hydrocarbons (mostly C 7 -C 11 ). 16. Weigh the mass of the remaining liquid products (C >11

EXPECTED OUTCOMES
In this section, we describe in detail the characterization of the hydrocarbon products formed in the catalytic conversion of PE (0.120 g, M w = 3.5 3 10 3 g mol -1 , Ð = 1.9) over Pt/F-Al 2 O 3 (0.200 g, 1.6 wt % Pt, 0.8 wt % F) at 280 C under Ar, as a representative example. Results for other catalysts were presented in Sun et al. 1 After 8 h, the yields of hydrocarbon gases (C 1 -C 8 ), volatile liquids (C 7 -C 11 ), and heavy liquids (C >11 ) are 7 (1) , 7 (1), and 64 (1) wt %, respectively. The solid residue (12 (1) wt %) is mostly coke, based on the temperature of its oxidation according to TGA. 1 The uncertainties are presented in parenthesis based on duplicates. Further characterization of the gas and liquid products is detailed below.
Gas chromatographic analysis of gaseous hydrocarbons GC-FID chromatograms of the gas products can be recorded before and after addition of an internal standard, shown in Figure 2. The initial chromatogram (Figure 2A) shows the presence of alkanes (C 1 -C 8 ) but no propene (absence of a signal at 3.2 min), suggesting that propene is a good choice for internal standard.
Gas chromatographic analysis of volatile liquid hydrocarbons (C 7 -C 11 ) GC-FID analysis of the liquid products dissolved in CH 2 Cl 2 can be conducted with and without solvent removal. The broad, intense solvent peak obscures other signals that may be present with retention times between 2 to 4 min ( Figure 3A). Comparison to the chromatogram after solvent evaporation ( Figure 3B) shows that most of the hydrocarbons lost during solvent removal are in the range C 7 -C 11 , which we refer to as volatile liquid hydrocarbons.  Quantitative NMR analysis of heavy liquid hydrocarbons (C >11 ) Solution-state 1 H and 13 C NMR spectra of the liquid products can be recorded at 20 C-25 C (we used a Varian Unity Inova AS600 spectrometer and a Bruker Avance NEO 500 spectrometer for 1 H and 13 C NMR, respectively). A typical 1 H NMR experiment was performed at an acquisition time of 2.5 s with 64 accumulated scans. The 1 H NMR spectrum reveals signals for aromatic protons at 6.5-9.0 ppm and for benzylic protons (H a , including the benzylic positions of alkyl substituents on fused aromatics such as naphthalenes and phenanthrenes) at 2.0-3.5 ppm, indicating the presence of alkylaromatics ( Figure 4A). 15,16 To ensure that the 13 C NMR information is quantitative, Cr(acac) 3 is added as a relaxation agent to reduce long 13 C spin-lattice relaxation times (T 1 ). The spectrum of a model compound, dodecylbenzene (60 mg), is measured with an inversion recovery pulse   Since all peaks in Figure 5 are positive for t R 0.8 s, the estimated value of T 1 is 1.2 s. Quantitative 13 C NMR spectra therefore require a relaxation delay t R 6.0 s. The NMR sample of heavy liquid hydrocarbons resulting from PE depolymerization was prepared at the same Cr(acac) 3 concentration and the T 1 and t values determined in the inversion recovery experiment with dodecylbenzene are then used to record quantitative 13 C NMR spectrum at an acquisition time of 1.5 s with 4096 accumulated scans. Figure 4B shows signals for aromatic carbons in the region from 118 to 150 ppm. 18 Their integration will be used to estimate aromatic yields and selectivities (see below).

GC-MS analysis of heavy liquid hydrocarbons (C >11 )
GC-MS can be used to determine the average carbon number for different classes of hydrocarbons present in complex mixtures. Characteristic ion chromatograms for each type of hydrocarbon (alkylbenzenes, alkylnaphthalenes, alkylphenanthrenes, and alkanes) are extracted from the total ion chromatogram, as shown in Figure 6.
Size exclusion chromatographic analysis of heavy liquid hydrocarbons (C >11 ) While PE analysis generally requires access to high temperature gel permeation chromatography (GPC), it is possible to analyze low molecular weight PE and depolymerization products using more widely available room temperature GPC, performed in a solvent such as chloroform or tetrahydrofuran. The instrument is typically equipped with RI (refractive index) and/or UV detectors. The system is usually calibrated with either PE or PS standards (the use of more readily available PS standards requires a conversion process to obtain PE molecular weights). 19 Typical results for the heavy liquid products resulting from the conversion of PE (0.120 g, M w = 3.5 3 10 3 g mol -1 , Ð = 1.9) catalyzed by Pt/F-Al 2 O 3 (0.200 g, 1.6 wt % Pt, 0.8 wt % F) after 8 h at 280 C under Ar are shown in Figure 7A for a low molecular weight PE and its depolymerization products , revealing a decrease in M w and dispersity (Ð = M w /M n ). A comparison of results using RI and UV detectors shows how the UV-active chromophores are distributed relative to the total hydrocarbons ( Figure 7B). In this case, both detection methods give similar distributions, suggesting that aromatic products are evenly distributed over the entire molecular weight range of liquid hydrocarbon products.

Thermogravimetric analysis of solid hydrocarbons
To characterize solid hydrocarbons, TGA is performed in air while ramping the temperature at a heating rate of 10 C min À1 from 50 C to 700 C. Intact PE is oxidized at ca. 300 C, while oxidation

QUANTIFICATION AND STATISTICAL ANALYSIS
In this section, we describe methods to quantify the overall mass balance as well as specific types of hydrocarbon products in the gas, liquid, and solid phases. It includes the method for calculating the selectivities of various types of hydrocarbons (aromatics vs. alkanes, etc.) in the heavy liquid products. We also describe how to assess the depolymerization rate, by determining the rate of C-C bond scission.

Hydrocarbon distribution
Quantification of gas products (C 1 -C 8 ) 1. Calculate the individual and total amounts of gas products based on the GC-FID analysis. a. Assign the peaks for each carbon number (i) according to their retention times (assigned by comparison to individual norm-alkane standards from C 1 to norm-C 8 ).
i. Assign all peaks that elute between norm-C i-1 H 2i and norm-C i H 2i+2 to have carbon number C i .

Note:
The norm-alkane has the highest boiling point of all alkane isomers.
b. Calculate the amount of propene added using ideal gas law (Equation 3), where n propene is the moles of propene added as internal standard, P is the propene partial pressure, V is the injected volume, R is the ideal gas constant and T is the absolute temperature.
Note: Propene was selected as our calibration standard because we typically did not detect it among our gas phase products. If propene is a reaction product, it can still be used for calibration, but its yield will need to be quantified using the method of standard additions. c. Calculate the number of moles (n i ) and mass (m i ) contributed by each species to the gas phase products, using its peak area (A i ) relative to that of propene (A propene ) using Equations 4 and 5.
The volume expansion factor, V 2 /V 1 , is the total volume (V 2 ) including the reactor, Schlenk line, and flask, relative to the volume of the Schlenk flask (V 1 ).
Note: It is reasonable to assume the same relative carbon response factor for all hydrocarbons in GC-FID analysis. 21 d. Calculate the total moles and the total mass of gas products by summing the contributions of each gas phase species.
Quantification of volatile liquids (C 7 -C 11 ) and heavy (less volatile or non-volatile) liquids (C >11 ) 2. Calculate the individual and total amounts of volatile liquid product based on the GC-FID analysis. a. Assign carbon numbers to each peak according to its retention time, by comparison to the retention times for a standard mixture of norm-alkanes (C 7 -C 40 ). i. Assign all peaks eluting between norm-C i-1 C 2i and norm-C i C 2i+2 to have carbon number i. b. For each carbon number in the region corresponding to C 7 -C 11 (Figure 3), calculate the difference in peak areas (A) before and after solvent removal by evaporation. Calculate the moles of each species with a given carbon number species (n i ) using the external standard method according to Equation 6, where V 0 is the solution volume and f is the relative carbon response factor.
Note: norm-octane in CH 2 Cl 2 can be used as an external standard. A calibration curve is constructed using the peak areas for norm-octane at different concentrations.
Note: It is reasonable to assume the same relative carbon response factor (f) for all hydrocarbons in GC-FID analysis. 21 c. Estimate the mass of each species with a given carbon number (m i ), assuming its molecular weight to be that of the alkane (Equation 7). Here, we assume no mass loss of catalyst compared to the initial charge. Note: Since no mass loss was observed by TGA between 600 C and 700 C (Figure 8), coke was assumed to be fully oxidized below 600 C and the remaining mass (m 600 %) was considered to correspond to the catalyst. Depending on its structure, complete oxidation of coke may require temperatures higher than 600 C.
d. Use the results from step 6c to obtain the percentage of mass of solid carbon residue (unreacted PE and coke) relative to initial PE, Equation 14.
e. Check that the mass of carbon residue derived from the TGA measurement is consistent with the total solid hydrocarbon mass obtained directly by weighing in step 5 (Equation 10).
Average rate of C-C bond scission Note: Using M n,PE to calculate the number of chains in the solid residue is inaccurate if the residue contains partially depolymerized, but still insoluble, polymer chains. However, these chains contribute little to the number of chain cleavage events, relative to the gas and liquid products, therefore the error is small relative to other uncertainties inherent in the method.
Note: If the mass balance is incomplete (i.e., less than 100 wt %), the missing mass must be assigned to one or more hydrocarbon components prior to using Equation 16. We assumed the missing mass to be unreacted PE. If a different assumption is made, the masses of the other presumed contributions to the total product mass should be scaled as appropriate.
Note: Typically, the overall mass balance is 90 wt % or better using the mass recovery method described above. A poor mass balance will lead to a large error in the calculation of number of moles of C-C bond scission events.
b. Calculate N(0) using Equation 17, where m initial PE is the initial mass of PE and M n, PE is its initial number-averaged molecular weight.
Nð0Þ = m initial PE M n;PE (Equation 17) 8. Calculate the average rate of C-C bond scission (r C-C ) using Equation 18.
Note: This approach assumes the rate law is pseudo-zeroth-order in the number of C-C bonds, which is appropriate for our reaction. If the reaction order is not zero, Equation 18  Note: 1 H NMR analysis does not readily differentiate naphthenes from alkanes, or alkyltetralins from alkylbenzenes. For simplicity, the possible naphthenes and alkyltetralins contributions are simply combined with the contributions of alkanes and alkylbenzenes, respectively, in calculating alkane and aromatic selectivities.
Note: Further identification of individual molecular components can discerned from the mass (FD-MS) 6 and mass fragmentation (GC-EIMS) patterns for alkanes, cycloalkanes, alkylbenzenes, alkylnaphthalenes. Note: More information about the types of aromatic ring structures present in the liquids can be obtained by mass spectrometry and UV-vis spectroscopy. For example, three-ring structures were assessed as likely alkylphenanthrenes rather than alkylanthracenes and the presence of larger fused aromatics (e.g., chrysenes) was deemed unlikely on the basis of the UV-vis spectra. iii. Determine the average number of alkyl substituents (n) by comparing the experimental ratio H a /H arom and C arom-H / C arom-C to the predicted value for the major type of aromatic product (or products, as appropriate), according to the value of q found in step 10b. The example of q = 2 (i.e., on average, C a H 2 R substituents) is shown in Figures 9 and 10. For simplicity, we assume that each type of aromatic product (e.g., alkylbenzenes, alkylnaphthalenes) in the liquid product mixture has the same number of alkyl substituents (see limitations below). Thus, the lower and upper bounds of H a /H arom for a mixture of alkylaromatics are enclosed in the red, dashed box ( Figure 9). For H a / H arom = 1.5, the number of substituents could be 3 or 4 (dashed line in Figure 9), depending on the relative concentrations of alkylbenzenes, alkylnaphthalenes, and alkylphenanthrenes. Combining this information with the experimental ratio C arom-H / C arom-C (where C arom-H / C arom-C = 1.0 in our study), we estimate that each aromatic molecule has, on average, 3 alkyl substituents.
Note: This method works best when n % 4, or the major types of aromatic products are % 3 rings. Alkylaromatics with more than 4 substituents or higher polycyclic aromatics are not considered here.
d. If desired, predict the likely locations of the alkyl substituents based on the most plausible mechanism of aromatic ring formation. 1 Note: It is challenging to precisely identify the locations of substituents on aromatic rings present in a complex mixture of hydrocarbon products. Thermodynamics can provide guidance about the most stable substituent positions. For our methyldialkylaromatic products, we proposed the following plausible structures, based on the likely formation mechanism 1 : e. Estimate the average carbon number for each type of hydrocarbon product, using MS i. Calculate the number-averaged carbon numbers for different types of hydrocarbons (e.g., alkanes, alkylbenzenes) using mass spectrometry with either hard ionization (GC-EIMS, with ion chromatograms isolated for individual classes of hydrocarbons) [22][23][24][25] or soft ionization (FD-MS, with the corresponding molecular ion chromatograms). An example of data obtained using GC-EIMS is shown in Figure 6, where the individual ion chromatograms for alkylbenzene, alkylnaphthalene, alkylphenanthrene, and alkane products ( Figures 6A-6D) were generated by summing the individual ion intensities (I) for each ion characteristic of a particular product group. The results allow us to calculate a separate number-averaged carbon number for each hydrocarbon type.
Note: Alkylbenzenes do not elute in precisely the same range as alkanes with the same carbon numbers. While branched alkanes typically elute just before the norm-alkane with the same mass, alkylbenzenes may elute either before or after the corresponding norm-alkane. Most alkylbenzenes have slightly longer retention times than the norm-alkanes, with retention times mostly between those of the norm-alkanes with carbon numbers i and i+1. 26 Thus, when alkylbenzenes are a major component of the products, the average carbon number may be underestimated by ca. 1. This error is generally less than the accuracy of the analysis. Alkylnaphthalenes and alkylphenanthrenes generally elute later than alkybenzenes with the same carbon number, although they usually represent a small fraction of the alkylaromatic products and therefore do not have a large impact on the calculation of the average carbon number.
ii. The calculations use one of two equations for each type of hydrocarbon, depending on the assumption made about the MS response factors (in the absence of individual response factors for each species). The example shown here applies to alkylbenzenes. Equation 20 assumes that the MS response factor is a function only of the type of hydrocarbon (e.g., alkane vs. alkylbenzene), not on its molecular weight (M). Alternatively, Equation 21 assumes that the MS response factor for each type of hydrocarbon also depends on its molecular weight. Typically, the estimated average carbon number calculated for each type of hydrocarbon is similar using either method (see Sun  Note: Assign the carbon number of each species using norm-alkane standard compounds, according to the GC-FID method described above (see hydrocarbon distribution).
As an alternative to step 10e.i, the average carbon number for heavier hydrocarbons can be identified by GPC using Equation 22, where i represents the number-averaged carbon number. Note: GPC analysis can be performed using both RI and UV detectors to reveal the distribution of aromatics (which absorb light) relative to the total hydrocarbon distribution (which includes alkanes that do not absorb UV light). The comparison may confirm that aromatics are evenly distributed across the molecular weight range of the liquid products, or reveal that aromatics are concentrated in a particular molecular weight range.
Note: For simplicity, we assume the molecular formula to be C n H 2n , based on the most intense signal at 1.
n ben ; n nap or n phe = À S ben ; S nap or S phe Á À m heavy liq M n;liq Á (Equation 33) n arom = n ben + n nap + n phe (Equation 34) 11. As an alternative to the method described in step 10, estimate aromatic yields and selectivities using quantitative 13  , where C total is the sum of C ali and C arom .
C total = n ben;C $ i ben + n nap;C $i nap + n phe;C $i phe + n alk;C $i alk (Equation 36) d. Calculate the fraction of aromatic carbons, C arom /C total , based on the integration of the quantitative 13 C NMR spectrum. The integrated ratio can be set equal to that established from the chemical formulas and molar yields for each aromatic structure type, in accordance with Equation 37 .
C arom C total = 6 n ben;C + 10 n nap;C + 14 n phe;C n ben;C $i ben + n nap;C $i nap + n phe;C $i phe + n alk;C $i alk (Equation 37) Note: Since quantitative 13 C NMR does not readily differentiate between types of aromatic rings, the molar ratios of the individual aromatic types must be assumed to be the same as those obtained from 1 H NMR analysis.
e. Calculate the selectivity for alkanes and for each aromatic structure type by solving Equations 32 and 37 simultaneously. f. Compare the total aromatic selectivity obtained using quantitative 13

LIMITATIONS
This protocol considers polyolefin depolymerization product mixtures that contain alkylaromatics (alkylbenzenes, alkylnaphthalenes and alkylphenanthrenes) and alkanes as the major components.

OPEN ACCESS
Other types of hydrocarbons including naphthenes, alkyltetralins, and compounds with multiple, unfused aromatic rings are potential contributors to 1 H NMR signals. For the purpose of this study, any naphthenes were combined with alkanes and all compounds containing aromatic rings were combined with alkylbenzenes in the calculations of selectivities towards alkanes and aromatics. However, if their contributions are significant, it could affect the corresponding calculations.
The locations of alkyl substituents on aromatic rings can be proposed based on plausible mechanisms for their formation (Scheme 1). 1 The number of alkyl substituents is assumed to be similar for each aromatic structure type, based on the depolymerization mechanism. This assumption may not accurately represent the precise contributions of aromatic protons from each structure type. Characterization of the number of substituents and their positions for each aromatic structure type would improve the accuracy of the calculations. This characterization might be accomplished using GCxGC or other advanced separation methods, if available.
The number-averaged carbon number i for each type of hydrocarbon determined by mass spectrometry based on fragmentation patterns is semi-quantitative, because hydrocarbons with different structure types and chain lengths may have different ionization efficiencies in the mass spectrometer.

TROUBLESHOOTING Problem 1
Unclosed mass balance due to incomplete recovery of gas products in Step 7 of the section titled ''recovery and quantification of hydrocarbon reaction products'' under ''step-by-step method details.'' This problem leads to an error in the calculation of number of C-C bond scission events.

Potential solution
Check for leaks before expanding the gas from the reactor to the Schlenk flask. Very small leaks in the reactor, glass joints, rubber septa or the syringe connecting the reactor to the Schlenk collection flask via the Schlenk line can cause significant gas losses. Therefore, the reactor itself and each part of the set-up should be carefully leak-checked.

Problem 2
Unclosed mass balance due to loss of volatile liquid products to the filtration in Step 13 of the section titled ''recovery and quantification of hydrocarbon reaction products'' under ''step-by-step method details.'' This problem also introduces error in the calculation of number of C-C bond scission events.

Potential solution
Either direct distillation or cold filtration of the liquid products after reaction may be performed as an alternative to vacuum filtration to recover the volatile liquid products.

Problem 3
Inability to quantify C 5 -C 7 hydrocarbon products in step 2 of the section titled ''hydrocarbon distribution'' under ''quantification and statistical analysis,'' due to overlap of their signals in the GC-FID with the solvent signal. This issue reduces the apparent recovery and underestimates the number of C-C bond scission events.

Potential solution
Methylene chloride can be substituted by norm-decane for extraction in an otherwise duplicate experiment. The mass of light hydrocarbons and the number of C-C bond scission events can then be evaluated by combining results for the different hydrocarbon ranges accessible using each extracting solvent.

Problem 4
Incomplete mass balance as a result of a leak in the reactor for Step 1 of the section titled ''polyethylene depolymerization'' under ''step-by-step method details.'' A leaky reactor may result in a poor mass balance and potential exposure of the catalyst to air, which would affect reactivity.

Potential solution
Since the depolymerization reaction is conducted under 1 atm of inert gas, it might be difficult to detect a leak in the reactor during reaction. Therefore, the reactor can be checked for leaks prior to use by pressurizing the reactor (> 1 atm) with inert gas and monitoring for pressure drop. The pressure should stabilize for an extended period. A leak typically happens at the connection between the upper part of the reactor and the reactor's vessel. Replacing the O-rings should remedy a leaky reactor.

Problem 5
Usage of methylene chloride (a hazardous chemical) as the extracting solvent in Steps 11-15, 17 and 19 of the section titled ''recovery and quantification of hydrocarbon reaction products'' under ''stepby-step method details.''

Potential solution
Hexane and ethyl acetate can be used as alternatives for extraction of the liquid product from the depolymerization reaction.

RESOURCE AVAILABILITY
Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Mahdi M. Abu-Omar (mabuomar@ucsb.edu).

Materials availability
This study did not generate new unique materials.

Data and code availability
The published article (Sun et al.) 1 and its extensive Supplemental Information include all data generated and analyzed during this study. No code is reported in this work.