+

Preliminary Draft of Chapter for “Green Chemistry” Book – 12/16/96h BNL-63763

Transition Metal Catalyzed Reactions of Carbohydrates:

A Nonoxidative Approach to Oxygenated Organics

Mark Andrews, Chemistry Department

Brookhaven National Laboratory, Upton, NY 11973-5000

TABLE OF CONTENTS

I - Background and Introduction ................................ ........................................ 2

Oxygenated Organics: Problems with Hydrocarbon Oxidation Approaches .....2

Oxygenated Organics: Advantages of Non-Oxidative Biomass Approaches .....3

Oxygenated Organics from Biomass: Methods and Research Opportunities ....4

Carbohydrate Catalysis Research Needs ................................ .............................5

II - Progress to Date . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . **o* . . . . ..*.. *o**oo.. *.. o* . . . . . . . . . . . . ..o . . . ..o* . . ..*oo . . ..o... o ‘7

Prior Literature ................................ ................................................................ ......7

Aldose Decarbonylation ................................ ......................................................... 8

Catalytic Hydrocracking of Sugars ................................ ....................................... 9

Pt and Pd Diolate and Alditolate Complexes ................................ ....................... 10

Catalytic Diol Deoxydehydration ................................ .......................................... 14

III - Opportunities for the Future ................................ ....................................... 15

Complexation of Cyclic and Disaccharide Sugars/ Other Metals ....................... 15

Development of Catalytic Polyol Disproportionation ................................ ........... 18

Development of Catalytic Carbohydrate Ionic Hydrogenolysis ........................... 19

Development of Aqueous Organometallic Carbohydrate Chemistry ..................22

N- Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

N- Bibliography *.. *... *.. o.*.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ●..0. 24

1

I - BACKGROUND AND INTRODUCTION

Oxygenated Organics: Problems with Hydrocarbon Oxidation

A prime target for new, environmentally benign,

Approaches

organic synthesis

technologies is the production ofoxygenated organics, key compounds in the fuel,

commodity, andspecialty chemical markets.1 Most of these products are currently

derived from the partial air oxidation of petroleum or natural gas fossil feedstocks.

While hydrocarbons and oxygen are presently cheap raw materials, their utilization

poses long-standing, recalcitrant problems, many with unfavorable environmental

repercussions. The following commercial example is illustrative. Adipic acid, a two

billion lb/yr nylon precursor, is currently produced2’3 by a multi-step process which

begins with the hydrogenation of carcinogenic benzene to cyclohexane, followed by

catalytic air-oxidation to give a cyclohexanone / cyclohexanol mixture in about 75%

yield at 5% conversion. After recycle of unreacted starting material by distillation,

this ketone-alcohol mixture must be further oxidized with corrosive nitric acid to

produce not only the desired adipic acid, but nitrogen oxides, which are only

partially recyclable. Thus, the overall process involves a reduction, two oxidations,

and a distillation, and generates several environmentally hazardous waste streams.

Furthermore, the air oxidation of organic materials almost always involves

potentially explosive conditions, which require expensive, specially designed reactor

systems to account for this latent safety/environmental problem.

The fundamental problematic issues, which adversely affect not only process

economics but the environment and public perception, are summarized in

Highlight 1. While solutions to some of these problems may come from current

research efforts, other problems are inherent in the approach itself. For this reason,

we believe that alternative “green chemistry” strategies merit serious exploration.

DISCLAIMER

This report was prepared as an account of work sponsoredby an agency of the United States Government. Neitherthe United States Government nor any agency thereof, norany of their employees, make any warranty, express orimplied, or assumes any legal liability or responsibility forthe accuracy, completeness, or usefulness of anyinformation, apparatus, product, or process disclosed, orrepresents that its use would not infringe privately ownedrights. Reference herein to any specific commercialproduct, process, or service by trade name, trademark,manufacturer, or otherwise does not necessarily constituteor imply its endorsement, recommendation, or favoring bythe United States Government or any agency thereof. Theviews and opinions of authors expressed herein do notnecessarily state or reflect those of the United StatesGovernment or any agency thereof.

DISCLAIMER

Portions of this document may be illegible

in electronic image products. Images are

produced from the best available original

document.

Highlight 1. Fundamental Issues in Hydrocarbon Oxidation

Technological Issues

● The chemical inertness of hydrocarbons leads to high reaction activation

energies, often resulting in poor selectivities due to greaterreactivity of primary oxidation products.

● Poor air-oxidation selectivity wastes valuable feedstock.● Other oxidants, while frequently more selective, are much more

expensive or, if produced from oxygen (e.g., R02H), generatestoichiometric co-products when used.

● Oxidations with air or peroxides pose explosion hazards.

Soc ietal Issues

● Hydrocarbons are a non-renewable, fossil resource.

● Crude oil production is fraught with political and environmental concerns.● Transportation of natural gas and petroleum products poses safety and

environmental hazards.● Poor air-oxidation selectivity generates significant toxic waste streams.

Oxygenated Organics: Advantages of Non-Oxidative Biomass Approaches

One such alternative approach is the production of oxygenated organics from

biomass. Biomass carbohydrates are not only our most abundant organic feedstock,

but they are renewable and “pre-oxygenated”. This leads to a complete shift in

strategies for producing oxygenated organics compared to hydrocarbon oxidation.

The objective now generally becomes the partial removal or redistribution of the

oxygen already present in these feedstocks, rather than its problematic addition.

Advantages of this non-oxidative approach are summarized in Highlight 2. The

societal advantages of a biomass approach are particularly significant, as these

issues are ever-growing contributors to governmental regulations and industrial

decision-making considerations. In particular, environmental costs can radically

change the overall economics of a process, such that feedstock prices no longer

dominate the equation. While there are potential

associated with biomass production, such as monocrop

environmental concerns

plantings and the use of

3

.

+

fertilizers, herbicides, and pesticides, these issues should be more tractable than

those associated with fossil feedstocks.

Highlight 2. Advantages of Biomass Route to Oxygenated Organics

Technological Advantaae~

● Rich functionality in biomass fosters low reaction activation energies,

permitting mild reaction conditions and hence lower capital andoperating costs.

● Hydrogen-bonding characteristics of carbohydrates should be veryhelpful in promoting highly selective reactions.

● Conversion processes will primarily involve reductive reactions, which

are typically more selective than are oxidative reactions.

I Adv~

● Biomass is an abundant, widespread, sustainable resource.● Production and transportation of biomass are relatively safe and

environmentally friendly.● Biomass process solvents and waste streams are likely to be relatively

biocompatible.

Oxygenated Organics from Biomass: Methods and Research Opportunities

There are two major traditional approaches to biomass conversion

(Highlight 3), fermentation and high-temperature pyrolysis.4 “We believe that both

of these will be significant contributors to future oxygenated organics technologies.

While both approaches can effectively utilize raw, multi-component biomass, they

sufTer from limitations on the range of products that can be obtained and require

substantial energy input. Recently, we have proposed a new strategy based on

homogeneous transition-metal catalysis,5-7 a technology that has become

increasingly significant in the latter part of this century. 2

1----—Highlight 3. Biomass Conversion Strategies

● Fermentation● High-temperature pyrolysis* Homogeneous transition-metal catalysis

4

.

.

Our approach recognizes that the high selectivity and reactivity of homogeneous

transition-metal catalysts are an excellent match to the rich and fragile

functionalities (polyhydroxyl, carbonyl, and acetal) present in biomass

carbohydrates. As such, homogeneous catalysis should be an effective means of

converting specific carbohydrates, obtainable from sugar-rich crops and biomass

fermentation, into valuable oxygenated organics. The potential advantages of a

homogeneous catalysis approach are its greater product versatility and lower energy

requirements compared to pyrolysis and straight fermentation. At present,

however, the design of appropriate catalytic conversion systems is strongly inhibited

by a lack of information on the interactions of native carbohydrates with

organometallic homogeneous catalysts. We believe that this situation presents a

unique basic research opportunity. Highlight 4 gives illustrate ve examples of the

types of carbohydrate conversions that we have begun to explore that might be

feasible using transition-metal catalysis:

Highlight 4. Potential Metal-Catalyzed Carbohydrate Conversions

● Hydrocracking to glycerol, ethylene glycol, and methanol.

● Deoxydehydration to allylic alcohols, e.g., xylitol to pentadienol.● Disproportionation to unsaturated aldehydes or acids, e.g., glycerol to

acrolei n.● Dehydroxylation to cx,~-diols and derived compounds, e.g. glucose to

adipic acid, erythritol to tetrahydrofuran.

Carbohydrate Catalysis Research Needs

Highlight 5 gives the fundamental objectives identified as being essential to

effective attainment of transition-metal catalyzed transformations of carbohydrates.

While our primary focus has been on exploring and understanding the interactions

of organotransition metal complexes with carbohydrates, we have also found it

advantageous to model these interactions by studying the simpler case of metal–diol

5

Highlight 5. Fundamental Metal–Carbohydrate Research Objectives

● Synthesis and characterization of organometallic carbohydrate anddiolate complexes as models for catalytic intermediates.

. Understanding factors affecting complexation stabilities and selectivities.● Exploring reactivities of carbohydrate/diolate complexes.● Determining solvent effects on complexation and reaction chemistries.● Discovering catalytic cycles to effect carbohydrate/diol redox chemistry.

chemistry.s’g

concerns here

Solvent issues are also a significant secondary focal point.

include simple volubility problems, interference with catalysis

The

from

solvent coordination via donor atoms or hydroxyl groups, effects of solvent

hydrogen-bonding interactions, and product isolation problems. Some of the

carbohydrate redox transformations and catalytic intermediates that we envision

and will describe later include the following:

/

LmM’H +Hz

+H2 / –H20

r

RH2C-CH2R

~1HO R

I

LmM’H +H2

HO R2 RCH20H

.

.

For sugars, R = CH20H, etc.

6

.

II - PROGRESS TO DATE

Prior Literature

One of the first reported homogeneously catalyzed reactions of carbohydrates

was by Kruse of ICI in 1976, the RuHCl(PPhs )3 mediated hydrogenation of aldose

and ketose sugars to the corresponding alditols in amide solvents: I“’ll

0“H o

HOH

OHH—HO

H

H, ~0c

H

*

OH

HO H

H OH

A :,~’owfip:t:Kruse(ICI) GLUCOSE

H2*

RUHCL3

CHZOH

H

$

OH

HO H

H OH

H OH

CHZOH

GLUCITOL

Note that, despite the small proportion of open-chain aldehydo-glucose present in

solution, 12 the metal catalyst is capable of hydrogenating the “latent” carbonyl

group. Prior to our studies, there were also several related reports of catalytic

transfer hydrogenation and disproportionation reactions of sugars, 13-1* and one

example of an isolated organometallic carbohydrate complex, (1,2-

CGHA(P*P~e)2)Pt(CsHsOs),19 derived from the simplest possible carbohydrate,

glycerol. While diastereoselectivity is observed in the formation of this complex,

complexation of glycerol poses minimal regioselectivity considerations (only 1,2- vs

1,3-) compared to up to 15 possibilities in an asymmetric C fj alditol. In the reported

study, only 1,2-glycerol complexation was observed:

.

Aldose Decarbonylation

Our first work in metal-mediated carbohydrate chemistry involved a novel

approach to carbohydrate “descent of series” chemistry, 20 the decarbonylation of

aldose sugars (e.g., glucose, R = H), by Wilkinson’s catalyst, Rh(PPhJsCl: 21’22

7H20H H. Oc+ $H20H

H+

OH CH20H

tH20H +CH20H

Q

oRhCl(CO)L2

HO ~H3

~= OH NOSolvent = u (NMP)

OHAndrews,Klaeren&Goul(

While the reaction is not catalytic nor particularly suited to the production of

oxygenated organics, the study confirmed Kruse’s early observation 1 that amide

solvents, such as lV-methyl-2-pyrrolidinone (NMP), are an effective medium for

conducting organometallic transformations of native carbohydrates. Again, the

metal reagent has the notable ability to selectively react with the tiny fraction

(< 0.01%) of open-chain, aldehyde form of the sugar present without resort to

temporary protecting groups. 23 The reactions are clean and predictable, offering a

convenient and economical synthesis of authentic reference samples of certain types,

of alditols (sugar alcohols) for comparison with natural produ~ts. For example,

glycosylpentitols can be prepared in a simple single step from readily available

disaccharides (as illustrated above for lactose), a process that would otherwise

constitute a multi-step synthetic challenge with generation of concomitant waste

8

.

*

streams. With ketose sugars, we observed decarbonylation and dehydration

products, including a variety of furans. 24 We have also observed formation of

furandimethanol (FDM) when the Kruse hydrogenation system is applied to

fi-uctose.6 FDM is a difunctional aromatic monomer used to manufacture flame-

resistant polymers, and is also a precursor to tetrahydrofurandimethanol, an

increasingly important diol.25

Catalytic Hydrocracking of Sugars

Our second major study also involved chemistry dependent upon the latent

carbonyl group present in monosaccharide sugars. This work demonstrated that

aldose and ketose carbon–carbon single bonds could be catalytically hydrocracked

by H2Ru(PPhs)A via a combination of a retro-aldol reaction and carbonyl group

hydrogenation, resulting in the formation of lower polyols: 5J6

CH20H Hexitols(64%) CH20H

$0

t

\HOH

RUH2L4 (2%)CH20H

HO H

+2H OH +

HO H

H OH Hz (300 psi) H OH

100 “C, 24 hCH’20H

H OH H OH

CHZOHNMP Glycerol (15%)

CH’20H----- ----- ----- ----- ----- ------ ------ --

The hydrocracking of fmctose is very selective for glycerol; the only other significant

J- ORUH ~2

0

HO HO+

/ RUH2

species present are unreacted fructose and the two hexitols derived from simple

9

.

hydrogenation of the fructose carbonyl group. Addition of a basic co-catalyst (KOH)

accelerates the reaction and increases the selectivity for hydrocracking over simple

hydrogenation (3 l% yield ofglycerol). Unfortunately the mass balance and cracking

selectivities are then somewhat reduced, due to sugar degradation and enhanced

formation of Cz, Cd and C5 fragments. Key findings of these studies are that the

transformation occurs under much milder conditions (100 ‘C/ 300 psi Hz) than with

previously patented heterogeneous catalysts (2OO ‘C / 2000 psi Hz )26-34 and with

greater selectivity (no formation of partially deoxygenated products, e.g., 1,.2-

propanediol). An improved version of this chemistry would be desirable as glycerol

is potentially a good precursor to a variety of oxygenated organics.

Pt and Pd Diolate and Alditolate Complexes

Though it is clearly possible to do catalytic chemistry involving carbohydrate

carbonyl groups, the most characteristic functionality present in these

polyhydroxylic compounds is the diol unit. This grouping is found even in

polysaccharides such as starch and cellulose where the carbonyl group is fully

masked as an acetal, rather than as the equilibrating hemiacetal found in

monosaccharides. Building on the brief literature reports of (Lz)Pt(diolate)

complexes, 19’35we have now made an extensive study of bis(phosphine) platinum(H)

diolate and sugar alcohol complexes. 8’9 The key to this entire study was provided by

the development of a new method for synthesizing diolate complexes, reaction of a

bis(phosphine) platinum carbonate, L2Pt(COs),36 with the diol or alciitol:

10

This reaction is nearly thermoneutral, hence various exchange reactions allow

determinationof relative complexation constants as afunction ofthesubstituents

on the diol and ancillary phosphine ligand. For different 1,2-diols with Lz =

l,3-bis(diphenylphosphino)propane (dppp), these cover arangeof almost 100. For

different phosphines with ethane-l,2-diol ,therelative complexation constants cover

a similar range. The total range, from (1,2 -bis(dicyclophexylphosphino)ethane)-

Pt(pinacolate) at the low end to (cis-1,2-bis(diphenylphosphino)ethene)-

Pt(phenylethane-l,2-diolate)at the high end, is estimated tobe over 10s. In all

cases, electron-donating groups on both the diol and phosphine lead to lower diol

binding constants. Complexationof l,3-diols is about afactorof100–1000 lower

than that of 1,2-diols.

With alditols, isomeric complexes are possible, even after requiring

complexation to occur via an a, ~-diol linkage. In practice, significant

regioselectivities are observed, which vary strongly with alditol stereochemistry,

favoring coordination to internal threo diol units (e.g., 2,3-galactitol, 3,4-mannitol)

over erythro and terminal diol units (3,4-galactitol, 2,3- and 1,2-mannitol):

11

.

—

P

CH20H HoH2y~ 83% 1,2-Isomer 14%~ ~ 1670

:: H+C)H~17% 2.3-Isomer 86-

H+OH . . .

H+

OH

@El cH@H

[

Ph Ph\/

~p ~ RI\ /0/p(\

o R2/p\

Ph ‘Ph

] “+OHHOH2C

EZl

= 1 J @zmlcH@H~11%

HOH2C7% ~

+

1,2-Isomer 1%~

$

~ 3%H OH HO H

89V0~ ~ 71~0 2,3-Isomer 17%~ ~ 5%HO H

470 ~ ~ 18% 3,4-Isomer 82%~

t

HO H

H ‘H t_cD,a2_J ‘: ~ “r

yridine CH20H HOH2CAndrews& VOSS Pyridine

These ratios are thermodynamic, not kinetic, as isomerization is fast on the

laboratory time scale but slow on the NMR time scale. Hydrogen-bonding

interactions are important contributors to these regioselectivities,as illustrated by

the x-ray diffraction determined structure of (dppp)Pt(3,4-mannitolate ):9

~h ~h Hs. ?

: , .O.. OH

< >J

‘\?’ ),,,,\\i

OH/ptJ \

/! ] .0‘h Ph Ho”

Detailed studies show that the hydrogen-bond acceptor (HBA) strength of the

platinum diolate oxygens is comparable to that of the strongest neutral 13BA’s.~7

The intramolecular hydrogen-bonding interactions in the Pt alditolate complexes

are thus sufficiently strong that they are retained to a large extent even in neat

12

HBA solvents such as pyridine, leading to only minor changes in complexation

isomer ratios as a fi.mction of solvent (cf. comparisons in above Figure).

While the (L2)Pt diolate and alditolate complexes are remarkably stable

thermally, they undergo facile photochemical oxidative-cleavage of the diol C–C!

bond to give two carbonyl fragments and a reactive (L2 )PtO species.g When the

photolyses are run in the presence of hydrogen and catalytic HzRu(PPhs)A, the

ketones oraldehydes can bereduced tothe corresponding alcohol. Synthesis of the

starting diolate ii-omthe platinum carbonate canalso be integrated into the overall

reaction. The net transformation then corresponds to hydrocracking of the starting

diol carbon–carbon single bond. Thus it is possible to selectively convert mannitol

into glycerol: 6

$OH

HOit I

OH

HO

‘“OHHO

Hz,hv~2

1.0 (dppe)Pt(COs)

0.1 HzRu(PPhs)4

NMP

EOH

OH

OH

While this reaction is clearly not of practical utility, the results demonstrate that

carbohydrate complexation selectivity, facilitated by the rich functionality of

carbohydrate substrates, can be converted into reaction selectivity.

Since Pd(II) is a much more catalytically active metal than Pt(II), we have

investigated palladium carbohydrate chemistry. Initial results (S. K. Mandal and

K. S. Koenig) were disappointing, however, as competition studies show that diols

and alditols bind about a factor of 100 times poorer to (L2 )Pd(II) than to the

corresponding (L2 )Pt(II) center. Furthermore, while (dppp)Pd(3,4-mannitolate) has

been isolated, it is very sensitive to decomposition by water and shows no sign of

thermal oxidative-cleavage of the central C–C bond, possibly due to Woodward -

Hoffiann symmetry constraints. 8

13

Catalytic Diol Deoxydehydration

Since Pt(II) and Pd(II) diolate complexes exhibited no signs of thermal redox

activity, we looked elsewhere for possible systems to effect catalytic carbohydrate

transformations. Based on several stoichiometric reactions reported by Herrmann38

and Gable39 in conjunction with their studies of alkene oxidation, we have now

designed and implemented the catalytic cycle shown to the left below:7

R,mdrews & Cook

HO OH

)--’

-2~0 Cp*Re03* ~hF

Ph + 90“c +PPh~ O=PPh3

50, t“ 1

II ttt4.2 “~.O.I h ,“ A

.,. r i- 7 ‘/—\

6 4 8 12 16

Time (h)

For the simple diol phenyl- l,2-ethanediol, the reaction to give styrene proceeds very

well in solvents such as benzene or chlorobenzene using triphenylphosphine as a

sacrificial reductant (above right). The initial rate of the reaction is essentially

equal to that observed by Gable for alkene extrusion from pure

Cp*ReO(phenylethanediolate),40 suggesting that extrusion is the rate-limiting step

in the catalytic cycle. In donor solvents, such as tetrahydrofuran (T13F) and NMP,

the catalyst dies after a few turnovers. A primary cause of this has been identified

14

(over-reduction of the catalyst to a Re(III) species) and ways to prevent this from

occurring have been discovered (use of an acid co-catalyst or a weaker sacrificial

reducing agent). The reaction is stereospecific as shown by the conversion of the

protected sugar l,2:5,6-diisopropylidenemannitol to the corresponding trans alkene.

Alditols are also deoxydehydrated, glycerol yielding allyl alcohol and the C4 alditol

erythritol giving not only 3-buten- 1,2-diol and cis -2-buten-l ,4-diol, but the fully

deoxygenated product butadiene as well. This technique is at least as good as most

current methods for converting diols to alkenes, the typical approach being reaction

with thiophosgene to give a thiocarbonate, which is then heated with a phosphine to

give the alkene, phosphine sulfide, and carbon dioxide.41 For carbohydrate

substrates, the reaction shows promise of selectively converting specific hydroxyl

groups into another reactive but readily differentiated functional group, an alkene,

or more specifically, an allyl alcohol. This methodology could significantly reduce

the need for inefficient, waste generating temporary protecting groups in

carbohydrate syntheses. Future implementations that would be even more

environmentally benign could potentially be developed based on carbon monoxide as

the sacrificial reductant. Improvements of this sort might then lead to viable routes

to biomass-derived commodity oxygenated organics as well fine chemicals.

Ill – OPPORTUNITIES FOR THE FUTURE

Complexation of Cyclic and Disaccharide Sugars/ Other Metals

A key extension of our previous (L2)PtII alditol complexation studies is the

preparation of compounds derived from cyclic and disaccharide sugars, which are

more typical of primary biomass sugars such as glucose, fructose, and sucrose, as

well as being better models for the biomass polymers starch and cellulose. They are

also more complex and will provide a demanding test of complexation stereo- and

15

regio-selectivity since the a- and ~ -anomers of both pyranose and furanose ring

forms are in equilibrium via the open-chain form of the sugar:

yH20H 7H20H

~-Pyranose

$

HO H a-Pyranose

H OH

HO 1 /

H+OH ~

CH20H HOT

HO

e

0>0OH aldehydo- HO

Glucose

b

0>0

OH

fl-Furanose ‘HOH

a-Furanose

Our prior model studies with cis - and trans-cycloalkanediols show that

complexation of the cis isomer is favored over the trans isomer by a factor of seven

for cyclohexane- 1,2-diol.9 A somewhat smaller ratio (ea. 2:1) is observed for the one

sugar prototype we have examined (Z. H. Shriver), internal cis/trans competition in

methyl- cx-mamopyranoside. For cyclopentane- 1,2-diol, complexation of the cis

isomer is four times stronger than for cis -cyclohexane - 1,2-diol, while the trans

isomer is essentially unreactive, suggesting that very high selectivities should be

achievable for cis vs trans coordination in furanose sugars. For underivatized C G

sugars such as glucose, there are theoretically twelve possible isomeric a, ~-diolate

complexes that can be formed, not counting the four from the open-chain form.

Based on our previously observed trends, the higher acidity of the anomeric

hemiacetal hydroxyl protons should favor complexation involving this site. It is also

probable that complexation to the metal will significantly alter the equilibrium

16

.

ratios of the sugar isomers present, for example, increasing the furanose form over

the pyranose form by preferential binding to a cis-furanose diol unit or to the exo -

cyclic dihydroxyethyl side-chain in a C (j furanose sugar.

The common disaccharide sucrose will actually be a much simpler case since

its carbonyl group is masked as a full acetal and only the two trans- (x,&diol units in

the glucopyranose fragment would be expected to coordinate, the trans-diol unit in

the fructofuranose ring presumably being unreactive. Two products are

experimentally observed in a 3:1 ratio (M. A. Andrews). The common

polysaccharides cellulose and starch have only one unique U,&diol unit to complex,

a 2,3-trans-diol glucopyranose unit.

HOH2C

~

ct,~-Diol Units in Sucrose, Starch and Cellulose

O.O ~ ~

HO 2 trans-Pyranose1 m.wzs-Furanose I

OHo

HOH2C

w

O>OH ~os”)

nCH20H

OH 1 ?rans-Pyranose / Repeat Unit

Another essential future objective is determining the complexation and

reactivity characteristics of carbohydrates with other metals besides Ptll,

particularly those that are more redox active and those from other parts of the

periodic table. In particular, the carbohydrate coordination chemistry of early,

oxophilic transition metals should be quite different from that of late metals with

have studied to date. Lone pairs on the coordinated diolate oxygens would be

expected to participate in n-backbonding to the electron-deficient metal center,

greatly reducing their hydrogen-bond acceptor ability. This, and the decreased

polarization of the metal–oxygen bond, should significantly alter the coordination

17

selectivities. Results from both studies with more biomass relevant carbohydrates

and with other metals will be helpful in predicting and understanding the kinds of

catalytic reaction selectivities that may be achievable with native sugar substrates.

Development of Catalytic Polyol Disproportionation

We have intriguing preliminary results (G. K. Cook) which suggest that it

may be possible to effect catalytic diol deoxydehydration without addition of an

external reducing agent via a disproportionation reaction. Thus 1,2-

phenylethanediol is converted to styrene in about 60% yield in the presence of only

catalytic Cp*Re03 and a suitable “initiator”, also present in catalytic amounts. One

of the most effective initiators is the product styrene itself, hence the reaction

exhibits auto catalytic behavior. Cross-over experiments using ct-methylstyrene as

the initiator and 1,2-phenylethanediol as substrate show that the a-methylstyrene

is not consumed.

presence of styrene

these experiments,

Similarly, conversion of 1,2-hexanediol to l-hexene in the

as an initiator proceeds without oxidation of the styrene. From

we conclude that the formation of alkene from the diol must

proceed via diol disproportionation, though we have not yet identified the oxidized

diol product(s). While there is obviously still much that we don’t understand about

these reactions, development of the following type of catalytic polyol

disproportionation reaction seems potentially attainable and is certainly

stoichimetrically achievable with a suitable catalyst:

‘07-0” - -“ +2“20OH L02=+%x”

Proposed

18

This chemistry could be used to produce valuable unsaturated aldehydes and

carboxylic acids from biomass feedstocks, the glycerol coming from either fatty acid

esters or from the hydrocracking of fructose or other carbohydrates.

Development of Catalytic Carbohydrate Ionic Hydrogenolysis

Our colleagues at Brookhaven National Laboratory have developed an ionic

hydrogenation chemistry that accomplishes the stoichiometric hydrogenation of a

number of organic substrates, ranging from alkenes42 and alkynes43 to ketones, 44

acetals,45 and alcohols,4G utilizing a metal hydride and a strong acid, e.g.

(C&0JW(CO)3H + triflic acid (HOTf). For alcohols, this reaction accomplishes the

net removal of oxygen, presumably by a mechanism involving protonation of the

alcohol, loss of water, and hydride transfer to the resulting carbenium ion to

generate the hydrocarbon product:

OH1.2 CF3S03H

+

22 ‘C, 5 rninuws+ Cp(CO)3WH ~

[i,,] -*P,::)3W,0TDBullock & .%ng

The qualitative reaction rates reflect the stabilities of the intermediate carbenium

ions. Thus, deoxygenation of tertiary alcohols (above) occurs within minutes at

room temperature. Deoxygenation of secondary alcohols requires hours at room

temperature, while most primary alcohols are inert under these conditions.

Based on these results and other organometallic literature, catalytic versions

of these reactions are very plausible, as illustrated for a diol:

19

OH

\R

H2

YOxidativeAddition

LnM@

4

@LnM

,H

‘H

OHR

<

Y

OH

ProtonTransfer H20

\

Proposed R

There are a large number of cationic dihydrides / molecular dihydrogen complexes

now known,47-49 many of which have sufficient acidity to protonate substrates such

as diols. The resulting intermediate carbenium ion in this case would be stabilized

by lone-pair donation from the adjacent hydroxyl group. Subsequent hydride attack

should occur at the more positively charged internal carbon, leading to a primary

alcohol that should be stable towards further reduction. With polyols as substrates,

this reactivity would lead to cx,co-diols, which are valuable compounds in their own

right or they may be converted to other products such as tetrahydrofuran (from the

C4 alditol erythritol) or adipic acid (from CG alditols, e.g., sorbitol and mannitol,

readily available from the hydrogenation of glucose and/or fructose). Such a

synthesis of adipic acid would clearly be a much more environmentally benign route

to this product than the current route described in the introduction.

A number of problems will need to be addressed by experimentation to

develop this chemistry. The first is to show that diols and polyols will undergo this

20

type of ionic hydrogenation., Preliminary experiments (N. M. Brunkan) show that

phenylethane- 1,2-diol is reduced by (C5H5)W(CO)3H + HOTf in dichloromethane

solution at room temperature to give the expected 2-phenylethanol as the primary

product (via the intermediate complex [Cp(CO)s(HOCH2CH 2Ph)]+[OTfJ-). Styrene

epoxide, a more direct route to the proposed pronated epoxide intermediate, is

reduced to 2-phenylethanol more quickly under the same conditions. Simple diols,

such as 1,2-hexanediol, were not reduced, however, NMR spectra suggesting that

the reaction stopped at the protonated diol stage. Propylene oxide was reduced to

n-propanol, further indicating that dissociation of water from the protonated diol is

the rate inhibiting step. The second problem will be to conduct the ionic

hydrogenation in solvents that will dissolve sugars. The problem here is to avoid

excessive leveling of the acidity of the acid by protonation of the solvent. A good

choice might be acetic acid, whose conjugate acid [CH3C(OH)2]+ has a pKa = –6, ca.

four orders of magnitude more acidic than protonated alcohols. Here test

experiments indicate that the problem will be esterification of the diol or polyol.

The final step of the catalytic cycle, reformation of the metal dihydride, could also

pose a problem. Thus, initial attempts (Song and Bullock) to react metal triflates

with Hz have not been very successful, apparently because the triflate counterion

binds too strongly to the vacant metal site required for effective oxidative-addition

of hydrogen. A potential solution to this problem is to use a much more weakly

coordinating counterion such as BAr’A– (Ar’ = 3,5-bis(trifluoromethyl)phenyl)).50 In

summary, while the preliminary results for ionic hydrogenation of diols discussed

here are somewhat disappointing, they represent very limited explorations of the

wide range of parameter space available, e.g., with respect to the metal and its

associated ligands which together are known to have a strong effect on the hydricity

of the metal hydride (a range of 10s in rates of reaction with trityl cation).5 I

21

Development of Aqueous Organometallic Carbohydrate Chemistry

Given the high water volubility of sugars and water’s biocompatibility and

low cost, the development of aqueous carbohydrate homogeneous catalysis systems

is a desirable goal for industrially benign organic synthesis applications. While

most organometallic chemistry has traditionally been done in organic solvents,

water-soluble complexes are now well-known 52-54and one aqueous process utilizing

them as catalysts, propylene hydroformylation, has been commercialized. 55 For

carbohydrate transformations that rely on the carbonyl functionality of sugars, such

as our fructose retro-aldol mediated hydrocracking, we see no inherent problems

associated with conversion to an aqueous medium. For catalytic cycles based on

coordination of a carbohydrate hydroxyl functionality, however, complexation

competition problems could arise, as water contains the same functionality. The

chelate effect operative with a carbohydrate diol functionality offers some entropic

advantage over water, but whether this is sufficient for the carbohydrate to

successfully compete with water for binding to the metal center remains to be

determined. For those systems for which water proves detrimental, there are still a

number of solvents, such as 2V-methyl-2-pyrrolidinone (NMP) and dimethylsulfoxide

(DMSO), which are well-suited to metal-carbohydrate chemistry and are considered

reasonably attractive under today’s health and environmental standards.

Some hydroxyl reactions, however, such as ionic alcohol hydrogenolysis, do

not depend on coordination. Recently an aqueous, homogeneously catalyzed alcohol

carbonylation system has been reported and applied to a carbohydrate-derived

substrate, hydroxymethylfurfural (HMF) to give the new compound 5 -formylfuan-

2-acetic acid. 56 Under different reaction conditions, the catalyst is also capable of

deoxygenating the alcohol group in HMF, giving 5-methylfurfural:

22

.

H+OH

LH20H

Fructose

Pdll J P(C~H4S03- )3+Co

H+/H20

A c02

‘“’CGCH”‘3CT5’CH;heldon

This latter carbonylative reduction chemistry may be an alternative to ionic

hydrogenolysis as amethod ofdehydroxylating carbohydrates.

IV - SUMMARY

There is a critical need for new environmentally friendly processes in the

United States chemical industry as legislative and economic pressures push the

industry to zero-waste and cradle-to-grave responsibility for the products they

produce. Carbohydrates represent a plentiful, renewable

processes might economically replace fossil feedstocks.

biomass to~uels, is still not generally economical,4 the

resource, which for some

While the conversion of

selective synthesis of a

commodity or fine chemical, however, could compete effectively if appropriate

catalytic conversion systems can be found. Oxygenated organics, found in avariety

ofproducts such as nylon and polyester, are particularly attractive targets.57 We

believe that with concerted research efforts, homogeneous transition metal

catalyzed reactions could play a significant role in bringing about this future green

chemistry technology.

23

.

,-

IV - BIBLIOGRAPHY

(1) KirsChner, E. M. Chem. Z3ng.News 1996, April 8, 16-20.

(2) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis: The Applications and

Chemistry of Catalysis by Soluble Transition Metal Complexes; 2rd cd.;

Wiley: New York, 1992.

(3) Da~s, D. D.; Kemp, D. R. in Kirk-Othmer Encyclopedia of Chemical

Technology; 4th ed.;Wiley and Sons: New York, 1991; Vol. 1, pp 466-493.

(4) Bridgwater, A. V. Appl. CataL A-Gen 1994, 116,5-47.

(5) Andrews,M. A.; Klaeren, S. A. J. Am. Chem. Sot. 1989, 111,4131-4133.

(6) Andrew, M A.; Klaeren, S. A.; Gould, G. L. in Carbohydrates as Organic

Raw Materials II; Descotes, G., Ed.; VCH: New York, 1993; , pp 3-25.

(7) Cook, G. K.; Andrews, M. A. J. Am. Chem. Sot. 1996, 118, 9448-9449.

(8) Andrews, M. A; Gould, G. L. Organometallics 199310,387-389.

(9) Andrews,M A.; Voss, E. J.; Gould, G. L.; Klooster, W. T.; Koetzle, T. F. J.

Am. Chern. Sot. 1994, 116, 5730-5740.

(10) Kruse, W. M. U. S. Patent 3,935,284, Jan, 1976.

(11) Kruse, W. M.; Wright, L. W. Carbohydr. Res. 1978,64,293-296.

(12) Angyal,S. J. Adu. Carbohydr. Chem. Biochem. 1991,49,19-35.

(13) Ra.iwwal, S.; Vancheesan, S.; Rajaram, J.; Kuriacose, J. C. Indian J. Chem.

1979, 18B, 293.

(14) Pillai, S. M.; Vancheesan, S.; Rajaram, J.; Kuriacose, J. C. J. Mol. CataL

1982,16,349-358.

(15) Rajagopal,S.; Vancheesan,S.; Rajaram, J.; Kuriacose, J. C. J. Mol. CataL

1983,22, 137-144.

(16) RajagqA, S.; Vancheesan, S.; Rajaram, J.; Kuriacose, J. C. J. Mol. CataL

1983,22, 131-135.

24

(17) Wssoui, M.; Beaupere,. D.; Goethals, G.; Uzan, R. J. Mol. Catal. 1985,29,7-

12.

(18) Rajagopal, S.; Vancheesan, S.; Rajaram, J.; Kuriacose, J. C. J. Mol. CataL

1993,81,185-194.

(19) Appelt, A.; Willis, A. C.; Wild, S. B. J. Chem. Sot., Chem. Commun. 1988,

938-940.

(20) 130ugh, L.; Robinson, A. C. in 27w Carbohydrates: Chemistry and

Biochemistry; Pigman, W. W. and Horton, D., Ed.; Academic Press: New

York, 1972; Vol. IA, pp 127-138.

(21) Andrews, M. A.; Klaeren, S. A. J. Chem. Sot., Chem. Commun. 1988, 1266-

1267.

(22) Andrews, M. A.; Gould, G. L.; Klaeren, S. A. J. Org. Chem. 1989,54,5257-

5264.

(23) Binkley, R. W. Modern Carbohydrate Chemistry; Marcel Dekker: New York,

1988, pp 113-166.

(24) Andrews, M. A. Organometallics 1989,8,2703-2708.

(25) McKillip, W. J.; Sherman, E. in Kirk -Othmer Encyclopedia of Chemical

Technology; 3rd ed.;Wiley: New York, 1980; Vol. 11, pp 499-527.

(26) van Ling, G.; Ruijterman, C.; Vlugter, J. C. Carbohydr. Res. 1967,4,380-386.

(27) Hanford, W. E. U. S. Patent 2,209,055, July, 1940.

(28) Conradin, F.; Bertossa, G.; Giesen, J. U. S. Patent 3,030,429, Apr, 1962.

(29) Kasehagen, L. U. S. Patent 3,369,199, Aug, 1968.

(30) Sirkar, A. K. U. S. Patent 4,380,678, Apr, 1983.

(31) Arena, B. J. U. S. Patent 4,401,823, Aug, 1983.

(32) Arena, B. J. U. S. Patent 4,496,780, Jan, 1985.

(33) Tanikella, M. S. S. R. U. S. Patent 4,404,411, Sep, 1983.

(34) Dubeck, M.; Knapp, G. G. U. S. Patent 4,476,331, Ott, 1984.

25

.

ii

(35) Bryndza, H. E.; Calabwse, J. C.; Marsi, M.; Roe, D. C.; Tam, W.; Bercaw, J.

E. J. Am. Chem. Sot. 1986,108,4805-4813.

(36) Andrew, M. A; Gould, G. L.; Klooster, W. T.; Koenig, K. S.; Voss, E. J. Inorg.

Chem. 1996,35,5478-5483.

(37) Andrews,M. A.; Cook, G. K.; Shriver, Z. H. J. Am. Chem. Sot. Submitted.

(38) Herrmann,W. A.; Marz, D.; Herdtweck, E.; Schafer, A.; Wagner, W.;

Kneuper, H.-J. Angew, Chem., Int. Ed. EngL 1987,26,462-464.

(39) Gable, K. P. Organometallics 1994,13,2486-2488.

(40) Gable, K. P.; Juliette, J. J. J. J. Am. C%em.Sot. 1995,117,955-962.

(41) Block, E. Org. Reactions 1984,30,457-566.

(42) Bullock, R. M.; Song, J.-S. J. Am. Chem. Sot. 1994,116,8602-8612.

(43) LUan, L.; Song, J.-S.; Bullock, R. M. J. Org. Chem. 1995,60,7170-7176.

(44) Song, J.-S.; Szalda, D. J.; Bullock, R. M.; Lawrie, C. J. C.; Rodkin, M. A.;

Norton, J. R. Angew. Chem., Int. Ed. EngL 1992,31, 1233-1235.

(45) Song, J.-S.; Szalda, D. J.; Bullock, R. M. J. Am. Chem. Sot. 1996,118, 11134-

11141.

(46) Song, J.-S.; Bullock, R. M. unpublished results.

(47) Heinekey, D. M.; Oldham, W. J., Jr. Chem. Rev. 1993,93,913-926.

(48) Jessop, P. G.; Morris, R. 1% Coord. Chem. Rev. 1992,121, 155-284.

(49) Kubas, G. J. Ace. Chem. Res. 1988,21, 120-128.

(50) Brookhart, M.; Grant, B.; Volpe, A. F., Jr. Organometallics 1992,11,3920-

3922.

(5U Cheng, T.-Y.; Bullock, R. M. Organometallics 1995,14,4031-4033.

(52) ROundhiLD. M. Adv. Organometal. Chem. 1995,38,155-188.

(53) Kalck, P.; Monteil, F. Adv. Organometal. Chem. 1992,34,219-284.

(54) Barton, M.; Atwood, J. D. J. Coord. Chem. 199L24,43-67.

(55) Haggin, J. Chem. Eng. News 1994, Oct. 10,28-36.

26

(56)

(57)

Papadogianakis, G.; Maat, L.; Sheldon, R. A. J. Chem. Sot., Chem. Commun.

1994,2659-2660.

Busche, R. M. Biotech. Prog. 1985,1, 165-179.

Acknowledgments. The author would like to thank his co-workers for their

major contributions, experimental and intellectual, to the studies discussed in this

Chapter: Nicole Brunkan, Gerald Cook, George Gould, Stephen Klaeren, Wim

Klooster, Kristina Koenig, Santosh Mandal, Zachary Shriver, and Eric Voss. This

research was carried out at Brookhaven National Laboratory under contract DE-

AC02-76CHOO016 with the U.S. Department of Energy and supported by its

Division of Chemical Sciences, OffIce of Basic Energy Research.

27