Separate basic chemicals into HVC, chlorine, methanol and ammonia

[nworbmot]

The category of "basic chemicals" comes from the JRC IDEES database. It consists to our knowledge of high value chemicals (HVC), chlorine, methanol and ammonia. HVC are used for plastics synthesis and consist of ethylene, propylene and BTX (benzene, toluene, and the three xylene isomers).

It is necessary to separate out these chemicals because current and future production routes are different. For example, the hydrogen necessary for ammonia synthesis could be replace by green hydrogen. As another important example, primary HVC production could be reduced by mechanical and chemical recycling of plastics.

Previously ammonia production had already been separated from basic chemicals.

In this pull request (PR) in addition methanol and chlorine are also separated, based on energy consumption data from (DECHEMA, 2017). What is left is HVC production. The resulting specific energy demand for HVC in resources/industry_sector_ratios.csv (2.9 MWh_elec/tHVC, 4 MWh_CH4/tHVC, 0.2 MWh_heat/tHVC, 14 MWh_naphtha/tHVC) is relatively close to the numbers quoted in (Lechtenböhmer et al, 2016) (4.7 MWh_misc_energy/tHVC, 14 MWh_naphtha_feedstock/tHVC).

In addition, the option to specify mechanical and chemical recycling of plastics is given. This reduces the primary HVC production, and has a dramatic effect on costs. Assuming plastics demand stays constant, in a net-zero-CO2-emission scenario for Europe, producing 100% of HVC via primary route is 73 bnEUR/a or 14% more expensive than a scenario with 20% primary production, 30% mechanical recycling, 10% chemical recycling and 40% reuse (reuse has no energy use). A reasonable circular economy scenario, following Material Economics (2019) page 133 could be e.g. 25-25-25-25% for these different options.

What about process emissions from recycling plastics? In the current model setup, this is taken care of in that all carbon in primary production lands in the atmosphere. Suppose x+y+z carbon units of plastics are produced in a year, x primary, y mechanical and z chemical. Feedstocks required: x(1+c) primary (feedstock: naphtha), y(1+d) mechanical (feedstock: waste plastics), z(1+e) chemical (feedstock: waste plastics). There are process emissions of xc for primary production, yd for mechanical recycling and ze for chemical recycling. Of the x+y+z units in produced plastics, y(1+d)+z(1+e) are recycled for next year and x - yd - ze are burned in waste-to-energy plants. Thus total emissions are xc + yd + ze + (x - yd - ze) = x(1+c). The model currently allows the xc process emissions to be captured and the x goes into the atmosphere, correctly taking account of everything. We could allow process emissions from mechanical and chemical recycling to be captured.

Some loose ends that could be fixed:

• Process emissions from methanol production are ignored and are all landing at HVC. Could separate: (DECHEMA, 2017) has 1.5 tCO2/tMeOH with process emissions from feedstocks, i.e. originating from CH4 input. It's a very small effect. Separate process emissions from methanol production from HVC production. pypsa-eur#582
• Allow process emissions from mechanical and chemical recycling to be captured. These need to be specified first; at the moment they are landing in the atmosphere. Allow process emissions from mechanical and chemical recycling of plastics to be captured. pypsa-eur#583
• The separation of basic chemicals hasn't been applied to the scripts for "today's industry energy consumption". This is not necessary for future scenarios, but would be useful for future pathway investigations. Implement separation of basic chemicals for today's industry energy consumption pypsa-eur#581
• We need a reliable source for the energy demand for mechanical recycling. (Accenture, 2017), cited by (DECHEMA, 2017), has on page 35 something approaching 7 MWh/tHVC, which seems high, around same as chemical recycling.
• Adding waste-to-energy (WtE) plants to consume non-recycled plastics (x - yd - ze in above). I've already added this in a test, but the effect is small. WtE plants are expensive and only built if waste is forced to be used. The model chooses unabated WtE over WtE with carbon capture. Could argue this is outside system boundary and continue assuming all carbon in primary plastics lands in atmosphere. Add waste-to-energy (WtE) plants to consume non-recycled plastics pypsa-eur#580

References:

• Lechtenböhmer et al, 2016: Decarbonising the energy intensive basic materials industry through electrification - Implications for future EU electricity demand
• DECHEMA, 2017: Low carbon energy and feedstock for the European chemical industry
• Accenture (2017): Taking the European Chemical Industry into the Circular Economy
• Material Economics (2019): Industrial Transformation 2050 - Pathways to Net-Zero Emissions from EU Heavy Industry

[fneum]

Suggested change

	MWh_elec_per_tHVC_mechanical_recycling: 3. # estimate
	MWh_elec_per_tHVC_mechanical_recycling: 0.547 # from SI of https://doi.org/10.1016/j.resconrec.2020.105010, Table S5, for HDPE, PP, PS, PET. LDPE would be 0.756.

Adds the energy use of mechanical recycling.

I have assumed energy demand for recycling of HPTE, PET, PS, PP as they are all very similar whereas LDPE is higher as well as more difficult to recycle.

https://www.generalkinematics.com/blog/different-types-plastics-recycled/

[nworbmot]

Sounds good! Also for reference, another source doi:10.1016/j.scitotenv.2017.05.278 with somewhat lower energy demand around 0.3 MWh/tPlastic. We can take the newer higher number.

[nworbmot]

Could we also check that the numbers from https://doi.org/10.1016/j.resconrec.2020.105010 are consistent with our chemical recycling assumptions? @brynpickering warned about reliability of Material Economics numbers in a private email.

[fneum]

The paper describes for routes of chemical recycling:

1. refinery feedstock: liquefy plastic -> use as a substitute for oil or naphtha in refinery and steam cracking processes
2. fuel production: gasification / pyrolysis -> gaseous / liquid fuels
3. monomer production: polyolefins -> monomers (e.g. ethylene) through thermal/catalytic pyrolysis
4. chemical upcycling

Route 3) comes closes to the Material Economics study (which uses pyrolysis and electric steam cracking and methanol synthesis to produce HVC from HVC for which they need 6.9 kWh/kg electricity with a yield of 0.91 kg/kg).

Route 3) in the paper following table S10 in the SI requires 3 MJ/kg (0.78 kWh/kg) of heat to turn HDPE into ethylene with a yield of 0.85 kg_ethylene / kg.

Not sure how conclusive this comparison is. Seems Material Economics numbers are a bit high?

*polyolefins: LDPE, HDPE, PP
*HVC: ethylene, propylene, polyolefins etc.

[fneum]

Resulting change in overall energy demands (2050):

	old	new
electricity	1697	1721
coal	0	0
coke	0	0
solid biomass	702	702
methane	380	381
hydrogen	204	195
low-temperature heat	59	59
naphtha	778	728
process emission	127	127
process emission from feedstock	27	25
current electricity	1040	1040

Little more electricity, little less hydrogen, less naphtha.

[fneum]

For HVC_primary_fraction, HVC_mechanical_recycling_fraction and HVC_chemical_recycling_fraction you can now also set a pathway, e.g.:

HVC_primary_fraction:
    2020: 1.0
    2030: 0.8
    2040: 0.5
    2050: 0.0