454626 Highly Selective Conversion of Biogenic Waste Materials to "Biogenic Formic Acid" As a Green Platform Chemical

Sunday, November 13, 2016: 5:18 PM
Union Square 21 (Hilton San Francisco Union Square)
Jakob Albert, Chair of Chemical Reaction Engineering and Peter Wasserscheid, Chair of Chemical Reaction Engineering, Erlangen, Germany

The selective oxidation of complex, water-insoluble and wet waste biomasses from second and third generation to formic acid including effective catalyst recycling is reported. Additionally, the relevance and limits of potential contaminants are illustrated by different experimental approaches. By using a very robust homogeneous polyoxometalate catalyst in aqueous solution, molecular oxygen as oxidant and an acid as solubilizer, it is possible to convert different lignocellulosic and algae feedstock into formic acid and pure carbon dioxide. The applied green oxidation system benefits from its low reaction temperature (below 100 °C) and its very selective nature. The upcoming requirements for sustainable production of chemicals will lead to an increasing substitution of fossil fuels by biorenewable feedstock. This is mainly due to the fact that almost all use of hydrocarbon fossil fuels in production processes is directly or indirectly linked to carbon dioxide (CO2) emissions that contribute to global warming. However, biomass conversion is usually characterized by complex feedstock compositions leading to sophisticated processes with many parallel and consecutive reactions involved. This often results in low yields of the designated products. Consequently, only few processes for directly converting biomass into value-added chemicals have been established in academic research. The high oxygen content of biomass requires typically a number of reductive conversion steps in biomass valorization sequences leading almost unavoidably to drawbacks like formation of tar or gluey byproducts. Moreover, the ultimate goal of producing valuable chemicals in high yields by direct conversion of complex biomasses, such as e.g. lignocellulosic substrates or algae is still far ahead. Major difficulties arise from the fact that such substrates are not well soluble in conventional solvents and are very resistant to chemical and biological transformations.

Other important considerations for the chemical use of biogenic feedstocks arise from conflicts with the use of biomass as food and animal nutrition. To characterize the potential for such conflicts, the broad range of biogenic feedstock conversion processes has been classified into three generations: The first generation of biomass conversion processes deals with feedstock that is in direct conflict with the supply of food. Examples are sugars, starch and vegetable oils. Conversion processes that utilize lignocellulosic biomass like wood or herbage do not use feedstock that is in direct conflict with food supply. However, these biomasses grow on farmland and are therefore in indirect conflict with food supply. The use of microorganisms, such as algae, to produce fuels or chemicals is denoted as third generation biomass conversion processes. These processes are preferred as future biomass conversion processes as no direct or indirect conflict with food and nutrition applications arises from a large scale, industrial realization.

Formic acid (FA) is a basic chemical that is widely used in chemical, textile, leather, pharmaceutical, rubber and other industries. Recently, FA is discussed in addition as fuel for direct FA fuel cells, and as hydrogen storage material. FA can be easily and selectively decomposed to hydrogen and CO2through metal catalysed processes under very mild conditions. In addition, thermal decomposition of FA (above 373 K) forms CO and water. Thus, FA can be regarded as a liquid syngas equivalent and conversion of biomass to FA can be regarded as up-grading an abundant natural resources into a green energy and synthesis equivalent.

Oxidative conversion of wet biomass to formic acid using homogeneous POM catalysts has been developed by our group in the recent five years (OxFA-process). The concept essentially overcomes all major problems of classical biomass gasification or reforming processes. The interesting difference of the OxFA route compared to reductive biomass valorization approaches is that all thermally induced formation of solid or gluey by-products is completely avoided. Moreover, all heteroatoms in the biomass are converted into their highest oxidized, water-soluble forms (e.g. sulfates, nitrates) in contrast to the formation of H2S and ammonia encountered under reductive conditions.

The OxFA-process is mildly exothermic and operates under mild temperature conditions of typically below 373 K. As water is applied as solvent, biomass of different origin, composition and humidity can be applied without drying. Formic acid is produced as the only product in the liquid phase avoiding tedious liquid phase separation processes. The only side-product is pure carbon dioxide that forms in the gas phase and can be re-used for the production of new biomass substrate (e.g. for the production of algae in an algae farm). As the OxFA process converts a very broad range of biogenic raw materials into only two products that separate nicely into gas and liquid phase its simplicity and robustness are clear advantages compared to other biomass valorization technologies. The OxFA process applies homogeneous polyoxometalate (POM) complexes as redox catalysts. POMs are characterized by strong Brønsted acidity, high proton mobility, fast multi-electron transfer, high solubility in various solvents and excellent resistance against hydrolytic or oxidative degradations in solution. The named properties make this class of catalysts very attractive for the oxidative conversion of biomass in aqueous media. In particular, their tunable redox potential and acidity offers enormous potential for their future use in oxidation chemistry. To increase the efficiency of the OxFA process for the conversion of water-insoluble, complex biomass, we recently introduced the use of sulfonic acids as additives. The latter have been found to act as efficient depolymerization catalysts, solubilizers and promotors. These findings have paved the way for the oxidative processing of e.g. lignocellulosic biomass, a class of substrates that is notoriously difficult to convert in aqueous phase reforming.

Oxidation experiments for expanding the substrate scope

In this contribution, we significantly expand on the scope of biogenic substrates that can be processed with the OxFA-process. Therefore, we first studied a range of second generation feedstocks like cane trash, pomace, fruit pulp, bark from hard and softwood, leafs, straw, grass clippings and aerated material. As representatives for third generation biomasses, we tested green algae like chlorella and ulva lactuca, brown algae like ascophyllum nodosom, blue-green algae like spirulina or cyanobacter and chondrus crispusas red algae. To enlarge the list of feedstocks even to the treatment of problematic and contaminated biological waste materials, we also applied effluent sludge and deinking sludge from a paper manufacturing plant under typical OxFA process conditions. Additionally, we used railway sleepers and beech condensates as substrates for the process. The effluent sludge contained about 80 wt% organic materials with a high nitrogen content of about 10 wt%. This substrate also contained large amounts (ca. 10 wt% in total) of inorganic materials and salts like calcium oxide and different phosphates. The applied deinking sludge was of even more inhomogeneous composition. Besides 35 wt% of organic material, it contained many different heavy metal salts like copper, lead, cadmium and nickel compounds. Furthermore, this substrate contained highly volatile halogenated hydrocarbons, like chloroethane and vinyl chloride as well as polycyclic and polychloric hydrocarbons. The applied railway sleepers were characterized by a high content of aromatic compounds. The applied beech condensate was also rich in aromatic compounds.

Based on our previous findings, we applied for the screening experiments of the named new substrates molecular oxygen as the oxidant, the polyoxometalate H5[PV2Mo10O40] (HPA-2) in aqueous solution as a catalyst and para-toluenesulfonic acid (TSA) as an additive and solubilizer. The conversion of all substrates was performed in a 600 mL Hastelloy autoclave with a gas entrainment impeller and additional cooling to deal with the exothermicity of the reaction. Effects of different liquid volumes in the reactor (100 mL of reaction mixture in 600 mL of reactor volume) can be excluded as the reaction takes place only in the liquid phase where the oxygen is entrained by the impeller. In particular, there is no further reaction in the gas phase. All biogenic feedstocks were processed for 24 hours at 90 °C under 30 bar oxygen pressure. The initial composition of the reaction mixture in each oxidation experiment was 100 g of water as a solvent, 1.74 g (0.9 mmol) of HPA-2 polyoxometalate as a catalyst, 1.90 g (11 mmol) of TSA as additive/solubilizer and 3.30 g of the substrate under investigation. To enable the calculation of yields based on the comparable basis of carbon atoms available in the substrate, we performed an elemental analysis of each substrate prior to our catalytic experiments (results see Table 1).

The results in Table 1 clearly show that the applied homogeneous oxidation system is able to convert all applied 2nd and 3rd generation biomasses to predominantly formic acid and carbon dioxide. Under the applied conditions no CO formation from FA decomposition and no catalytic FA decomposition to CO2and hydrogen were detected (see Table 1, entry 19). The yield in FA was found to depend strongly on the structure of the substrate and the number of oxygen-functionalities in the carbon-framework. Among the tested new substrates, the best raw material for producing formic acid was found to be pomace with 55 % FA-yield (Entry 1) followed by cane trash with 49 % FA-yield (Entry 2) and fruit pulp with 45 % FA-yield (Entry 3). Note that all these latter three origin from commercial and industrial biomass conversion processes making the here presented technology an interesting add-on to widely practiced processes of the food industry.

Table 1: Oxidative conversion of complex, water-insoluble biomass of 2nd and 3rd generation in presence of the additive TSA using a HPA-2 POM-catalyst.

Entry

Substrate

Molecular compositiona

Combined Yieldb FA+CO2 [%]

FA-Yieldb [%]

FA:CO2-selectivityb [%]

1

Pomace

C0.24H1.59O1

97.3

54.6

56:44

2

Cane trash

C0.92H2.04O1

97.5

48.9

50:50

3

Fruit pulp

C1.00H1.92O1

99.0

45.3

46:54

4

Spruce chips

C1.30H1.98O1

89.5

34.6

39:61

5

Poplar splint

C1.32H1.98O1

66.0

30.8

47:53

6

Grass clippings

C1.17H2.05O1

74.6

29.9

40:60

7

Chondrus crispus

C0.71H1.30O1

48.3

21.7

45:55

8

Chlorella

C1.69H2.92O1

53.7

21.6

40:60

9

Oak bark

C1.21H1.87O1

56.0

21.1

38:62

10

Willow bark

C1.34H2.15O1

59.5

16.1

27:73

11

Cyanobacter

C0.72H1.32O1

27.4

15.8

58:42

12

Spirulina

C1.70H2.99O1

40.6

15.7

39:61

13

Birch bark

C1.39H2.29O1

41.0

15.5

38:62

14

Ascophyllum nodosum

C0.82H1.42O1

29.2

14.8

51:49

15

Straw

C1.04H1.62O1

53.1

11.9

22:78

16

Nettle leafs

C1.10H1.98O1

54.2

11.0

20:80

17

Ulva lactuca

C0.75H1.53O1

16.4

7.6

46:54

18

TSA

C7H8O3S

-

-

-

19

FA

CH2O2

-

-

-

20

Neat pomacec

C0.24H1.59O1

52.0

12.5

24:76

21

Neat cyanobacterc

C0.72H1.32O1

14.8

4.3

29:71

Reaction conditions: 3.3 g substrate, 1.9 g (11 mmol) additive and 1.74 g (0.9 mmol) catalyst dissolved in 100.0 mL H2O, 30 bar O2, 90 °C, 24 h, 1000 rpm; a Determined via C, H, N, S elemental analysis; b Yield and selectivity determined by means of 1H-NMR using benzene as an external standard according to n (FA)/n(C-atoms feedstock); c Reference experiments without TSA additive.

Interestingly, also substrates from the third generation like red algae (Chondrus crispus) or green algae (Chlorella) gave surprisingly high FA-yields. Note, that for these substrates the selected reaction time of 24 hours was definitely too short, so not all the substrate was converted to FA and CO2. From our experience, it can be expected that prolonged reaction times would lead to a final FA yield for these substances close to the FA selectivity obtained after 24 h reaction time. These experiments could also confirm the very positive effect of the toluenesulfonic acid (TSA) additive and solubilizer on the POM-catalyzed biomass oxidation. Experimental evidence is provided by the reactions with and without the TSA additive (compare Table 1, entries 1 and 11 against entries 20 and 21). For pomace, the experiment without TSA (Entry 20) showed only very little FA formation (Y (FA) =12.5 %) compared to a FA-yield of 54.6 % (entry 1) with TSA for otherwise identical conditions. For cyanobacter, the experiment without TSA (entry 21) yielded only 4.3 % FA after 24 h reaction time, in contrast to the experiment with the TSA additive in entry 11 (15.8 % FA-yield). Based on these results, it is obvious that the TSA additive does not only act as a solubilizer but also as reaction accelerator and selectivity enhancer. The latter can be explained by the solubilizing effect of the TSA additive giving the homogeneous POM catalyst more access to the functional groups within the carbon-framework. As these groups are normally carbonyl- or hydroxy-groups this leads to an increase of the FA-yield based on our previous findings with different funtionalized C2-model compounds. These results are in good agreement with previous findings concerning the suitability of the TSA additive for the transformation of weekly water soluble biomass substrates using homogeneous POM catalysts in water. A baseline experiment using pure TSA as the substrate showed again no oxidation neither to FA nor to CO2(see Table 1, entry 18).

In an additional set of experiments, we tested conversion of chemically contaminated substrates in the POM catalyzed oxidation. For this purpose, we processed effluent sludge, railway sleeper, beech condensate and deinking sludge under typical OxFA conditions. In these experiments, the initial composition of the reaction mixtures contained 100 g of water as a solvent, 1.74 g (0.9 mmol) of HPA-2 as a catalyst and between 1.85 g and 10 g (100 mmol) of the substrate dependent on its water content. The additive TSA was not used in these experiments as the unknown composition of the applied “real life” substrates was expected to open additional but unknown reaction pathways for TSA (e.g. salt formation with metal ions). This would lead to consecutive reactions and uncontrollable loss of the additive which would have made interpretation of the obtained results rather difficult. For all experiments of this series the reaction temperature was kept at 90 °C in order to prevent thermal FA decomposition. The oxidation reactions were carried out at 30 bar oxygen atmosphere with stirring of 1000 rpm for 24 h reaction time. The obtained results are summarized in Table 2.

Table 2: Oxidative conversion of chemically contaminated biomass substrates with the HPA-2 catalyst.

Entry

Substrate

Molecular compositiona

Combined Yieldb FA+CO2 [%]

FA-Yieldb [%]

FA:CO2-Selectivityb [%]

1

Effluent sludge

C6.31H11.35O6.63N1

46.1

15.7

34:66

2

Railway sleeper

C1.25H1.90O1

39.5

11.2

29:71

3

Beech condensate

C1.21H1.86O1

2.5

1.0

41:59

4

Deinking sludge

C0.38H0.47O1

-

-

-

Reaction conditions: 100 mmol substrate (1.85 g for entry 1; 5 g for entry 2; 10 g for entry 3 and 5.35 g for entry 4) depending on molecular weight; 1.74 g (0.9 mmol) catalyst dissolved in 100.0 mL H2O; 30 bar O2, 90 °C, 24 h, 1000 rpm. a Determined via C, H, N, S elemental analysis; b Yield and selectivity determined by means of 1H-NMR using benzene as an external standard according to n(FA)/n(C-atoms feedstock).

It is obvious that the HPA-2 catalyst also works with contaminated feedstock and is highly robust against many impurities. Thus effluent sludge with the highest amount of organic and oxidisable matter yields nearly 16 % FA at a combined FA + CO2 yield of 46 % after 24 h reaction time. The high nitrogen content within the sludge from a wastewater treatment plant has a negative effect on the conversion rate but does not prevent the oxidation reaction. Also for railway sleepers containing high amounts of aromatic compounds, the oxidation to FA and CO2 is observed. A FA + CO2combined yield of 40 % is observed after 24 h resulting with FA selectivity of 29 %. In contrast to the latter two substrates, highly diluted beech condensate and deinking sludge show almost no oxidative conversion after 24 h reaction time.

Defined poisoning experiments for the POM catalyst

In order to elucidate the complete lack of reactivity with the deinking sludge substrate we performed a couple of poisoning experiments to learn more about the sensitivities of the applied POM oxidation system. Our first hypothesis was that the heavy metal content in the applied deinking sludge (Palm GmbH & Co. KG paper manufacturing company) could be responsible for the observed catalyst deactivation. Such heavy-metal cations are expected to form water-insoluble and inactive complexes with the anion of the polyoxometalate catalyst. To confirm this hypothesis, we added on-purpose water-soluble salts of higher transition metals like copper, lead, cadmium and nickel as contaminants to a glucose oxidation experiment. Moreover, similar contamination experiments were performed with other typical components of deinking sludge, such as e.g. gypsum, kaoline, TiO2 or ZrO2. For these experiments, 5 g (25 mmol) of glucose as a water-soluble substrate and 1.74 g (0.9 mmol) of the catalyst dissolved in 100 mL water, together with 8 mmol (0.63-2.18 g) of the impurity were processed under 90 °C, 30 bar oxygen with 1000 rpm for 6.5 h in a 600 mL high-pressure autoclave. All results are displayed in Table 3; the baseline result for a non-contaminated glucose oxidation experiment is shown as entry 9.

Table 3: HPA-2 catalysed glucose oxidation - defined poisoning experiments with typical impurities from paper manufacturing to elucidate the lack of catalytic activity for deinking sludge as substrate.

No.

Impurity

Potential source of origin

Combined Yielda FA+CO2 [%]

FA-Yielda

[%]

FA:CO2-Selectivitya [%]

1

CuSO4

printer’s ink

39.8

-

0:100

2

CdCl2

printer’s ink

87.6

44.4

51:49

3

NiCl2

printer’s ink

n.d.

n.d.

n.d.

4

PbCl2

printer’s ink

6.2

4.4

71:29

5

Gypsum

filler

89.9

47.5

53:47

6

Kaoline

filler

93.7

49.2

53:47

7

TiO2

filler

72.3

44.4

61:39

8

ZrO2

paper varnish

83.0

38.8

47:53

9

-

-

100

46.2

46:54

Reaction conditions: 5 g (25 mmol) glucose as a substrate; 1.74 g (0.9 mmol) HPA-2 as a catalyst dissolved in 100.0 mL H2O; 8 mmol (0.63-2.18 g) of the impurity depending on molecular weight; 30 bar O2, 90 °C, 6.5 h, 1000 rpm. a Yield and selectivity determined by means of 1H-NMR using benzene as an external standard according to n(FA)/n(C-atoms feedstock).

The results in Table 3 clearly indicate a strong influence of Pb- and Cu-salts on the catalytic activity for glucose oxidation. Other heavy metal ions, like Cd2+, appear to be of little effect. Interestingly, the presence of copper ions does completely inhibit the formation of FA from glucose oxidation, only CO2 is detected as a product. This goes along with the observation that copper ions form water-insoluble complexes with the catalyst heteropolyanion in form of small blue particles at the bottom of the reaction vessel. Unfortunately, the oxidation result with the added NiCl2 could not be quantified by the usual NMR-method due to the strong deformation of the spectra due to the paramagnetic nature of Ni (II) in the sample. Nevertheless, also for this case, the formation of FA could be qualitatively confirmed by the help of 13C-NMR. In contrast to the experiments with Cu, no formation of water-insoluble crystals was observed in the case of NiCl2 addition. Finally, the addition of Pb (II) species to the aqueous glucose solution resulted in an immediate color change from orange to a blue-green solution. The 51V-NMR spectra of the obtained solution strongly indicate complexation of the heteropolyanion with the Pb2+ cations. In addition, the formation of yellow crystals at the bottom of the reaction vessel was observed thus removing the active catalyst from solution. In general, 51V-NMR spectra of the reaction solutions containing late transition-metal cations show significant changes after addition of Cu-, Ni- and Pb-salts. In contrast, the 51V-NMR spectra with added fillers do not show significant changes of the polyoxometalate catalyst. Addition of these water-insoluble fillers, like gypsum, kaoline, titanium dioxide and zirconia, resulted in an increase of the viscosity of the reaction mixture; however, this had very little effect on the observed glucose oxidation reaction.

Our contribution gives strong support to the scientific and economic attractiveness of homogeneously catalyzed biomass oxidation to formic acid by polyoxometalate catalysts. We could expand the substrate scope for the reaction considerably by converting a very wide range of substrates including many complex water-insoluble biogenic feeds and also contaminated materials. While waste from the food industry, effluent sludge and even railway sleeper was found to be effectively converted under typical OxFA reaction conditions (90 °C, 30 bar O2, 24 h reaction time), a clear and so far undetected limitation has been found for metal ion containing feedstocks. Whenever contaminated biological feed with metal ions forming stable water-insoluble complexes with the catalytically active polyoxometalate ion were present in the applied feedstock, the reactivity of the oxidation system was strongly reduced or almost completely eliminated. The favorable role of the TSA additive for supporting depolymerization and solubilisation of 2nd and 3rd generation was impressively underlined by comparative oxidation experiments with pomace and cyanobacter. From all results of this study we anticipate a high potential of POM-catalyzed biomass oxidation for future decentralized biomass conversion processes of wet and contaminated substrates. As formic acid is a chemical product used in high amounts by farmers (e.g. in the preparation of silage) and waste streams from farming or food manufacturing represent very attractive feedstocks, the technology appears particularly attractive if linked to agricultural activities or small to medium sized food industries.


Extended Abstract: File Not Uploaded