Publication - Research and analysis

Deep decarbonisation pathways for Scottish industries: research report

Published: 21 Dec 2020

The following report is a research piece outlining the potential pathways for decarbonisation of Scottish Industries. Two main pathways are considered, hydrogen and electrification, with both resulting in similar costs and levels of carbon reduction.

Deep decarbonisation pathways for Scottish industries: research report






5. Decarbonisation and Energy Efficiency Roadmaps Scottish Assessment.pdf






11. Scope 2 emissions associated with purchased electricity or steam which is generated off-site by a third-partyand scope 3 emissions to other parts of an industry's supply chain are beyond scope. For further detail on the differences between scope 1, 2, and 3 emissions see

12. A breakdown of all Scottish emissions (in and out of scope) is provided in Appendix 8.3. Emissions from all Scottish industries defined by the 'industry' CCP mapping. The original dataset can be accessed at

13. This includes combustion of all fossil fuels as well as process emissions, but not emissions from the combustion of bioenergy sources (e.g. biomass or biogas) which are known as 'biogenic' and are considered carbon neutral for accounting purposes.

14. Ricardo (2020).

15. The Kinneil Terminal within the Forties Pipeline System is the only site within this subsector which was kept within scope because of its tight connection to other sites within the Grangemouth complex.

16. Of these, 60% relates to nitrogen oxides (N2O) and 40% to methane (CH4).

17. Specifically, the 7 largest sites are: the Grangemouth refinery and chemical plant, the Fife ethylene plant, the Dunbar cement plant, the Kinneil Terminal within the Forties Pipeline System, and two large combined heat and power plants in Grangemouth.

18. Sums may not add due to rounding.

19. Petrochemicals mostly refers to olefins / ethylene.

20. Within the analysis, emissions from each CHP plant are re-allocated to industrial users of the heat and power which they produce.

21. The 'Other non-metallic mineral products' subsector exists both within the chemicals sector, where it refers to a manufacturer of flame retardant construction materials, and within the 'other EIIs' sector, where it refers to various asphalt producers and a brick manufacturer.

22. This includes 127 SWA member sites (>95% of subsector emissions) and 2 breweries.

23. I.e. all of the sites indicated in footnote 17 except for the cement plant.

24. Note that some of the sites on the map are partly or fully hidden. For instance, the Grangemouth chemical plant (orange disc) partly hides the Grangemouth refinery (yellow disc), and completely hides the Kinneil gas terminal. Also, SWA member sites not included in the NAEI dataset are not mapped here.

25. This includes combustion of all fossil fuels as well as process emissions, but not emissions from the combustion of bioenergy sources (e.g. biomass or biogas) which are known as 'biogenic' and are considered carbon neutral for accounting purposes. It is also noted that some of the site-level data is estimated, rather than reported by site operators.

26. This data was published by Ricardo (2020) and was analysed as explained in Appendix 8.3.



29. A full bibliography of the sources reviewed for this task is provided in Appendix 8.1.

30. Including both blue and green hydrogen. See note on terminology on page 12 for definitions.

31. With the exception of the cement industry, where both are deployed in both pathways for reasons discussed in Section 4.2.1.

32. A list of the stakeholders who were happy to be named is shown in the Acknowledgments.

33. Sites directly interviewed were responsible for about 80% of all CO2 emissions in 2018, and an additional 15% of the emissions arose from sites conducting activities similar to those of interviewed stakeholders.

34. Committee on Climate Change (2019).

35. This statement applies to the UK as a whole, rather than to Scotland alone. However, the closure of the Ravenscraig steelworks in 1992 represents a clear example of how this applies to Scotland. Source: Committee on Climate Change (2018).

36. Owen et al. (2020). 'Territorial emissions' exclusively refers to emissions arising from activities based in the UK, whereas consumption-based emissions account for all emissions embedded in goods purchased in the UK.

37. This accounting standard arises from the assumption that all emissions relating to bioenergy use are offset by absorption of atmospheric CO2 during plant growth through photosynthesis.

38. The term 'purchased' is used to differentiate these fuels from the 'internal fuels' discussed below.

39. Mostly coal but also other solid fuels, like the waste-derived fuels burnt in the cement kiln.

40. This is for instance the case in the cement kiln. While gas-fired kilns also exist, these are generally considered less efficient.

41. This is the core process for olefins (e.g. ethylene) production and is found at thepetrochemical plants in Grangemouth and Fife.

42. For instance, fuel gases and ethylene are produced at a fixed ratio for a given feedstock (e.g. naphtha). Slight differences in this ratio may arise due to changes in feedstock, and increased use of ethane can for instance slightly reduce the proportion of internal fuels that are co-produced with ethylene. When this happens, more natural gas must be purchased to offset the reduced internal fuel production, and this could in turn be fuel-switched to hydrogen.

43. Improvements in energy efficiency will however reduce the total amount of energy required, which can reduce the use of any purchased fossil fuels that is co-fired with the internal fuels (e.g. natural gas in the case of steam cracking, as discussed in the previous footnote). It is also noted that fuel switching could be part of the solution to reduce emissions from internal fuels, but only when it is combined with CCUS as discussed in Box 4.

44. Note that this is grey hydrogen, since CCUS is not currently installed.

45. Options to reduce flaring emissionsare assessed in Element Energy (2019a).

46. For further detail see Brownsort (2018).

47. The main processes include steam cracking for olefins production and various furnaces at the refinery.

48. The 'energy carrier' carries the heat from where fuels are combusted (e.g. the boiler) to one or multiple processes requiring heat.

49. The percentage is expressed on an energy-content basis (i.e. based on the MWh of the various energy sources). Source: analysis of energy consumption statistics by BEIS,

50. Stakeholder interviews suggested that the share of energy use linked to electricity is around 5-15% in these sectors.

51. See Appendix 8.1 for the complete Bibliography and Section 6.4 for a summary of the feedback received from the interviewed stakeholders. It is noted that this study does not make use of confidential site data.

52. The waste-heat recovery options considered here refer to on-site use of the recovered heat. However, it is noted that it is also possible to use the recovered industrial heat off-site, e.g. in district heating.

53. It is noted that, the EU-defined list of Best Available Technologies may not apply after Brexit, and that only BATs relevant to increasing energy efficiency are considered here.

54. The key sources for this analysis are: (1) Benner et al (2011); (2) ICF & Fraunhofer ISI (2019); (3) IEA, ICCA, & DECHEMA (2013); (4) Griffin et al (2018); (5) WSP & DNV GL (2015); (6) Mineral Products Association (2013); (7) ETC (2018); (8) Brownsort (2018); (9) British Glass (2014); (10) British Glass (2017). The full references as well as a complete list of all other sources consulted for this task are included in Appendix 8.1.

55. Major overhauls were only specifically listed for sectors where stakeholders indicated that their implementation would be not be considered an 'incremental' improvement.

56. See

57. Element Energy, Ecofys, & Imperial College London (2014).

58. For modelling purposes, it is further assumed that investments in efficiency measures can achieve a 5-year payback, on average, also thanks to the assumed incentives (not modelled explicitly). Note that this does not imply that commercial viability could be attained without the incentives.

59. See note on terminology on page 12 for definitions of green and blue hydrogen.

60. The presence of biomass in waste-derived fuels was reported by the Dunbar cement plant operator at

61. The characterisation of the mixed-fuel kiln was informed by stakeholder consultation. The corresponding cost was calculated as the weighted average cost of the individual technologies (i.e. the kiln co-firing biomass, hydrogen furnace, and electric plasma gas, see Table 4 for individual costs), with weights corresponding to the percentages quoted above.

62. Specifically, it is assumed that 45% of the energy output current derived from CHP plants is provided by electric boilers, 15% by heat pumps, and 40% by the grid. Heat pumps are mostly used in the context of waste-heat recovery (rather than from ground or air source).

63. Based on network cost calculations from Ricardo (2019), including civil and installation costs.

64. See Element Energy, Advisian, & Cardiff University (2019).

65. For examples of anaerobic digestors used in combination with biogas CHP plants at Whisky distilleries see Scotch Whisky Association (2012).

66. Ricardo (2018).

67. Committee on Climate Change (2018).

68. The 2-year delay is an estimate of the time it would take a newly demonstrated technology to become more broadly available commercially. It is noted that Scottish sites could potentially deploy these technologies sooner if strong incentives existed, and indeed they could be among those sites who pilot new technologies even before they have reached TRL 9.

69. Thus, for example, a technology that is at TRL 5 in 2020 could be expected to reach TRL 9 in 2028.

70. The possibility of retrofitting certain hydrogen technologies also helps with their uptake (see corresponding feedback from industry stakeholders in Section 6.4 and uptake assumptions in Section 5.3).

71. See

72. i.e. from 2028, see Section 4.2.3.

73. Sources: 1) Element Energy & Jacobs (2018); 2) Ricardo Energy & Environment (2019); 3) Parsons Brinckerhoff (2011).

74. The characteristics of electric and hydrogen ovens were assumed to be comparable to those of process heaters (except for their lifetime, assumed to be equal to that of natural gas ovens).

75. The later date refers to furnaces for glass smelting.

76. Natural gas and hydrogen CHP costs based on the "medium" and "high" estimates for "NOAK Small GT based CHP", respectively.

77. See page 28 for a description of the mixed-fuel kiln.

78. Since neither the use of electricity nor hydrogen combustion release greenhouse gases.

79. Only those emissions associated with the generation of the additional electricity demanded by the electrification of industrial processes are considered here. Scope 2 emission related to the baseline level of electricity demand are instead not considered.

80. 2017 data and 2035 projection from BEIS updated energy and emissions projections (2018).


82. These assumptions are broadly in line with low-end estimates from van Cappellen et al. (2018) and Mohd et al. (2019).

83. A discussion on these embedded emissions related to green hydrogen is provided by E4Tech (2019).

84. The Health and Safety Laboratory (2015) reports that "the differences in the behaviour of methane mixed with up to 20% hydrogen and that of pure methane are small and unlikely to present a significantly greater hazard in practical situations" and that gas appliances are "capable of operating safely (at least in the short term) with a hydrogen content of ≤ 20%", although some of the long-term impacts are not well understood, including the effect of burner heating due to higher flame speeds. Likewise, Ofgem's summary of the HyDeploy project states that "All appliance sold post 1993 must comply with the 1990 Gas Appliance Directive 90/396/CCE (GAD), which demonstrates that they can operate on a wider range in gas quality than specified in the GS(M)R".

85. Produced via the reforming of natural gas in combination with CCUS, which assumes that CCUS is deployed as indicated in Section 5 and that one or more green hydrogen projects are also commissioned by then.

86. Note that the production cost does not include transportation and distribution costs, billing costs, taxes and levies.

87. Hydrogen Council (2020).

88. Speirs et al. (2017).

89. Element Energy (2020). Project configurations assessed include hydrogen production from a high voltage grid connection, from direct connection to a wind farm, from non-exclusive connection to a wind farm (i.e. some electricity is fed to the grid), and from offshore electrolysis. Factors like the electricity price, levies, and connection costs vary depending on the configuration.

90. Mid-range values by Bloomberg NEF (2020) calculated with a levelised cost of electricity of $28/MWh in 2030 and $17/MWh in 2050. For the best locations (e.g. Australia) they instead estimate a price of 3.2 p/kWh in 2030 and 1.8 p/kWh in 2050.

91. Electricity price for 'Large manufacturing sites (low-carbon support compensation)', gas price for 'Large (ETS) manufacturing'. Source: CCC (2017).

92. To which one must add the costs and energy losses related to conversion from gas to hydrogen, as well as CO2 compression, transport, and storage. Note that carbon capture is expected to cost relatively little in the case of blue hydrogen due to the high CO2 concentration in the flue gases.

93. A reader interested in assessing the impact of different energy prices can simply scale the results on the additional cost of energy (see Table 15) by the updated difference in energy prices to obtain an approximate result.

94. See Appendix 0 for the carbon cost assumptions based on projections by BEIS.


96. This may also include deployment of electrification options instead of hydrogen fuel-switching technologies, or vice-versa.

97. Note that the replacement cost for fossil-fuelled appliances is also incurred in the BAU scenario.

98. Ricardo (2020).

99. As was discussed in the previous chapter, process emissions and those from the combustion of internal fuels cannot be abated by fuel switching and are therefore prime candidates for CCUS.

100. A thorough review of all capture technologies can be found in Element Energy, Carbon Counts, PSE, Imperial College, & University of Sheffield (2014).

101. A capture rate of 100% is assumed for the refinery SMR and for the cement process emissions since they both result in high CO2 purity streams.

102. Levelised values were calculated using a social discount rate of 3.5%.

103. Cost estimates based on the assumptions reported in Appendix 8.7, an assumed energy price of 4.9 p/kWh and 6.4 p/kWh for hydrogen and electricity, respectively, and a project lifetime of 20 years. A substantially lower cost of capture applies to high CO2 purity streams (see footnote 101).

104. Note that the second option does not apply to petcoke combustion in the refinery's fluid catalytic cracker (FCC).

105. Pale Blue Dot, Axis Well Technology, & Costain (2016). Excludes capture, compression, and onshore transport.

106. As an alternative, CO2 could be shipped from Grangemouth directly to the Acorn Storage site.

107. Exceptions the refinery SMR and for the cement industry are discussed in footnote 101.

108. IEAGHG (2019).

109. Element Energy (2019b).

110. This process converts carbonaceous materials into hydrogen (and oxygen, carbon monoxide, and carbon dioxide) at high temperatures (>700°C) and without combustion. A complete description of the process can be found at

111. Speirs et al. (2017).

112. Committee on Climate Change (2018).

113. Element Energy (2017b).


115. COVID-19: UK Economic Update on 29 April.

116. ONS Manufacturing sector performance, UK: 2008 to 2018.

117. Certain economists predict "long-lasting negative effects on unemployment […] because the lockdown disproportionately disrupts the employment of workers who need years to find stable jobs" in the USA ( Similar conclusions were drawn by a study looking at the Chinese economy (

118. It is however important to reiterate once more that other reasons (e.g. economic or logistics) may make it strongly preferable to choose hydrogen over electrification (or vice versa) for certain processes, as discussed in the next chapter in the context of a possible hybrid pathway.

119. See Section 3.2.

120. Although it was noted above that partial reconversion of the gas infrastructure is thought to be possible, this is not considered to be a likely option for initiating fuel switching at large industrial sites, also due to the difficulties in managing varying levels of hydrogen in the gas blend highlighted by Navigant (2020). Instead, it is expected that new hydrogen-only infrastructure would need to be developed.

121. ACT Acorn Consortium (2019). The import capacity of the Feeder 10 pipeline is of up to 3 MtCO2 per year. It is noted that CO2 shipping could potentially offer a faster route to the completion of the Grangemouth CCUS project, should the availability of the Feeder 10 pipeline be a bottleneck.

122. Note that this SMR is used to produce hydrogen for internal refining processes but it is not assumed that the same reformer will become the blue hydrogen production hub to serve local industries. Rather, a new reformer is assumed to be developed.

123. The study does not make assumptions around who would develop the blue hydrogen production facilities. It is however noted that the Grangemouth refinery already produces (grey) hydrogen within its SMR.


125. Emissions embedded in the low-carbon energy used for CCUS operations are netted off from its abatement potential.

126. To capture flaring emissions, it would be necessary to invest in equipment that mostly sits idle, and which may be utilised even less if the reliability issues are resolved. This is likely to imply a very high abatement cost for CCUS on flaring-related emissions.

127. British Glass (2017).

128. See

129. Due to differences in product quality.

130. See Element Energy, Advisian, & Cardiff University (2019).

131. The fact that the steps appear smoother in the Electrification pathway is due to the fact that the carbon intensity of the electricity grid is assumed to significantly reduce over time, whereas that of hydrogen only reduces marginally (see Figure 7). This in turn causes a steady reduction in the emissions trajectory of the Electrification pathway even in years where no fuel switching deployments occur.

132. The key fuel-switching deployments abate approximately 1.0 MtCO2e and the four CCUS deployments abate 2.6 MtCO2e. Combined, this represents just under two thirds of the total abatement achieved by 2045 in both the Electrification and Hydrogen pathways (5.6 and 5.5 MtCO2e, respectively).

133. This is when the Acorn project expects to be able to receive up to 3 MtCO2/year from Grangemouth through Feeder 10.

134. See 4.2.2 for assumptions on the share of blue and green hydrogen.

135. See footnote 84.

136. Sums may not add due to rounding.

137. Includes the steam crackers.

138. Assumed to be in Grangemouth, only in the Hydrogen pathway.

139. Based on an estimated 0.25 MtCO2e of emitted for each TWh of blue hydrogen produced. Source: Mohd et al. (2019).

140. Energy demand calculated from site emissions and assuming average fuel emission factors. Note that energy demand slightly reduces over time due to efficiency improvements and that it is lower for the Electrification pathway, due to the higher efficiency of electrical appliances.

141. Fuel switching is assumed to be rolled out progressively across the Grangemouth sites.

142. The cement industry could begin fuel switching to 70% biomass sooner with the right support in place. Here it is conservatively assumed that switching to the mixed-fuel kiln occurs all at once.

143. Curly braces indicate assets not quantitatively assessed in this study.

144. The possibility of combining natural gas with biogas and/or biomass gasification is discussed in Box 5. The quantitative analysis presented here assumes that only natural gas is used.

145. CO2 could potentially be utilised instead of stored; this does not affect the quantitative analysis presented below.

146. It is expected that nearly all of the electricity generated in 2045 will be from renewable sources, which is why non-renewable generation sources were omitted from the table.

147. The electricity demand calculated here is additional to that for renewable electricity drawn from the grid already today.

148. Energy demand related to CCUS at the blue hydrogen production facility is not included here as its contribution is accounted for in the context of the primary energy discussion when estimating the natural gas demand for CCUS (see next sub-section).

149. The mixed-fuel kiln assumed to be used by the cement industry (see page 28) is responsible for all biomass use, all electricity for industry in the Hydrogen pathway, and all hydrogen for industry in the Electrification pathway.

150. Load factor refers to Scottish average for wind power. Source: BEIS (2020) Table 6.1c Renewable electricity capacity and generation: Scotland.

151. Accounts for 8% transmission and distribution losses. Source: Written evidence submitted by Citizens Advice (NTC0019),

152. Assumes that the electrolysers are directly connect to a dedicated wind farm, hence the load factors coincide.

153. Although electrolysers can achieve higher efficiencies (closer to 70%), the value assumed here is considered more representative in light of the fact that electrolysers powered by dedicated renewable generation will not always be operating at peak performance.

154. Representative of an average hydrogen production efficiency according to Antonini et al. (2020).

155. According to BEIS (2020), Scotland had 11.8 GW of renewable generation capacity in 2019, of which 9.3 GW was wind power.


157. University of Edinburgh (2009).

158. Assumption on grid upgrade requirements discussed on page 29.

159. Speirs et al. (2017).

160. Volumetric energy densities: 12.7 MJ/m3 for hydrogen; 40.0 MJ/m3 for gas. Source: Bossel and Baldur (2003).

161. Other costs which might be incurred in the decarbonisation pathways include one-off costs related to the disruption to site operations when switching fuels or installing CCUS as well as potentially increased operational complexity once the systems are operating. These hard-to-estimate costs are assumed to be negligible compared to the cost factors quantified in this study.

162. Key financing assumptions: assets financed over their entire lifetime, at a 10% weighted-average cost of capital (WACC). Methodology reported in Appendix 8.9.

163. Scottish Annual Business Statistics (2017). Note that the scope of this study largely (but not fully) overlaps with the 'manufacturing industry' definition used in the referenced statistics.

164. Using a 3.5% social discount rate and referred to 2018 pounds.

165. Note that this excludes any financing requirements related to investment in the infrastructure discussed in Section 6.2.3.

166. HM Treasury: Green Book supplementary guidance: optimism bias.

167. The term 'bankable' implies that a bank would be willing to offer debt financing against such a revenue stream. In other terms, the revenue stream is considered highly reliable and the risk of default on debt repayment is considered low.

168. It was conservatively assumed here that complete appliance replacement would be necessary, as discussed in Section 4.2.1.

169. This is due to the high cost and low utilisation of sizing heat pumps for peak demand, considering that 80% of the energy needs could already be met via heat pumps that are only 1/3rd of the size. Source: Ricardo (2020).

170. Also known as Border Adjustments or Border Tax Assessments.

171. Rosenbloom et al. (2020).

172. Mazzucato (2013) discusses the conditions under which direct equity investments might be preferable.


174. Element Energy (2018).

175. The emitter is paid (or refunded) the difference between a CO2 strike price contractually agreed (in £/tCO2 abated, fixed for the duration of the contract), and the prevailing CO2 market certificate price (or carbon tax). The quantity of CO2 abated is determined relative to an industry benchmark.

176. 'Properly incurred' refers to costs that are consistent with, and were negotiated freely in, the market.

177. Except for the cement kiln, which is retrofitted.

178. The illustrative example of ethylene (used in soda bottles) is extracted from a recent report by the Energy Transitions Commission (2018). Both price increases correspond to the added cost of decarbonisation.


180. Just Transition Commission Interim Report (2020).

181. The Grangemouth refinery currently produces (grey) hydrogen with an SMR, though without capturing the related carbon emissions. Carbon capture is already implemented at the Kinneil gas terminal (and likely at other gas terminals), where CO2 is separated from other feedstock gases before being released to the atmosphere. (Note that even though this is technically CO2 separation and not capture, the two share the same core technology.)

182. Source: Scottish greenhouse gas inventory 2018.

183. Sums may not add due to rounding.

184. Lenaghan, M., & Mill, D. (2015). Industrial Decarbonisation and Energy Efficiency Roadmaps: Scottish Assessment.

185. SLR (2020).

186. The global warming potential of each GHG relative to that of CO2 is provided at

187. Information about the different GHGs can be found at

188. Adapted from Element Energy, Jacobs. (2018). Industrial Fuel Switching Market Engagement Study.

189. See

190. See e.g.

191. I.e. the difference in total fuel cost comparing after and before the measure is implemented. This also factors in differences in energy prices.

192. Costs and abatement were calculated up to and including 2070 to ensure the analysis would capture the entire lifetime of all equipment deployed in 2045.