UOP LLC

25 E. Algonquin Rd.
Des Plaines, IL 60017-5017
Tel: 847.391.2000
Fax: 847.391.2253
www.uop.com

California Air Resources Board

Sacremento, CA
November 16, 2018

Dear Anil:

Honeywell UOP would like to formally submit the following response to the
information presented during the most recent Co-processing Work Group meeting held at
CARB on October 19, 2018:

The presentation and work cited by REG assessed the measurement accuracy of
carbon 14 quantification (14C) as described in ASTM D-6866 using known blends of
biogenic (renewable diesel derived from vegetable oils) and non-biogenic (fossil)
containing diesel components. This study unfortunately did not address the primary goal
of this work group and that is to identify methods that can be used to quantify the biogenic
components in fuels generated from co-processing operations in a refinery process unit
(such as from a FCC unit). This study also failed to present results on the measurement
of 14C in products / fuels from co-processing operations from a FCC unit, products
containing volatile components (as present in gasoline or lighter gas streams, which per
Bucholz is more difficult to measure), or fuels / products containing less than 0.45 wt%
pmC (percent modern carbon). Although this study aimed to establish 14C as an accurate
and credible technique for measuring percent modern carbon above detection limits, it
did not address the issues that are relevant to the real practice of co-processing bio-
containing feedstocks in refinery processes (specifically a FCC unit) at levels currently
contemplated by industry. It is expected that many, if not all the products out of the FCC
unit will contain less than 0.45 wt% pmC which raises doubts on the using this method for
measuring pmC in this situation.

It should also be noted that the REG presenter cited that REG currently does not
employ use of 14C measurements in their operations to validate biogenic carbon contents
of the fuels that they produce and/or sell to industry.

It was stated by both Bucholz and Rocke in their presentations during the Work
Group meeting that Atomic Mass Spectrometry (AMS) Limits of Detection (LOD) or Limits
of Quantification (LOQ) are important for differentiating sample 14C from processing
background 14C that can be present in the lab. For any stream containing percent
modern carbon below this limit of detection, it can be statistically stated that this stream

-1-
may or may not contain modern carbon. This is particularly relevant to the commercial
practice of co-processing of biocrude in a FCC unit where addition rates of biocrude are
planned less than 5 wt% absolute (less than 2.5 wt% percent modern carbon) in the feed
to the unit. One of the main concerns that exists with use of 14C measurements, that is
specific to co-processing in a FCC unit, is that it doesn’t provide the level of statistical
precision at the pmC (percent modern carbon) that is anticipated to be present in the FCC
unit products. Unlike a hydrotreater unit which typically will have 2-3 carbon containing
product streams (gas, naphtha, diesel), the FCC unit can have up to 8-10 carbon
containing product streams (fuel gas, C3, C4, LCN, HCN, LCO, HCO, Bottoms, Coke) all
which can contain modern carbon. This presents a challenge when attempting to account
and balance to 100% of the modern (biogenic) carbon charged in the process in the
products. Dr. Rocke referenced the LOD of pmC > 0.4406 wt% as “the minimum
detectable value…that should always generate a measurement above the critical level”
for applicability of this measurement method for determination of modern carbon. This
suggests that any measurement less than this limit of detection likely cannot be reliably
used in the establishment of a modern carbon balance around the unit. The limit of
detection of pmC referenced by Bucholz was stated to be lower than that shown by Dr.
Rocke, however this was in specific reference to the Lawrence Livermore National
Laboratory Atomic Mass Spectrometry (AMS) unit and was not in reference to limits of
detection of other commercial labs employing use of ASTM D6866 that would be used by
the refining industry to measure 14C in the products from their processes

Bucholz also described the complication of global nuclear testing on the amount of
14C present in the atmosphere since the 1950’s. It is well understood that the time of
synthesis of plant determines the characteristic 14C that will be present in the biomass.
If, for example, a tree that is used as the biomass source grew from 1960 until it was
harvested in 2017, the tree would have a 14C concentration from each of the growing
cycles represented by the underlying atmospheric 14C concentration levels. The
characteristic amount of 14C present can also vary depending on which portion of the
biomass is used in co-processing. One could hypothesize that for a given biomass source
that there could be many characteristic 14C’s present in the biomass which, depending
on the portion of the biomass used in co-processing, could be from contemporary or much
older periods that would have vastly different percent modern carbon levels. It should be
noted that information presented by REG looked at contemporary sources of biomass
where the variation in 14C is relatively consistent over the short growing cycles of biomass
used to generate the biogenic containing fuels. Contrast this tothe biogenic feedstocks
likely to be used in FCC co-processing operations that are derived from long growing
cycle (non-contemporary) biomass. This adds to the complexity and analysis of using
14C for the purposes of determining the biogenic content of the fuels produced from this
source of biomass.

Vegetable oils, which are being used in co-processing applications in hydrotreaters

in Europe (and believed to the feedstock sources used in the REG study) is typically
derived from plants that are well-controlled in terms of growth period and human inputs in
cultivation, in mono-culture systems (soybeans and corn) that practically have a human
defined pedigree. They have been planted and harvested in the same known year. This
very well defines the 14C atmospheric content and the 14C carbon content of the plant
materials. Note that each various plant type has different selectivities of 13C and 12C
fixing (and therefor 14C fixing selectivity from the atmosphere) so knowing the plant type
and the year in which it is grown is important to know the 14C content to back-calculate

-2-
the total percent modern carbon represented for that particular piece of biomass when
diluted into a petroleum stream. .

Contrast this to 2nd generation lignocellulosic biomass sources, which are

specifically contemplated and used in the application of FCC co-processing in which many
of these variables are less controlled. Managed forests can have a pedigree as they were
likely grown under human control. However, they are the product of decades of growth
so an averaged atmospheric 14C content needs to be used. These sources can be
mixtures of tree and plant varieties adding more uncertainty to the amount of modern
carbon represented by a certain 14C signal. Additionally, there are a lot of effects that
can influence the 14C/12C ratio of biomass. Location is one factor. Where there are
large emissions of anthropomorphic carbon from fossil fuels (cars, power plants, industry,
lots of people, i.e. cities and heavily urbanized areas, the natural 14C level in the
atmosphere is diluted with 12C carbon dioxide. Biomass from the N.E. of the USA would
have a 14C level lower than biomass grown in the N.W. USA and would appear older than
actual based upon global 14C/12C ratios. The northern hemisphere also has different
14C/12C ratio to southern hemisphere due to different atmospheric circulations and more
anthropomorphic carbon emissions. On the local level, environmental factors such as
drought, rainfall, fast or slow growth can all affect somewhat isotopic fractionation in
plants. Variation in the isotopic ratios in tree rings have been used to estimate climatic
conditions over time. The type of biomass also needs to be taken into consideration as
there are differences between C3 (trees & shrubs) and C4 (Grasses, Sugar cane, Maize,
Sorghum) photosynthetic pathways.

An example of the present difficulties of using ASTM D6866 to measure the

amount of modern carbon present in FCC co-processed products is illustrated in Table 1
on the following page. The yields summarized in this table are taken from a FCC co-
processing consortium report completed under a Department of Energy study:

(https://www.energy.gov/sites/prod/files/2015/04/f21/thermochemical_conversion_chum
_242303.pdf, Pg 13)

If one assumes that the biogenic carbon present in the biocrude feedstock evenly
distributes across the entire spectrum of FCC products, one can estimate the pmC
concentration that is likely in each of these product streams. The data suggests that at
biocrude injection levels < 5 wt% on feed to the FCC unit, a majority if not all of the
products, with possibly the exception of gasoline, is predicted to have less than 0.4406
pmC present. Per expert opinion, the amount of pmC present in these streams would be
below the Limit of Detection for 14C measurement via AMS and therefore these streams
would be deemed to not contain modern carbon. Even at an injection rate of 5 wt% of
biocrude on feed to the unit the total amount of modern carbon that could be measured
is only 33% of the total percent modern carbon fed to the unit. This highlights the difficulty
of using 14C at low injection levels in co-processing operations. It isn’t because modern
carbon is not present in any of these product streams, but rather the amount present is
statistically below the limits of detection for the ASTM D-6866 method.

-3-
 Table 1: Co-Processing FCC Yields and Estimated pmC in Product Stream

Biooil Biooil
Base Substitution Substitution
Feedstock Case 2.5 wt% 5 wt%
VGO wt-% total feed to unit 100 97.5 95
VGO wt% Carbon in VGO 88 88 88
VGO Percent Modern Carbon (pmC) 0 0 0

Biooil wt-% total feed to unit 0 2.5 5

Biooil wt% Carbon in Biooil 50 50 50
Biooil Percent Modern Carbon (pmC) 100 100 100
pmC in Total Feed wt-% total feed to unit 0 1.25 2.5

Product Yields Measured Interpolated Measured

Dry Gas wt%FF 3.5 3.15 2.8
LPG (C3-C4) wt%FF 13.8 13.8 13.8
Gasoline (C5-220°C) wt%FF 39.9 40.25 40.6
Diesel (220-344°C) wt%FF 20.3 19.95 19.6
Bottoms (+ 344°C) wt%FF 16.1 15.25 14.4
Coke wt%FF 6.4 6.2 6
CO wt%FF 0 0.5 1
CO2 wt%FF 0 0.2 0.4
Water wt%FF 0 0.7 1.4
Total 100 100 100

pmC in Products (estimate)

Dry Gas pmC in stream 0.00 0.03 0.06
LPG (C3-C4) pmC in stream 0.00 0.14 0.28
Gasoline (C5-220°C) pmC in stream 0.00 0.41 0.82
Diesel (220-344°C) pmC in stream 0.00 0.20 0.40
Bottoms (+ 344°C) pmC in stream 0.00 0.15 0.29
Coke pmC in stream 0.00 0.06 0.12
CO pmC in stream 0.00 0.21 0.43 *assumes 100% of product C is modern
CO2 pmC in stream 0.00 0.05 0.11 *assumes 100% of product C is modern
Water pmC in stream 0 0 0
Total 0 1.25 2.5

Percent Modern Carbon Calculated in FCC Products 0 0 0.82

Total Modern Carbon Mass Balance Closure 0% 33%

The use of 14C measurements may be possible in co-processing operations in a

FCC when the total amount of biogenic carbon being fed is large enough to ensure that
the products contain sufficient and statistically measurable pmC greater than the limit of
detection for AMS. It is possible that this threshold isn’t likely achieved until more than 10
wt% of the feedstock charged to the FCC unit is of biogenic nature.

Finally the relevant publications (reference #3) in the REG presentation (slide 13)
of G. Fogassy, N. Thegarid, Y. Schuurman and C. Mirodatos The fate of bio-carbon in
FCC co-processing products. Green Chem., 14 (2012) 1367-1371 looks to draw
conclusions that the addition of biocrude to FCC results in specific biocrude carbons
preferentially going to gas and coke. This paper concludes that this could be modified
with changes in catalyst, injection method, feed properties and operating conditions.
These changes not only effect the modern carbon deposition between cuts but also the
petroleum carbon deposition themselves which must be predominantly calculated in
specific yields to the each individual products. This article also goes on to state a case for
use of a mass balance approach in co-processing applications, "This also questions the
use of the ASTM D6866 protocol applied to fuels to verify the amount of renewables
processed in a (bio) refinery. Rather the overall process needs to be taken into account."

-4-
 Dr. Jim Rekoske, who was a previous CARB Work Shop participant from
Honeywell UOP, presented on the use of a statistically-based mass/carbon yield mass
balance methodology for the purposes of establishing yields from the co-processing
biocrude in a FCC unit at levels at or below 10 wt% on feed. The methodology outlined
by Rekoske is fully compatible with existing refinery operations and processes and can
be readily integrated into normal operating procedures within the refinery. Honeywell
UOP continues to support the use of this method at the levels anticipated to be practiced
by industry for the reasons previously provided to CARB.

Honeywell UOP is willing to bring experts in the field of statistical mass balance
methods to CARB to review and discuss in more detail the methodology originally
presented and proposed by Rekoske to aid in the guidance document currently being
considered by CARB staff. We believe that the statistically-based mass balance method
is an accurate, reliable and practical approach for assessing the biogenic contribution to
products out of an FCC when co-processing second generation (cellulosic-derived)
biomass feedstocks.

Honeywell UOP has been in the business of supporting the refining and
petrochemical process industry for over 100 years. Honeywell UOP’s business model
was built and continues to operate on the foundation of mass balances in the
development, engineering, and operation of chemical-based processes in the refining and
petrochemical industries. Mass balances are the industry standard means for making
critical decisions to the operations and profitability of these chemical-centered facilities.
If there was an alternate method available for assessing performance, Honeywell UOP’s
business model would be utilizing this. It is still believed that for this specific application
(co-processing of biomass feedstocks from second generation sources in a FCC unit at
the levels indicated), that the mass balance approach is best served for assessing the
contribution of these feedstocks on the production of renewable fuel components.

Very truly yours,

Chad Huovie
Director, FCC Co-processing
UOP LLC

-5-