25 January 2023

An article by Dr. Michael Has, dist. Prof.
The enclosed is an edited transcript of a talk given at the RadTech Conference on Circular Economy held in December 2022.


The following talk is somewhat off track in this conference about circular economy - this is because our todays topic would aim at re-assessing previously used materials to reduce the over-all footprint of a product and saving energy. What I am concerned with today are footprints and, in the following, especially the CO2 footprint of coatings and inks in printing. The CO2-footprint is primarily about recording energy consumption measured in emissions of greenhouse gases. The two fields are closely linked, however, because the production of materials naturally involves energy consumption, as does the processing of the materials used - such as in recycling, refurbishment or comparable measures. Both require energy, which is why the energy and material footprints are closely linked.

To get into the topic, we first need to clarify how footprints are calculated and how to deal with errors/uncertainties and estimate them. We will give examples and use the equation system to calculate footprints to discuss the contribution of coatings - a contribution that, I think, is mostly underestimated as intuition indicates that something so significantly thinner than the underlying material should not dominate net effects.

Footprint calculation

A number of footprints are distinguished in the literature - they all serve as KPIs to quantitatively describe the path to achieving sustainability development goals. Footprints exist to characterize human rights during production, water use, biodiversity, land use, and much more - the major ones of these sustainability goals are summarized in the Sustainable Development Goals, or SDGs. All of these goals characterize how a society manages its resources.

For the discussion pursued here, energy and material consumption in production, use, and disposal or recycling of raw materials and products are critical. The energy-related footprint is measured in kilograms of CO2 per product or per kilogram of product, and the materials consumed are often scaled by the location or individual products considered. This footprint is measured in a standardized way - reference are two ISO standards: ISO 14040/44 and 14064, which in turn uses the questionnaire offered in Greenhouse Gas Protocol. The carbon footprint is usually divided into three different contributions – called scopes,

Scope 1 is what a company emits through its own activities,

Scope 2 refers to incoming energy regardless of the form of delivery,

Scope 3 generally refers to everything that enters production from outside and, in whatever form, leaves the company.

A first look at the questions promises a simple answer - an impression that does not prove true on closer inspection. Answering the questions to be asked turns out to be complex; it usually requires the well-managed deployment of a larger team of experts. If, as in the case discussed here, a tangible product is involved, rule of thumb is that around 60 to 80 % of all emissions stem from the incoming and outgoing value chain. Describing this precisely is complicated above all because data from all suppliers must be requested to provide information on the goods they supply or receive, or even on waste.

Estimation of Errors

The underlying data are not provided with 100% precision, if at all, but are subject to errors and uncertainties: The literature distinguishes between aleatory (systemic) and epistemic (random) errors. The second source of uncertainty deals with values that are in principle accessible, but that are not precisely available. These are typically measurement errors, quantification errors, and the like missing data, perhaps conflicting information, conversion factors and their accuracy, rounding effects:

When measuring volumes, one usually has to allow for a measurement error of +/- 2% - 6%. A comparably low figure, but one that becomes relevant when, for example, one's own natural gas bill is the subject of this error. In addition, the exact conversion factors are not announced until about one to two years after the reporting date.

Aleatory uncertainty is well discussed in Literature (1), while for epistemic phenomena published knowledge is very limited in the context of LCA.

Focus here: epistemic uncertainties

Without discussing this in great detail here (2), in the practical calculation of footprints (within the statutory period of 4 months after the end of a fiscal year) it is probably rarely possible to achieve an error of less than 5-9%. This may sound small, but it is still relevant when one considers that the Science-Based Target Initiative and the UN recommend annual reductions in greenhouse gas emissions of 2% and 4%, respectively.

Three examples of footprints - an art museum and an airline describing a domestic flight within the Philippines and part of a material footprint in label printing.

Example 1

The first example displays the annual footprint of a Munich-based Arts Museum (Lenbachhaus). It is obvious that not all of the sub-criteria of the scopes are reflected in the footprint; some are unnecessary to describe the individual case and therefore are not listed, while others are: heating, which in turn requires the use of fossil energy, cooling, inbound transportation, and several other factors.
The total emission is the sum of the individual emissions. The accuracy would be minimum +/- 6% - 5%. So it's not really accurate what the calculation ends up with.
However, the calculation illuminates where the most significant emissions come from and thus gives strong indications of the savings potentials that are addressed in the following Ecodesign.



Estimated Percentual Contribution

Applying uncertainty/ most unclear Assumption


Scope 1




Heating, fossil Energy, Cooling

31 %

+/- .6 %


Scope 2




Electricity/ remote Heat

19 %

+/- .2 %
Source of energy known


Scope 3



Purchased goods

6 %

+/- 2 % / truncation errors


Resistance of wires


+ 1.1 % / Distance to source=far


Inbound Transport


+/- 1.4 % / age of fleet



.5 %

+/-  .2 %



3 %

+/-  .1 %


Business Trips

4 %

+/-  .7 % / age of fleet



.5 %

+/- .01 % / age of fleet


Total uncertainty

100 %

+ 6.4 % / - 5.3%

Table 1: Annual footprint of a Munich-based Arts Museum (Lenbachhaus).

Example 2

Another example - an airline describing a flight in the Philippines. Here, the data presented is incomparably more limited. The airline's goal was apparently to avoid making passengers feel better and also to present itself in an ecological light: The expected restriction of matching criteria from the scopes surprisingly offers no indication of aircraft depreciation, airfield, etc. . The various options for refuelling with fuels with more or less large footprint remains undiscussed and the total emissions are reduced to the aviation fuel consumed in this case - under the hardly realistic assumption that the aircraft always flies at 100% capacity with always the same route, flight altitude ... .

Scope 3



CO2 Emissions in 2019

In kg CO2


3.1 – 3.2

Purchased goods and service

Set to 0



Means of production / capital goods

Set to 0



Fuel and energy-related emissions

126 kg CO2 per seat

Biggest Source

Ave 900 liters Jet A1 consumed round trip;
Conversion factor for 1 liter Jet A1 = 2.52 kg/CO2

2,268 kg CO2 per round trip (Manila – El Nido - Manila)

3.4 ff


Set to 0


 Table 2: Carbon footprint for a return flight Manila – El Nido – Manila for one passenger.

Example 3

The material footprint was briefly mentioned above - here, too, an example published some time ago by FINAT (3), the European organization of label printers. This example refers to the waste generated during the production of label material. It is impressive that if the different phases of production are considered in a structured way, the production of the liner, the production of the face material, the production of the silicone, account for 40 to 65% of all waste. In the actual production of the label material, of course, energy and water are added. In the subsequent printing process, 20 to 40% of all waste is generated, and in the application of the labels and their use, only 3% each. This is not really a classic material balance, because it does not include which materials are meant exactly and which path the materials (can) take, for example, in a reuse - nevertheless, this representation offers good indications of the potential of the method to identify possible weak points in production methods and paths.


Table 3: Material Footprint – relative contribution of waste generated during the production of label material throughout the production and use of labels.

Coated Products

Coated products are not methodologically special as far as the calculation of footprints is concerned: The question list of the Greenhouse Gas Protocol or ISO is processed and the individual points are evaluated numerically. However, it is worth noting that even in this context it is not always advisable to rely on intuition:

There are many types of coatings - paints, varnishes, lamination, effect and functional coatings are often applied in one or more layers to a surface of a substrate. In most cases, the carrier material is thicker and, not only because of its thickness, also the material that is more mechanically resilient. Intuitively, therefore, it is initially obvious to assume that the energy footprint in the production of the coated substance is greater than that of the applied layer. The assumption is perhaps justified in individual cases, but not as a rule, as the following example shows.

In order to describe the general conditions a little more theoretically, a more general representation in formulas:

(1) CFproduct = ∑ CFsub-scopes

(2) CFproduct  = CFcoat + CFsubstrate + ∑ CFRest            
Emphasize coating and substrate
whereas the coating ma, of course consist of more than one layer – e.g. for the case of additive layers of ink (e.g. in cmyk) plus varnishing and plus lamination in (2) CFcoat would have to be replaced by
CFcoat =   CFink -c + CFink -m + CFink -y + CFink -k + CFvarn + CFlam

(3) CFsubstrate = msubstrate ksubstrate  = Asubstate g ksubstrate       
A=area; g=base weight; k=conversion factor

For the case of one-color print and printing on both sides CFink would develop so that

(4) CFink = mink kink  =  2 Rink dink, 1% COV Aink kink    
printing on both sides would lead to applying a factor 2. R=spec. weight,
dink, 1% =thickness for 1%, COV=Coverage

(5) To keep it simple - for printing in one color
CFproduct  = 2 Rink cmyk dink cmyk, 1% COV Aink cmyk kink cmyk 
                  + Asubstrate g ksubstrate  + ∑ CFRest            

Equation (5) indicates that, for the case of a printed product the carbon footprint of the printed product is a linear function of the coverage of the printed product. Reports (3) indicate that CFproduct of the printed product is dominated by the substrate, and consider the ink as a constant and comparatively small. To test this assumption, it is possible to calculate a "break-even point" for a coverage COV above which the carbon footprint of the ink is larger than that of the substrate:

(6)          COV   =   Asubstrate g ksubstrate / (2 Rink dink, 1%  Aink kink)

Additional treatments such as multiple layers, lamination, varnishing, metallic and functional applications may enter the equation and have to be treated like another layer with a specific coverage. For the case of varnishing applied on one side that would lead (6) to be changed to

(7)          CFproduct  =   CFink + Rvarn dvarn Avarn kvarn + CFlam + CFsubstrate +
                            + ∑ CFRest   

Also (7) implies a direct relationship between the thickness of the varnishing as coating:

(8)            dvarn =   (CFproduct - CFink - CFlam - CFsubstrate - ∑ CFRest) /
                            / (Rvarn Avarn kvarn)

It becomes obvious that the more layers there are on the surface the less the relative impact of the substrate is. In addition, in practice, the substrate must also be transported frequently between treatments, which also complicates the equation and must be calculated at the expense of the coating.

Using the approach on errors mentioned above leads to the insight that, if the errors in all footprints are comparable and, as often the case under practical conditions, a large fraction originate in conversion factors the error in the resulting judgement of critical thicknesses d of the coating is rather smaller than the errors contributing to the CO2-Footprints of individual layers or the substrate.

To add a practical example for printing (4): A print substrate printed essentially with black text - such as a daily newspaper - is used as a starting point. The area coverage COV in this application is practically limited to somewhat less than 10%. Under these circumstances, the contribution of the ink to the carbon footprint of the printed material is about 6-10%. In the case of other printed materials, namely in color printing, the area coverage is well above 300%. The question is obvious how an increase in area coverage to that extent would affect the relative contributions of the ink or substrate. Beyond this academic consideration, however, it is crucial whether such situations are actually found in practice.

For this purpose, standard printed products on the market in the USA were collected, and their usual basis weights and area coverages were compiled.



Print Volume share in US/2017

in %

Offset share

in %

Average g

in g/m2

Break Even Coverage

in %


Brochures, Flyers, Marketing Collaterals

22 %


90 - 120

117 – 235



  8 %


80 - 150

104 - 562


(Business Card, Report, Letterhead)

  2 %


90 - 120

117 - 235


Direct Mail

12 %


60 - 80

78 - 374



  6 %


80 - 100

104 - 468


Packaging (incl. labels, flexible and folding cartons, corrugated segment)

20 %


120 - 200

156 - 562

Table 4: Break-even area coverages calculated according to equation (5).

The conversion factor for the printing ink was set at 2.8 kg CO2 per kg offset printing ink, corresponding to the proportion printed in offset. The average volume shares refer to the year 2017 - should therefore be similar today. The table contains the break-even area coverages calculated according to equation (5). It is known that area coverages can exceed 300% in practical cases. The case where the contribution of the ink to the carbon footprint can be higher than that of the substrate is therefore quite realistic.

The example, thus, provides a clear indication that it can practically happen that coatings have a higher footprint than the underlying substrates. In addition to the basis weight of the substrate, the critical variables are the conversion factors of the coating and the substrate.

It is also worth mentioning that some coatings limit the recyclability of the coated good or even make recycling impossible. In that case, the coating would be even more unfavorable in terms of its carbon footprint.


The presented approach enables the calculation of CO2-footprints of prints. Epistemic errors are discussed.

The approach presented proves that coatings can contribute significantly to the CO2-footprint of coated substances - the areal coverage of coatings (e.g. ink) has to be considered when CO2-footprints of prints are reported. In connection with market figures, the approach also proves that the most important variables to be taken into account, in addition to the area coverage, are the basis weight of the substrate and the conversion factors.


(1)  Aleatoric and epistemic uncertainty in machine learning: an introduction to concepts and methods
https://link.springer.com/artic le/10.1007/s10994-021-05946-3

(2)   Michael Has, Uncertainties and Accuracy of Footprints, TAGA Annual Conference, Omaha, March 2023

(3)  Harmonized Life Cycle Assessment – Approach for the Self-Adhesive Label Industry;: https://www.finat.com/sustainability/life-cycle-assessment, assessed: December 18.2022

(4)   Michael Has, Contribution of Printing Ink to Carbon Footprint of Print, European Coating Journal, Spring 2023