Duo Wei, Xinzhe
Shi, Peter
Sponholz, Henrik
Junge*, and Matthias
Beller*
October 19, 2022
Manganese
Promoted (Bi)carbonate Hydrogenation and Formate Dehydrogenation:
Toward a Circular Carbon and Hydrogen Economy
Abstract
We report here a feasible hydrogen storage
and release process by interconversion of readily available (bi)carbonate
and formate salts in the presence of naturally occurring α-amino
acids. These transformations are of interest for the concept of a
circular carbon economy. The use of inorganic carbonate salts for
hydrogen storage and release is also described for the first time.
Hydrogenation of these substrates proceeds with high formate yields in
the presence of specific manganese pincer catalysts and glutamic acid.
Based on this, cyclic hydrogen storage and release processes with
carbonate salts succeed with good H2 yields.
Synopsis
Using
readily available raw materials (bi)carbonate and formate salts, we
report reversible chemical H2 storage and release
catalyzed by Mn in the presence of naturally occurring α-amino
acids.
Introduction
In contrast to the traditional
fossil-based linear economy, the circular economy is a model of
production and consumption which involves reducing, reusing, and
recycling, aiming to tackle the global resource shortage. Following
such a concept, human economy activity and life quality can be
sustained and improved while minimizing consumption of fossil
resources and emission of wastes.
(1) Efficient and economic carbon valorization is a substantial
practice for the circular carbon economy.
(2) Most research efforts, today and in the past, have focused
on the fixation of gaseous CO2, where a high
concentration and pressure of CO2 is commonly required,
for instance, in carbon capture and utilization (CCU), and direct
air capture (DAC) processes,
(3−5) making it expensive to catch carbon back into the ground
at a meaningful scale.
(6) Alternatively, improved carbon utilization might be provided
with the largely available solid bicarbonate and carbonate salts,
which are easily produced from CO2/base or so-called
“carbonate factories” in nature.
(7) The latter systems are widely used in the scenarios of
construction, health and diet, agriculture and aquaculture,
household cleaning, and pollution mitigation.
(8,9) Due to their inherent advantages, carbonate and to a
lesser extent bicarbonate salts are expected to play considerable
roles in the present and future circular carbon economy.
Apart from the production of urea,
(10) cyclic and polycarbonates,
(11−13) etc., which possess the same oxidation state as (bi)carbonates,
their effective reduction is a key step in not only sustainable
production of chemicals and fuels but also further implementation of
chemical hydrogen storage and release via reversible hydrogenation
of bicarbonates.
(14−16) As clean energy carrier, H2 is attracting
increasing attention, owing to its sustainable production from
renewable resources and green combustion in fuel cells providing
only water and energy.
(17−20) To avoid transportation and handling of gaseous H2
with low volumetric storage density, solid or liquid organic H2
carriers have emerged as substitutes to realize the on-demand
reversible chemical H2 storage and release.
(21−26)
Besides the well-known CO2/formic
acid (FA) based hydrogen storage system (Figure
1, left),
(27−35) the equivalent bicarbonate/formate cycle has been
investigated (Figure
1, right).
(15,36−41) Basically, all of these works make use of less
available precious metal catalysts for the corresponding (de)hydrogenation
reactions. One of the challenges applying bicarbonates in chemical H2
storage-release cycles is the undesired formation of carbonates and CO2
under specific conditions [eqs
1–3],
which limits the theoretical capacity of such a hydrogen
storage-release system after the initial cycle.
(42)
Figure 1
Figure 1. Different
concepts of chemical hydrogen storage and release based on the
interconversion of CO2/formic acid (left) and (bi)carbonate/formate
(right).
The hydrogen contents in FA (4.35 wt %)
and formate salts (1.02–2.85 wt %) are comparable to H2
storage alloys, e.g., magnesium hydrides (1–6 wt %).
(43) Nonetheless, the inferior hydrogenation/dehydrogenation
kinetics, life cycle, and harsh operation conditions (300–500 °C) of
such alloys make them currently not appropriate for most
applications.
(43,44)
Due to the lower reactivity of the
carbonyl group in carbonates compared to CO2 or
bicarbonates, their catalytic hydrogenation is more demanding.
Consequently, the transformation of inorganic carbonate to formate
salts is rarely reported. Examples generally proceed in low yields
(<10%) and poor selectivity using glycerol,
(45) hydrosilanes,
(46) and H2
(47,48) as reducing agents. In addition, both our group
(36) and Joó et al.
(49) reported the Ru catalyzed hydrogenation of carbonate salts
in the presence of CO2. Furthermore, a
transition-metal-free method was reported for the carbonate
hydrogenation to a mixture of formate, acetate, and oxalate with H2/CO2
(30/30 bar) at 230–320 °C.
(50)
The present situation and our general
interest in hydrogen storage technologies prompted us to investigate
the interconversion of (bi)carbonate and formate salts as a general
H2 storage-release method.
Results and Discussion
Hydrogenation of Bicarbonate to
Formate
We started the bicarbonate-to-formate
transformation in water and THF (v:v = 1:1) as co-solvent, H2
(60 bar), 90 °C and 12 h (Figure
2). Apparently, in the absence of any catalyst, no formate was
formed. Among all the tested complexes, the most successful one was
identified as Mn-2 bearing a methyl group at the triazine-based
pincer ligand, leading to formate in 95% yield (TON 55,000,
Figures S1–S2).
Figure 2
Figure 2.
Hydrogenation of potassium bicarbonate to potassium formate
catalyzed by base metal complexes. Standard conditions: KHCO3
(10 mmol), catalyst (0.1 mg), H2O/THF (5/5 mL), H2
(60 bar), 90 °C, 12 h. Yield of formate is calculated by (mmol
formate)/(mmol KHCO3) × 100%.
At this point, it is worthwhile mentioning
that mainly noble metals, Fe,
(48,51−58) Co,
(59) and Ni,
(60,61) catalysts have been reported in bicarbonate
hydrogenation. Manganese complexes have been rarely used in this
area, with limited examples in CO2 hydrogenation
(62−67) and FA dehydrogenation,
(68−71) despite its nature of abundance, nontoxicity,
biocompatibility, and environmental friendliness.
(72−77) In addition to Mn-2, Mn-1 and Mn-3
have produced formate in 61% and 84% yields, respectively while
other Mn-pincer/bidentate complexes and Fe and Co analogues gave
formate in yields up to 13%. Notably, no additional metal base
promoter, e.g., potassium tert-butoxide, is necessary in the
current reaction. Further hydrogenation of the product KHCO2
to methanol is not observed under current conditions. Analysis of
the gas phase after hydrogenation reactions revealed no detectable
CO2, CO, or CH4, indicating the distinct
selectivity of bicarbonate-to-formate transformation catalyzed by
the selected manganese complexes. Replacing THF by other organic
solvents, e.g., dioxane, triglyme, ethanol, 2-methyl-THF, or only
water as a single solvent, resulted in decreased formate yields
(2–77%,
Figure S3). Trials with three other bicarbonate salts based on
Na+, Cs+, and NH4+
cations led to moderate formate yields (46–69%) compared to KHCO3
(Figure
S4).
Carbonate Hydrogenation to Formate
After succeeding in the hydrogenation of
bicarbonates, we addressed the more desirable transformation of
carbonates to formates. Indeed, no hydrogenation of potassium
carbonate occurred applying Mn-2 complexes under the reaction
conditions shown in
Figure 2, and no formate was detected (Figure
3). To promote this less favored reaction, addition of a
carboxylic acid seems logical according to
eq 4. Compared to inorganic acids, e.g., HCl, the higher boiling
points of carboxylic acids are beneficial to maintain themselves in
reaction cycles. Thus, by simply adding propionic acid, some
conversion was observed, albeit the formate yield was low (19%).
Similarly, testing other dicarboxylic acids as well as α-amino acids
(AAs) revealed some reactivity. Surprisingly, the structure of the
AA has a strong influence on the formate yield. While in the
presence of the simplest AA glycine (Gly) a 11% formate yield was
observed, AAs bearing acidic side chains, e.g., glutamic acid (Glu)
and aspartic acid (Asp), led to much higher formate yields (up to
65%). In contrast, when utilizing basic AAs histidine (His), lysine
(Lys), and arginine (Arg), the yields of formate dropped
drastically. Apparently, a proper acidic media is important for
higher formate efficiency.
Figure 3
Figure 3.
Catalytic hydrogenation of potassium carbonate to potassium
formate. Conditions: K2CO3 (10 mmol),
Mn-2 (0.18 μmol), H2O/THF (5/5 mL), H2
(60 bar), additive, 90 °C, 12 h. Yield of formate is calculated by
(mmol formate)/(mmol K2CO3) × 100%.
Further investigations focused on Glu as
the best promoter. Variation of the Glu loading (0–150 mol %) led
to different formate yields (up to 65%,
Figure S6). Besides, CO2 was detected in the gas
mixture after hydrogenation reactions which derived from carbonate
salt. The optimal loading of Glu was found at 75 mol % based on
the formate yield. The Li+, K+, Rb+,
and NH4+ based carbonate salts gave the best
formate yields among the tested nine different carbonate species
(61–78%,
Figure S7). Finally, applying 5 μmol of Mn-2 catalyst,
formate was produced in 82% yield starting from K2CO3
(Figure
S8). Alternatively, using CO2 (10 bar) instead of
acid additives, a comparable formate yield (83%) was obtained (Figure
S19).
The positive influence of Glu on the
formate yield is explained by its dual function: (a) sufficient
acidity and (b) carbon dioxide capture ability. Indeed, control
reactions between K2CO3 and different acids
(Table
S1) showed that addition of propionic acid led to bicarbonate
as the main product (71%) with 8% CO2, while this ratio
was reversed in the presence of the stronger succinic acid with 3%
bicarbonate and 95% CO2. Interestingly, when Glu was
mixed with K2CO3, carbamate species were
observed in 26% yield along with bicarbonate (51%) and CO2
(16%), indicating the significant CO2 capture effect of
Glu. Apparently, the formation of carbamate species is substantial
to achieve a high yield in the carbonate-to-formate
transformation.
Hydrogen Production from Formate
For the development of
a round-trip hydrogen storage system, the release of hydrogen is
important, too. Hence, after having suitable conditions for
hydrogenate (bi)carbonates, we investigated hydrogen production from
formate under similar reaction conditions. In the absence of any
additive, Mn-2 among all the tested catalysts gave the best H2
yield (74%) and H2 purity (94.5%) besides CO2,
which results from the decomposition of bicarbonate (Figure
S9). To promote both the H2 yield and purity, the
effect of amino acid additives was investigated (Figure
4). After an equimolar amount of Lys
(67,78) was introduced to KHCO2, a quantitative H2
yield was found with more that 99% purity. Replacing Lys by Arg or
His, the yield of H2 dropped significantly. Interestingly,
a quantitative yield of H2 was obtained, albeit with a much
higher CO2 ratio (41.9%) by using Glu due to its increased
acidity.
Figure 4
Figure 4.
Catalytic hydrogen production from formates. Conditions: formate
(5.0 mmol), Mn-2 (5 μmol), additive (5.0 mmol), H2O/THF
(5/5 mL), 90 °C, 12 h. Yield of H2 is calculated by (mmol
H2)/(mmol formate) × 100%. The dotted lines serve as
guides to the eye.
In the absence of Lys, hydrogen was
produced generally in low purity (67–95%) applying various formate
salts based on Li+, Na+, K+, Cs+,
NH4+, Mg2+, and Ca2+
cations (Figure
S10). In contrast, promising results with both high H2
yield (>91%) and purity (>98%) were obtained in the presence of
Lys and the above-mentioned formate salts (Figure
4). According to
eq 3, carbonates are supposed to capture the released CO2
back to bicarbonates. To compare the CO2 capture
ability between K2CO3 and Lys, control
experiments were performed (Table
S4). Due to the presence of amino group in Lys, a significant
amount of carbamate species was obtained demonstrating the
superior CO2 (2 bar) capture ability of Lys (ca.
1.6-fold) compared to K2CO3 within 0.5 h.
Reversible H2 Storage
and Release Based on Bicarbonate/Formate Pair
For the implementation of a viable
hydrogen storage system, it is necessary to combine the two
individual processes, i.e., formate dehydrogenation and
bicarbonate hydrogenation and demonstrate the possibility of
stable hydrogen storage-release cycles (as illustrated in
Figure 1, right). Thus, starting from the selected formate
salt, H2 was generated in a 100 mL autoclave at 90 °C.
After completion of the reaction, a buret was used to collect H2,
and the autoclave was subjected to hydrogen storage under 60 bar
of H2 and 90 °C. The over pressure of H2 was
then released after cooling the reaction mixture to r.t., and the
dehydrogenation was repeated in the next cycle. Following this
protocol, several reaction systems were compared (Figure
5a,
Table S2). Starting from different formate salts based on Li+,
Na+, NH4+, Mg2+, Ca2+,
and Lys, although quantitative H2 yields (>95%) were
achieved in the initial dehydrogenation reaction, stepwise
decreased yields were observed after several cycles, which is
ascribed to the low efficiency in hydrogenation of corresponding
bicarbonate salts (Figure
S4). Interestingly, compared to other formate salts, applying
KHCO2, 80% of the initial H2 productivity
remained after five cycles with >99% H2 purity. The
formation of the bimetallic Mn–K intermediate
(79−82) is speculated to achieve such a good performance.
Figure 5
Figure 5. H2
storage-release cycles via interconversion of (a) bicarbonate/formate
and (b) carbonate/formate. Standard conditions: (a) formate/Lys
(5.0/5.0 mmol); (b) carbonate/Glu (5.0/5.0 mmol), Mn-2 (5
μmol), H2O/THF (5/5 mL), 90 °C, 12 h. H2
(60 bar) was applied in the hydrogenation step. H2
storage-release cycles start with (a) dehydrogenation and (b)
hydrogenation. H2 yields of each cycle are calculated
based on the initial loading of formate or carbonate salts,
respectively (each 5 mmol).
H2 Storage and
Release Cycles Starting from Carbonate Salts
Based on the relevant results of
hydrogenation of carbonate salts, we tested the cyclic
performance of H2 storage-release starting from K2CO3
and Glu. The favorable loading of Glu fell on 100 mol % (based
on carbonate salts) owing to the high H2 yields (up
to 94%) and reusability of the catalytic system (Table
S3). Indeed, from a K2CO3 and Glu
mixture at a pH of 7.9 (Table
S5), not only is less CO2 release expected, but
also the efficient CO2 capture by the amino groups of
Glu takes place under such basic conditions. As a result, the
amount of bicarbonate and carbamate after the first H2
storage and release cycle was measured to be 4.5 mmol (90% yield
based on initial loading of K2CO3,
Figure S20). Afterward, additional inorganic carbonates were
evaluated under the standard conditions (Figure
5b). Similar to K2CO3, a good
reactivity was also obtained using Na2CO3
and Rb2CO3 in four consecutive cycles with
up to 100% H2 yield. Carbonate salts based on Li and
Cs could be reused in at least three cycles. Moreover, it was
surprising that the easily available raw material MgCO3
could be utilized in the current H2 storage systems
achieving feasible efficiency (67% H2 yield) in the
first storage cycle, although decreased yields in the subsequent
runs were observed. On the other hand, calcium- and barium-based
carbonate salts gave H2 yields in up to 33% in the
initial cycles, due to their poor solubilities in water (ca.
0.02 mg mL–1 at 25 °C).
According to previous reports, FA/formates
dehydrogenation
(83,84) and its reverse reactions
(62,65,84−86) could be promoted by acids. In a detailed
study of FA dehydrogenation,
(83) the rate limiting step, i.e., decarboxylation of the
metal-formate intermediate (M–OOCH) is assisted by Lewis acid
(LiBF4) or Brønsted acid [Et3NH]+.
This lowers the activation energy of the decarboxylation
process, thus improving the reaction rates. In our earlier work,
(67,78) control experiments showed that the presence of an
α-amino acid group and an appropriate basic side chain in the
amine molecule are both crucial to facilitate the CO2
hydrogenation. Therefore, we propose that α-amino acids could
promote the H2 yield via stabilizing the Mn–OOCH
intermediate and accelerating the corresponding decarboxylation
process.
Conclusion
To conclude, we provide a viable Mn
promoted reversible hydrogen storage and release method via the
interconversion of largely available (bi)carbonate and formate salts
under comparably mild reaction conditions. For the first time,
low-cost carbonate salts could be applied as part of a H2
storage-release system with the help of naturally occurring AA
glutamic acid (Glu) as an additive, where the released CO2
could be ideally captured by the amino group of Glu and hydrogenated
back to formate to close the cycle. Notably, the overall system can
operate below 100 °C, making the utilization of so-called “waste
heat” possible.
(87) The dehydrogenation step of the resulting formate proceeds
smoothly without carbon dioxide liberation in the presence of
lysine. This enables hydrogen storage-release applications as shown
by several charge–discharge cycles with >80% H2 evolution
yield and >99% purity applying potassium formate, without reloading
of catalyst, solvent, and hydrogen carriers between each cycle.
Even though the hydrogen content of
formate (up to 2.85 wt %) is lower than that of FA (4.35 wt %), the
presented concepts have the inherent advantages of easy transport
and handling of the solid (bi)carbonate and formate salts compared
to the well-known carbon dioxide/formic acid couple (including our
previous work
(67,78)). Both the hydrogen acceptor and donor are nontoxic,
nonvolatile, noncorrosive, and nonacidic and show high solubility in
water.
(88) While the reported study paves the way for building up a
new H2 storage-release method, for larger scale
applications, it is desirable to improve the catalytic efficiency
even if an Earth abundant metal-based catalyst is applied.
