Liquid-Phase Dehydration of Lactic Acid for the Production of Bio-Acrylic Acid Development of a Multi-Step Process

Liquid-phase dehydration of lactic acid for the production of bio-acrylic acid Development of a multi-step process

Flüssigphasen-Dehydratisierung von Milchsäure zur Produktion von biobasierter Acrylsäure Entwicklung eines mehrstufigen Prozesses

Der Technischen Fakultät der Friedrich-Alexander-Universität Erlangen-Nürnberg zur Erlangung des Grades

DOKTOR-INGENIEUR

vorgelegt von

Matthias Kehrer, M. Sc. aus Nürnberg

Als Dissertation genehmigt von der Technischen Fakultät der Friedrich-Alexander-Universität Erlangen-Nürnberg.

Tag der mündlichen Prüfung: 17. Dezember 2018

Vorsitzende/r des Promotionsorgans: Prof. Dr.-Ing. Reinhard Lerch

Gutachter/in: Prof. Dr. Peter Wasserscheid Prof. Dr. Nicolas Vogel

To my family

“Obstacles don’t have to stop you. If you run into a wall, don’t turn around and give up. Figure out how to climb it, go through it, or walk around it.”

Michael Jordan

Preface i

Preface

The present work was carried out in the period from May 2014 to December 2017 at the Institute of Chemical Reaction Engineering of the Friedrich-Alexander-University Er- langen-Nürnberg, headed by Prof. Dr. Peter Wasserscheid. The results presented herein where achieved within the scope of the research project “Liquid-phase dehydration of lactic acid obtained via fermentation for the production of bio-acrylic acid”, which was carried out in close collaboration with the industrial partner Procter & Gamble.

At this point, I would like to dedicate personal thanks to a large number of people and their various contributions, involved in the development of this work:

First and foremost, I would like to thank my supervisor, Prof. Dr. Peter Wasserscheid, for giving me the opportunity to be part of this interesting and exciting project and for the excellent mentoring during the entire period of this thesis. Dear Peter, I really appreciate your challenges, your advice and your trust in my work –as a result I have developed myself on a personal and professional level.

Moreover, I would like express my thanks to Prof. Dr. Nicolas Vogel for kindly accept- ing the second review of this thesis and also to the other members of the PhD com- mittee.

I am very thankful to my research group head, Dr. Jakob Albert, for the valuable support, exciting discussions and the useful tips from the engineering perspective. In particular, I would like to thank the whole project team, namely Jens, Julian and Nicola for the energetic working atmosphere and the unlimited team spirit. Working with you was great and I enjoyed it tremendously!

I am also very grateful for the excellent scientific and financial support from our industrial collaboration partner P&G. Special thanks goes to Dr. Peter Dziezok and Dr.

Dimitris Collias for their inexorable interest and support, the fruitful discussions and the interesting insights into the work environment of P&G. In addition, financial sup- ii port by the Bundesministerium für Ernährung und Landwirtschaft through the Facha- gentur Nachwachsende Rohstoffe e.V. (FNR, Project No 22009614) is gratefully acknowl- edged.

I would further like to thank all students I had the privilege to get to know or even supervise during my time at the CRT. I especially thank Vera, Katrin, Julian and Simon for their support and the pleasant and friendly collaborations. I wish you all the best of luck in your future endeavors and scientific career.

Moreover, I would also like to extend my sincere thanks to all the people who have done the important and indispensable work in the background: Dear Mrs. Menuet,

Mrs. Singer and Mrs. Bittan, thank you for your support concerning all administrative tasks and for the entertaining conversations. Dear Michel, Achim, Julian, Sascha and

Alex, thank you for your help concerning technical and especially electrical tasks and the related and funny “sideshow”.

Furthermore, I like to dedicate a special mention to the motivating, professional and personal atmosphere in the CRT group and especially in the “Green Couch Office”. It was a great time and I will miss all the diverse activities and the nice and helpful colleagues I met here over the past few years.

The most important part of my acknowledgement is addressed to my whole family.

Thank you for your unlimited support and the spirit of unity! My greatest “Thank you” goes to my beloved wife Varina for her motivation, her patience and her love, which is beyond comparison! You are the best thing that ever happened to me!

Last but not least, I like to thank all others whose names are not mentioned here but who (in)directly contributed to the success of this PhD thesis. I really enjoyed the last

3.5 years at the CRT. All the (work) experiences I gained, the “friends for life” I made and the daily adventures are something I will remember and benefit from for the rest of my life!

Publications, contributions and supervised projects iii

Publications, contributions and supervised projects

Parts of this work have been previously published or presented.

Publications in scientific journals

M. Kehrer, J. Mehler, N. Taccardi, J. Nagengast, J. Kadar, D. Collias, P. Dziezok, P.

Wasserscheid and J. Albert, ChemSusChem, 2018, 11 (6), 1063-1072

J. Nagengast, S. Hahn, N. Taccardi, M. Kehrer, J. Kadar, D. Collias, P. Dziezok, P.

Wasserscheid, J. Albert, ChemSusChem, 2018, Accepted Author Manuscript. doi:10.1002/cssc.201800914

Conference contributions

M. Kehrer, J. Albert, P. Wasserscheid, D. Collias, oral presentation with the title “Novel liquid-phase technology for the production of bio-acrylic acid”, 2017,

Jahrestreffen Deutscher Katalytiker, Weimar, Germany

M. Kehrer, J. Albert, P. Wasserscheid, D. Collias, poster presentation with the title

„Liquid-phase dehydration of lactic acid towards bio-acrylic acid“, 2017, International

Symposium on Green Chemistry (ISGC), La Rochelle, France

Patents

J. Albert, P. Wasserscheid, N. Taccardi, J. Nagengast, M. Kehrer, J. Kadar, D. I. Collias,

Patent WO 2018022828, 2018

J. Albert, P. Wasserscheid, N. Taccardi, J. Nagengast, M. Kehrer, J. Kadar, D. I. Collias,

Patent WO 2018022826, 2018 iv

J. Albert, P. Wasserscheid, N. Taccardi, J. Nagengast, M. Kehrer, J. Kadar, D. I. Collias,

14739P, US, 2017, Patent pending.

Student Theses

Within the frame of the research work at the Friedrich-Alexander-University Erlan- gen-Nürnberg, the below listed student projects have been technically and theoretically supervised by the author of this dissertation. The following student theses are not necessarily published and may only be available from the university’s library

(FAU Erlangen-Nürnberg). Included raw data sets from supervised student work or references to the latter are marked with the respective number of the following direc- tory.

[A] K. Huber, “Catalytic esterification of fermentatively-derived lactic acid to various alpha-acyloxy derivatives”, bachelor thesis, 2015

[B] V. Haagen, “Alpha-acyloxy derivatives of fermentatively-derived lactic acid:

Synthesis, characterization and suitability as intermediate towards bio-acrylic acid”, bachelor thesis, 2015

[C] J. Mehler, “Synthesis of 2-bromopropionic acid from fermentatively-derived lactic acid and its derivatives”, master thesis, 2017

Abstract v

Abstract

This thesis deals with the development of a novel liquid-phase dehydration process of lactic acid (LA) for the production of bio-acrylic acid (AA). The approach of dehydrating LA in the liquid phase is expected to offer advantages over so far existing gas-phase dehydration processes and may therefore provide a technical as well economically feasible and sustainable alternative to the incumbent petro-based route.

The first part of this work provides a general introduction of the developed “NADA” technology, enabling the production of AA in a liquid phase process. The one-step and HBr-triggered dehydration of LA was realized in acidic bromide ionic liquids.

Subsequently, current limitations of the process using LA as substrate were addressed and a multi-step processing of the technology was conceptualized, based on the postulated reaction mechanism of the liquid-phase dehydration system.

In the main part of this work, a multi-step liquid-phase dehydration process for the production of bio-AA from LA, proceeding via brominated LA-species, was developed. The feasibility of splitting the “NADA” process into a spatially separated reaction sequence of the postulated mechanism was proven. Three individual reactions, namely bromination of LA, isomerization of 2- to 3-bromopropionic acid and dehydrobromination of 3-bromopropionic acid to AA were developed and largely optimized, respectively. In this context, bromide ionic liquids played a crucial role as reaction medium and/or reactant. After evaluation of the multi-step approach from LA towards bio-AA via brominated LA-species, possibilities for potential process shortcuts were finally addressed.

The presented results in this thesis contribute to the understanding of AA production from LA via brominated LA-species using (acidic) bromide ILs. The developed

“NADA” technology is a novel, selective and sustainable way for the production of

AA from LA, contributing to a future economic production of bio-AA.

Kurzzusammenfassung

Diese Arbeit befasst sich mit der Flüssigphasen-Dehydratisierung von Milchsäure

(LA) zur Produktion von biobasierter Acrylsäure (AA). Ein solcher und neuartiger Ansatz der Flüssigphasen-Reaktion bietet Vorteile gegenüber bestehender Gaspha- sen-Dehydratisierungsprozesse und kann damit als eine sowohl technisch wie auch

ökonomisch attraktive und nachhaltige Alternative zum bestehenden und auf fossi- len Rohstoffen basierten Produktionsprozess von AA angesehen werden.

Der erste Teil dieser Arbeit stellt eine allgemeine Einführung in die entwickelte

“NADA”-Technologie dar, welche eine Produktion von AA aus LA in der Flüssig- phase ermöglicht. Die einstufige und durch HBr ausgelöste Dehydratisierung von LA wurde in sauren Bromid-Ionischen Flüssigkeiten (ILs) realisiert. Anschließend wurden die aktuellen Limitierungen des Prozesses basierend auf dem Substrat LA her- ausgearbeitet und ein mehrstufiges Verfahren entworfen, wofür v.a. der postulierte Reaktionsmechanismus des Dehydratisierungssystems herangezogen wurde.

Im Hauptteil dieser Doktorarbeit wurde die mehrstufige Flüssigphasen-Dehydrati- sierung von LA zu bio-AA entwickelt, welche über bromierte LA-Derivate verläuft.

Innerhalb dieser Studie konnte die Machbarkeit der Auftrennung des „NADA“-Pro- zesses in eine räumlich getrennte Reaktionssequenz demonstriert werden. Genauer gesagt wurden die Bromierung von LA, die Isomerisierung von 2- zu 3-Brompropi- onsäure, sowie die Dehydrobromierung von 3-Brompropionsäure zu Acrylsäure, entwickelt und größtenteils optimiert. Im Zusammenhang mit der Entwicklung der ein- zelnen Reaktionsschritte spielten Bromid-ILs eine besonders wichtige Rolle. Sie wurden sowohl als Reaktionsmittel als auch als Reaktionspartner eingesetzt. Nach einer

Beurteilung der entwickelten und mehrstufigen Prozessführung der “NADA”-Tech- nologie wurden abschließend mögliche Optimierungen hinsichtlich einer zweistufi- gen Prozessführung untersucht.

Kurzzusammenfassung vii

Die vorliegende Arbeit trägt entscheidend zum Verständnis der Dehydratisierung von LA zu AA in (sauren) Bromid-ILs mithilfe von bromierten LA-Spezies bei. Die entwickelte „NADA“-Technologie stellt eine neuartige, selektive und nachhaltige

Möglichkeit zur Dehydratisierung von LA dar und liefert damit einen wichtigen Bei- trag zu einer künftigen ökonomischen Herstellung von bio-AA. viii

Contents

Preface ...... i

Publications, contributions and supervised projects ...... iii

Abstract ...... v

Kurzzusammenfassung ...... vi

Contents ...... viii

Nomenclature ...... xii

1 Introduction ...... 1

2 Fundamentals ...... 4 2.1 Acrylic acid ...... 4 2.1.2 State-of-the-art acrylic acid production ...... 6 2.1.3 Acrylic acid production from bio-renewables ...... 10 2.1.3.1 Production of bio-based propene ...... 11 2.1.3.2 Production of bio-based acrolein ...... 12 2.1.3.3 AA production from glycerol ...... 13 2.1.3.4 AA production from acrylonitrile ...... 15 2.1.3.5 AA production from 3-hydroxypropionic acid ...... 15 2.1.3.6 Further routes to bio-AA ...... 16 2.2 Lactic acid as substrate for bio-AA production ...... 18 2.2.1 State-of-the-art lactic acid production ...... 20 2.2.2 Lactic acid reactivity and potential products ...... 23 2.2.3 Dehydration of lactic acid to acrylic acid ...... 28 2.2.3.1 Gas-phase dehydration of lactic acid ...... 29 2.2.3.2 Dehydration of LA in near- and supercritical water ...... 32 2.2.3.3 Liquid-phase dehydration of lactic acid...... 33 2.3 Bromopropionic acids ...... 37 2.3.1 Conventional production of 2-bromopropionic acid...... 38 2.3.2 Bromination of alcohol functionalities...... 39 2.3.2.1 Basics of nucleophilic substitution of alcohols ...... 39 2.3.2.2 Synthetic methods for bromination of LA ...... 41 2.3.3 The role of 3-bromopropionic acid en route to bio-AA ...... 45 2.4 Ionic liquids and their role in bio-AA production from LA ...... 48 2.4.1 ILs as reaction medium for liquid-phase dehydration of LA: Tetrabutylphosphonium bromide ...... 49 2.4.2 Ionic liquids as reaction medium for bromination of LA ...... 52 Contents ix

2.5 Objective of this thesis ...... 55

3 Experimental section ...... 57 3.1 Experimental procedures ...... 57 3.1.1 Benchmark one-step dehydration of LA in the NADA molten salt reaction matrix ...... 57 3.1.2 Bromination reactions in aqueous reaction medium ...... 58

3.1.2.1 Bromination with HBr(aq) ...... 58 3.1.2.2 Extractor screening for aqueous bromination medium ..... 58 3.1.2.3 Bromination in biphasic reaction medium / In-situ LLE .... 59 3.1.2.4 Bromination in aqueous reaction medium with ionic liquid additive ...... 59 3.1.3 Bromination with zwitterionic HBr carriers...... 60 3.1.3.1 Preparation of zwitterions ...... 60 3.1.3.2 Loading of zwitterions with HBr ...... 60 3.1.3.3 Bromination with HBr-loaded ionic liquids ...... 61 3.1.3.4 LLE screening experiments for recycling of zwitterions and 2-BrPA isolation ...... 62 3.1.3.5 Recycling of zwitterions ...... 63 3.1.4 Isomerization reaction in ionic liquid reaction matrix ...... 63 3.1.5 3-BrPA conversion in TOA ...... 65 3.1.6 Thermally-induced HBr recovery from [TOAH]Br ...... 65 3.1.7 HBr recovery from [TOAH]Br via extraction ...... 66 3.1.8 Semi-batchwise dehydrobromination of 2-BrPA to AA ...... 66 3.2 Analysis, quantification and calculations...... 69 3.2.1 Characterization and quantification of liquid phase samples with NMR spectroscopy ...... 69 3.2.2 Characterization of biphasic reaction systems ...... 70 3.2.3. Characterization of HBr-loaded ionic liquids ...... 70 3.2.4 Calculation of HBr amount from measured pH value ...... 71 3.2.5 Off-line analysis of gaseous products ...... 71

4 Results and Discussion ...... 72 4.1 The Erlanger “NADA” process: A Nucleophile assisted dehydration to acrylates ...... 72 4.1.1 The NADA technology ...... 73 4.1.2 Benchmarking of the one-step NADA process with LA feedstock ...... 76 4.1.3 Motivation for a multi-step NADA process ...... 79 4.2 Development and optimization of a multi-step liquid-phase NADA process for the production of bio-acrylic acid ...... 82 x

4.2.1 Synthesis of 2-bromopropionic acid from lactic acid and lactide 82 4.2.1.1 Lactic acid bromination with aqueous hydrobromic acid . 83 4.2.1.2 Parameter variation of aqueous bromination medium ...... 86 4.2.1.2 Concepts to increase the performance of the aqueous bromination medium ...... 94 4.2.1.3 Zwitterionic HBr carriers for the synthesis of 2-BrPA from lactide and LA ...... 105 4.2.2 Isomerization of 2-BrPA to 3-BrPA in bromide ILs ...... 121

4.2.2.1 Isomerization of 2-BrPA to 3-BrPA in [PBu4]Br ...... 121 4.2.2.2 Variation of the IL isomerization matrix – the role of the bromide ...... 128 4.2.2.3 Cation variation of the IL used as isomerization matrix .. 130 4.2.2.4 Product isolation from isomerization matrix via distillation ...... 132 4.2.2.5 Summary of 2-BrPA isomerization study ...... 134 4.2.3 Dehydrobromination of 3-BrPA to AA...... 135 4.2.3.1 3-BrPA conversion in trioctylamine...... 136 4.2.3.2 HBr recovery from [TOAH]Br ...... 141 4.2.4 Evaluation of the multi-step NADA process ...... 143 4.3 NADA process shortcuts ...... 146 4.3.1 Dehydrobromination of 2-BrPA to AA in basic environment 146 4.3.2 Dehydrobromination of 2-BrPA to AA in bromide ILs ...... 149 4.3.2.1 Batchwise dehydrobromination of 2-BrPA to AA in bromide ILs ...... 150 4.3.2.2 Semi-batchwise dehydrobromination of 2-BrPA to AA in bromide ILs ...... 151 4.3.3 Summary of process shortcuts ...... 155

5 Summary ...... 157

6 Bibliography ...... 162

7 Appendix...... 172 7.1 Further data and information concerning LA bromination ...... 172 7.1.1 Summary of LA bromination data ...... 172 7.1.2 Bromide salt and precursor synthesis ...... 175 7.1.3 Error estimation of standard LA bromination with [MIMBS]Br 193 7.1.4 Synthesis of a [MIMBS]Br bromination agent with low water content and high HBr content ...... 194 xi

7.2 Further data and information concerning 2-BrPA isomerization to 3-BrPA ...... 195 7.2.1 Summary of isomerization data ...... 196 7.3 Further data and information concerning 2-BrPA dehydrobromination ...... 199 7.4 List of Figures ...... 200 7.5 List of Schemes ...... 205 7.6 List of Tables ...... 207

xii

Nomenclature

Letters and indices

0 initial (at t = 0) - c concentration mol/L g gaseous - i undefined species / compound - K partition coefficient - m mass g, kg M Molar mass g/mol n Molar amount mmol, mol p pressure mbar, bar S selectivity % t time (or at t = t) min, h

tr reaction time min, h T temperature °C

Tb boiling point °C

Tm melting point °C

Tr reaction temperature °C V volume mL, L X conversion % Y yield % δ chemical shift ppm ν stoichiometric coefficient -

Abbreviations

2-APA 2-acetoxypropionic acid 2-BrPA 2-bromopropionic acid 3-BrPA 3-bromopropionic acid 3-CNPA 3-cyanopropionic acid 3D three-dimensional 3-HP 3-hydroxypropinic acid 3-HPA 3-hydroxypropionaldehyde AA acrylic acid AcH acetaldehyde Nomenclature xiii

AcOH acetic acid aq aqueous / aqueous phase CRT Institute of Chemical Reaction Engineering (C)STR (continuously) stirred tank reactor DES deep eutectic solvent DiAA diacrylic acid E1 unimolecular elimination reaction mechanism E2 bimolecular elimination reaction mechanism EAPA 2-acetoxypropionate El electrophilic eq equivalent

Et(30) Solvent polarity scale / kcal mol-1 (molar transition energies of the standard betaine no. 30)

ETN dimensionless normalized solvent polarity scale FAU Friedrich-Alexander-University Erlangen-Nürnberg GHG greenhouse gases GC gas chromatography HBD hydrogen bond donor HCA halogenated carboxylic acid IL ionic liquid IS internal standard

L2A lactoyllactic acid LA lactic acid LAB lactic acid bacteria LCA life cycle assessment LHSV liquid hourly space velocity / h-1 LLE liquid-liquid extraction MAA methacrylic acid MEHQ hydroquinone monomethyl ether MSA methanesulfonic acid NADA Nucleophile Assisted Dehydration to Acrylates NMR nuclear magnetic resonance spectroscopy NREU non-renewable energy use Nu nucleophilic Org organic phase P&G The Procter & Gamble Company PA propionic acid PAA polyacrylic acid PLA polylactic acid PTZ phenothiazine xiv

ppm parts per million rpm rounds per minute STY space time yield

SN1 unimolecular nucleophilic substitution

SN2 bimolecular nucleophilic substitution TCD thermal conductivity detector TOA trioctylamine WHSV weight hourly space velocity / h-1

Abbreviations of ionic liquids

[1Bu4MePyr]Br 1-butyl-4-methyl pyridinium bromide [1Et1MePyrro]Br 1-ethyl-1-methyl-pyrrolidinium bromide [BIMBS]Br 1-(4-butanesulfonic acid)-3-butylimidazolium bromide [BMIM]Hal 1-n-butyl-3-methyl imidazolium halide

[BMIM]HSO4 1-n-butyl-3-methyl imidazolium hydrogen sulfate

[BMIM]HSO4 1-n-butyl-3-methyl imidazolium hydroxide

[BMIM]2SO4 di(1-n-butyl-3-methylimidazolium) sulfate [EIMBS]Br 1-(4-butanesulfonic acid)-3-ethylimidazolium bromide [EMIM]Br 1-ethyl-3-methylimidazolium bromide

[Et2NSF2]BF4 diethylaminodifluorosulfinium tetrafluoroborate

[Et4N]BF4 tetraethylammonium tetrafluoroborate

[Et4N]Br tetraethylammonium bromide

[Et4N]Hal tetraethylammonium halide [MIMBS]Br 1-(4-butanesulfonic acid)-3-methylimidazolium bromide [MIMPS]Br 1-(4-propanesulfonic acid)-3-methylimidazolium bromide

[NBu4]Br tetrabutylammonium bromide [OIMBS] 1-(4-butanesulfonic acid)-3-octylimidazolium bromide

[PBu3BS]Br tri-n-butyl-(4-butanesulfonic acid) phosphonium bromide

[PBu4]Br tetrabutylphosphonium bromide

[PBu4]Cl tetrabutylphosphonium chloride

[PBu4]NTf2 tetrabutylphosphonium bis(trifluoromethyl) sulfonylimide

[PBu4]OH tetrabutylphosphonium hydroxide [PMIM]Br 1-n-pentyl-3-methylimidazolium bromide [PyrBS]Br 1-(4-butanesulfonic acid) pyridinium bromide [TOAH]Br trioctylammonium bromide

1 Introduction 1

1 Introduction

“Charting a course to a more sustainable future”[1]

Today, sustainable and more environmentally-friendly chemical processing is gaining both commercial and societal interest. Indeed, the bio-renewable production of bulk chemicals is a hot topic in academia and industry and of high importance to overcome the dependency on fossil feedstocks and to reduce the emission of greenhouse gases (GHGs). However, moving to a sustainable society requires to meet high technological and economical challenges.[2]

This is precisely why large multinational companies have established specific sustainability goals towards their environmental management (e.g. climate, waste and water) to demonstrate their progress on the way to a sustainable future. The slogan

“Charting a course to a more sustainable future” heads the sustainability goals of one of the largest consumer goods companies worldwide, namely The Procter & Gamble

Company (P&G). To be more specific, the header reflects the company’s commitment to operating sustainably, driven by P&G’s long-term vision of “producing products that consumers love while maximizing the conservation of resources, powering all its plants with 100% renewable energy, using 100% renewable or recycled materials for all products and packaging and having zero consumer and manufacturing waste go to landfill.” One of the set and ambitious targets, for instance, is to “create technologies by 2020 to substitute top petroleum-derived raw materials with renewable materials […].”[1]

In this thesis and the related research project, which were both run in close collaboration with P&G, a renewable and large-scale production process of bio-acrylic acid

(bio-AA) was targeted. AA is a versatile bulk chemical and plays an extremely important role for the production of acrylate polymers (worldwide production > 5 million tons per year). Its primary use is in the paint, construction and superabsorbent industries, such as products used in diapers and other hygiene products.[2a, 3] 2

Currently, commercial acrylic acid is predominately manufactured by a catalytic gas- phase oxidation of petroleum-based propene. Due to enhanced process control and selectivity, a two-step oxidation via acrolein as intermediate has been established as the method of choice. The exothermic reactions are usually conducted in fixed-bed tubular reactors using molybdenum-based heterogeneous catalysts and typically characterized by high reaction temperatures. [4]

In principle, there are two conceivable solutions for AA production from bio-renewables. One is the adaption of the already existing, established and petroleum-based route by replacing the fossil-based feedstock with green propene or green acrolein.

The second one is the development of a novel process from alternative and bio-based materials. Relevant substrates for a sustainable production of AA are e.g. glycerol and

3-hydroxypropionic acid (3-HP). Currently, the most promising substrate for the production of bio-AA is lactic acid (LA), which has already become an essential platform chemical in the bio-based economy (variety of applications and their still growing market, e.g. bio-based and bio-degradable polymer polylactic acid). Over 90 % of the current LA production is based on bacterial fermentation of carbohydrates.[2a, 5]

LA can be converted to AA by dehydration. Since the first approach of gas-phase dehydration was reported in 1958, several groups have investigated the catalytic system and process operation of LA dehydration.[6] The so far highest yields of AA obtained from gas-phase dehydration of LA and its derivatives were reported by

Godlewski et al. and Velasquez et al. and were patented by P&G.[7] The experiments were conducted at reaction temperatures of approximately 375 °C and 25 bar with diluted and aqueous LA solution (20 wt%). The reported yields and the related selectivity towards AA were in the range of 90 %. Further approaches were based on dehydration of LA in near[8]- and supercritical[9] water or liquid-phase dehydration of LA[10], which is the by far least investigated approach of LA conversion to AA. However, liquid-phase dehydration of LA seems to be a promising route to bio-AA, which may overcome the known limitations of LA dehydration under harsh reaction conditions.

While the research of this thesis was ongoing, Terrade et al. described in 2017 a novel method for the production of bio-AA in the liquid phase, namely the catalytic cracking of lactide in bromide salts and in the presence of strong acids.[10b] The postulated 3

mechanism proceeds via brominated LA-species, which are in-situ converted to AA over the course of the reaction. At optimized conditions, AA yields of up to 58 % were obtained.

After more than half a century of research, LA dehydration to bio-AA has not yet been commercialized and to use the technology on a large scale, still further improvement is necessary.

In the context of this work and the developed multi-step liquid-phase dehydration process of LA to bio-AA, brominated LA-species are of particular importance. Cur- rently, brominated LA derivatives, namely 2- and 3-bromopropionic acid (2- and

3-BrPA) are manufactured from non-renewable substrates.[11] Thus, integration of these LA-species (2- and 3-BrPA) into a sustainable production route for AA includes the development of several reactions, like bromination of bio-based LA and dehydrobromination of 2- and 3-BrPA to AA. Due to their well-known and interesting properties, ionic liquids (ILs) offer immense opportunities as e.g. catalysts and novel solvents for chemical transformations.[12] Here, especially bromide ILs were used as reaction medium and/or reactant for the developed and investigated chemical reactions, e.g. dehydration of LA to AA, bromination of LA to 2-BrPA or isomerization of 2-BrPA to 3-BrPA.

The aim of this thesis is to develop a novel multi-step and liquid-phase dehydration process for bio-AA staring from biogenic LA. The developed liquid-phase technology is expected to offer economic and ecologic advantages over the gas-phase dehydration of LA or its derivatives because of lower temperature and pressure. Simultane- ously, the targeted establishing of a carbohydrate-based technology may address the pressure on the propene supply.

2 Fundamentals

The general chapter of this thesis is basically separated into two parts. In the first part, acrylic acid and lactic acid are introduced, representing the two protagonists of this thesis (chapter 2.1 and 2.2). After providing a general state of knowledge and technology, the focus is on the history of LA dehydration, to give an impression of the current status of this research area. The second part provides the basics of two essential building blocks of this work, namely the chemistry of bromopropionic acids and ionic liquids (ILs) and the related reactions and application possibilities, respectively

(chapters 2.3 and 2.4). Finally, the fundamental section is completed with the objective of this thesis (chapter 2.5).

2.1 Acrylic acid

Acrylic acid (AA) is the simplest olefinic carboxylic acid. High reactivity is created by the presence of two functional groups within the �, �-unsaturated carboxylic acid.

Accordingly, the free acid and their esters as well as the related methacrylic acid- based compounds (MAA) have a wide application potential. This group of substances is generally summarized as so-called “acrylates”. The carboxylic group can be either used for the production of acrylic esters or for the synthesis of different acrylamides and acrylic chlorides by common methods. The vinyl functionality of AA can readily undergo addition and polymerization reactions to form e.g. polyacrylic acid (PAA).

The polymerization and esterification reaction pathway is illustrated in Scheme 1.

This makes AA a versatile monomer and intermediate for a vast array of applications.

Particular note should be taken on the use of PAA in the cosmetic industry. Here, for instance, these so-called “superabsorbent polymers” are used as moisture absorbing material in disposable diapers (55 % of total AA in 2014). The remaining end-use market share of acrylates and PAA is in plastics, synthetic rubber, surface coatings, seal- ants, and as adhesives in paints.[2a, 3] 2 Fundamentals 5

The described spectrum of applications makes AA an essential raw material for industrial and consumer products. The production capacity of AA has exceeded 5 million tons per year and is still of rising tendency, placing AA among the top 100 chemicals produced worldwide. The global AA market (> $ 8 ��) is assumed to still increase with a forecast growth of over 5 % from 2016 to 2023. The current market price for AA is $1600-$2200 per ton, depending on the product purity (crude- and glacial-grade AA).[2a, 13]

Scheme 1: Polymerization and esterification as main product pathways starting from the AA monomer, according to Beerthuis et al.[2a]

In contrast to PAA, which is typically used as white and fluffy powder, AA is a colorless liquid at ambient conditions (Tm = 13°C, Tb = 141 °C) and further characterized as hygroscopic and acidic (pKa = 4.26). It is miscible with water and freely soluble in most organic solvents. However, intermolecular reactions of AA caused by high reactivity of the monomer can lead to spontaneous and exothermic polymerization, dis- playing the major challenge in AA production and handling (especially at elevated temperatures). Additionally, even at ambient temperatures, the formation of di-acrylic acid (DiAA) is observed over long storage periods. AA dimerization proceeds via Michael-type addition, where the carboxylate anion functions as Michael do- nor and the free acid as Michael acceptor. Therefore, commercially available AA is stabilized by polymerization inhibitors (approx. 200 ppm) like phenothiazine (PTZ) 6 or hydroquinone monomethyl ether (MEHQ). Monomer stabilization as well as prevention of polymerization, which is connected with tremendous heat and pressure generation as well as solidification of the material, plays a crucial role within AA production, downstream processing and storage.[4a, 14] Additionally, the high reactivity of

AA may cause further (consecutive) reactions, which need to be prevented to achieve high AA yields and efficient product isolation. Decarbonylation of AA to ethene

(C2H4) and carbon dioxide (CO2), for instance, is relevant in less acidic or even basic environment. Furthermore, in the presence of free hydrogen, AA is known to readily undergo hydrogenation to propionic acid (PA).[9b]

2.1.2 State-of-the-art acrylic acid production

Since the first industrial AA manufacturing was put into operation in the beginning of the 20th century (Röhm & Haas), different commercially practiced technologies were developed. In principle, there are several routes to AA proceeding on the basis of various starting materials (e.g. acetonitrile, ethylene cyanohydrin or �-propio- lactone). Worth mentioning is the Reppe process, which played a crucial role in worldwide AA production until the late 1960s. Here, acetylene reacts with water or an alcohol and carbon monoxide (CO) in the presence of a nickel carbonyl catalyst to directly produce acrylic acid or alkyl acrylates. The carbonylation of acetylene is now obsolete but accounted for almost one-half of the total acrylate production in the US in 1976. However, another route has become more and more important, which led to discontinuation of all traditional AA production processes in recent years.[4d, 4e]

Today’s commercial acrylic acid is exclusively produced by the oxidation of petro- based propene. Due to enhanced process control and selectivity, a two-step oxidation via acrolein as isolable intermediate has developed into the method of choice in most manufacturing facilities (see Scheme 2).[4d, 4e] 2 Fundamentals 7

Scheme 2: Two-step propene oxidation as reaction pathway for commercial production of AA.

In the preferred AA production route, optimum temperature and catalyst are different for each of the two steps. The first oxidation step of propene to acrolein requires a higher temperature (320-370 °C) than the second one (220-300 °C). Propene oxidation to acrolein is typically performed in the presence of steam and air at pressures between 1-2 bar. The exothermic reaction is usually conducted in a fixed-bed multi- tubular reactor (up to 22000 tubes per reactor). A typical molybdenum-based catalyst for the first stage of the oxidation sequence consists of molybdenum and bismuth as main components. Optional and varying elements are e.g. phosphorous, iron, tungsten, silicon and potassium. The second stage, namely acrolein oxidation to AA, is conducted in a similar reactor setup packed with a different catalyst and run at much lower reaction temperature (220-300 °C). Here, a molybdenum-based catalyst including optional components like vanadium, tungsten, iron, nickel, manganese or copper may be used. Accordingly, the catalysts of both stages are mixed oxides, typically supported on low surface areas and comprising molybdenum as main component.

Numerous companies like e.g. Nippon Shokubai[15], BASF[16] or Ube Industries[17], have developed two-step processes with optimal catalyst composition and process varia- bles. The processes are generally characterized by a moderate to high propene and acrolein conversion (70-97 %) and high product selectivity (85-99 %). Typically observed by-products of the two-step propene oxidation are acetic acid (AcOH), PA, acetaldehyde (AcH), acetone, CO and CO2. In the latter two cases, by-product formation results from further deep-oxidation of acrolein or acrylic acid. Due to the large amount of added water, the resulting AA product solution is approximately

20-25 wt% in water. The addition of water fulfills different tasks within the AA production process: Shifting the explosion limit of the process, improving desorption from the catalyst, facilitating the heat removal and lowering the AA partial pressure. 8

All these points are fundamental for the efficient and safe process operation, which is described in the following (see Figure 1).[4]

In the beginning of the gas-phase process, propene is mixed with air and steam and fed into the two-step oxidation reaction system. After passing the first oxidation reactor (1), which is operated at the above described conditions (320-370 °C, 1-2 bar,

Mo-based catalyst), the reaction products (main product is acrolein) are directly fed into the second oxidation reactor (2) for further oxidation to AA (220-300 °C, 1-2 bar,

Mo-based catalyst). The gaseous effluent from the second stage oxidation is fed to the bottom of the quench tower (3). Here, the product gas stream is quenched in water or aqueous AA solution and cooled to approximately 80 °C. A part of the residual gas obtained at the top of the absorber unit is incinerated, with the balance being recycled to the beginning of the oxidation sequence. The aqueous effluent (20-30 wt% AA) from the bottom of the absorber is sent to the downstream section for product recovery. AA is extracted from the absorber effluent by solvent extraction (4) and subsequently (vacuum) distilled in a solvent recovery column (5). Standard solvents for

AA extraction are e.g. butyl acetate, xylene or diisobutyl ketone but also approaches using high-boiling solvents like for instance a mixture of diphenyl and diphenyl ether are described in the literature. Columns are usually operated at low-bottom temperatures, short residence time and with the addition of polymerization inhibitors to minimize the formation of DiAA and PAA. AA is obtained in different product puri- ties from one (7) or more (10) further distillation steps. Alternatively, crude AA can be purified by a crystallization step (10), where the amounts of undesired by-products are reduced to 50 (PA) and 350 ppm (AcOH). Moreover, AA collection in crude (9) or glacial (11) product quality can be followed by functionalization steps like esterification to acrylates.[4d, 4e, 16a, 18] 2 Fundamentals 9

air (1) 1st oxidation reactor propene off-gas to (2) 2nd oxidation reactor steam incineration (3) quench tower (4) liquid-liquid extractor (5) solvent recovery column gas recycling (6) raffinate stripper (7) crude AA column (8) recovery column (9) crude AA tank (5) (8) (1) (10) product purification unit (11) glacial AA tank (3) (7)

air (4) (10)

(2)

(6)

(9) (11) waste water

Figure 1: Exemplary and simplified flowsheet of commercial AA production.

As can be clearly seen in Figure 1, downstream processing for recovery of AA represents a considerable, extensive and expensive part of the two-step oxidation process.

Due to the corrosive nature of AA, material resistance of the reactors and of all downstream process units plays an essential role, reflecting in cost of the process (e.g. tita- nium-clad steel for distillation towers).[4e] Moreover, further safety topics of the petro- based process are obviously the high exothermicity of the reaction, danger of polymerization, flammability of feed stream (contains propene and air) and the presence of a carcinogenic intermediate, namely acrolein. Furthermore, the dependency on petro-derived propene displays a major drawback of the process, especially in an epoch of rising green and renewable mentality and imperative need for sustainable processes. Propene is mainly produced as a by-product of ethylene production via steam-cracking of natural gas or as by-product of fluid-catalytic cracking of light crude-oil fractions. Hence, propene and thus AA production exclusively rely on fossil resources and are therefore non-sustainable. It should be noted that the increasing propene demand has triggered the development of new on-purpose processes for the production of the latter. Currently, the most important one is the so-called UOP Ole- flex process[19], where propene is produced via dehydrogenation of propane. Finally, the propene price tends to heavy fluctuations due to the strongly varying levels of 10 supply and demand of crude oil, which complicates forecasting economic viability of propene-based processes.[2a, 20]

2.1.3 Acrylic acid production from bio-renewables

Sustainable and more environmentally-friendly chemical processing is gaining both societal and commercial interest. Currently, the bio-renewable production of bulk chemicals is a hot topic in academia and industry and of high importance to overcome the dependency on fossil feedstocks and to reduce the emission of GHGs. In light of the large production quantities of acrylate polymers (e.g. in the field of superabsorbent polymers), the substitution of the fossil raw material propene by renewable feedstock is expected to afford a significant contribution to the reduction of the fossil energy and CO2 footprints of these products. Moreover, the sought substitution contrib- utes to the conservation of finite fossil resources, such as petroleum or natural gas.

This chapter deals with sustainable process opportunities for the production of bio-

AA.[2]

In principle, there are two different approaches to overcome the unsustainable issue of the commercial AA production process. The first one is the adaption of the existing route by replacing the fossil-based feedstock with green propene or green acrolein derived from biomass. This strategy brings the great advantage of applying bio-renewable feedstock within an already established and optimized process pathway. The second approach for bio-AA production is the development of novel processes from alternative and bio-based feedstock. The most relevant substrates for sustainable AA production are currently glycerol, 3-HP acid and LA. An overview of renewable and alternative routes towards bio-AA is shown in Scheme 3. In addition to better environmental attributes, a novel and renewable technology to bio-AA needs to keep up with or even overcome the commercial process on an economical and safety- related level. Up until now, this massive challenge has caused that none of the below presented routes has been commercialized yet. 2 Fundamentals 11

starch 2-APA LA lactide / LA oligomers

glutamic acid 3-HPA

acrylonitrile AA glycerol

3-HP

glucose acrolein

crude oil propene

Scheme 3: Overview of production routes to AA, starting from biomass feedstock (green) or crude oil (black) and proceeding via bio-based platform chemicals (blue) or petro-based intermediates (black); LA = lactic acid; 3-HPA = 3-hydroxypropionaldehyde; 3-HP = 3-hydroxypropionic acid; 2-APA = 2-acetoxypropionic acid; investigated route of this thesis is highlighted by bold arrows; adapted from Beerthuis et al.[2a]

2.1.3.1 Production of bio-based propene

The most obvious approach to allow for a bio-AA production would be a feedstock substitution of the commercial route from fossil- to bio-propene. In 2014, Global Bio- energies announced their success in the production of bio-sourced propene by the use of a proprietary prototype strain, which is able to convert glucose directly to propene.

Even though the process was so far only demonstrated on laboratory-scale, it was the first time ever an entirely biological production process was reported for propene. It should be noted that no comment on the economic viability and commercialization potential of the process has been given yet.[21] However, the production of isobutene from glucose has successfully been demonstrated in industrial pilot scale. Global Bio- energies’s demonstration plant in Leuna (Germany) is in successful operation since

September 2017.[22] A possible expansion of the process to bio-derived propene cannot be assessed at this time. In 2013, Iwamoto et al. described another renewable route for 12 propene production using bio-ethanol as substrate. The reported system achieved a propene yield of approximately 60 %, when a scandium-loaded indium oxide catalyst was used at 550 °C and atmospheric pressure. The reaction was performed in a fixed- bed tubular reactor for 45 to 140 minutes. Ethanol partial pressure was 30 vol% with nitrogen as carrier gas. In the presence of water and hydrogen mixtures as additives, catalyst life time as well as the performance of the reaction system was increased. It is worth mentioning that product selectivity tremendously decreased over longer reaction times. Observed by-products were acetone, (iso-)butenes and AcH. As no further investigations of the catalytic system are given, the reported system is not suitable for producing bio-based propene bio-ethanol on large-scale.[23]

2.1.3.2 Production of bio-based acrolein

Bio-based acrolein, which could be employed in the already established oxidative AA production, can be produced by two different methods. In both cases, glycerol, a cheap and easily available by-product of the production of bio-diesel[2a], is used as feedstock. In the first approach, glycerol is directly dehydrated to acrolein. The dehydration of glycerol can either be performed in liquid or gaseous phase, using homogeneous or heterogeneous catalysts.[24] The best results for liquid-phase dehydration of glycerol to acrolein were reported by Oliviera et al. in 2010.[25] The reported acrolein yield was 92 % with 100 % selectivity towards the desired product. The used feedstock was an aqueous glycerol solution, which was converted at 250 °C for 10 hours using a mordenite catalyst. Their investigation of different zeolite catalysts revealed that catalyst structure, porosity and strength of acidic sites were determining factors for the glycerol dehydration performance. However, batchwise operation mode and catalyst instability are assumed to cause technical inviability on large-scale. A further approach for liquid-phase dehydration of glycerol was published by Shen et al. in

2011.[26] Here, the influence of heteropoly acids on pure glycerol was investigated. The best results were obtained with silicotungstic acid as catalyst with an acrolein yield of up to 79 % within 40 min at 300 °C (at full substrate conversion). Observed by- products were acetic acid, hydroxyacetone and uncharacterized polymers. The reaction was operated semi-batchwise to directly remove the formed products from the 2 Fundamentals 13

reaction system. Nevertheless, undefined polymerized by-products hinted at intermolecular reactions of the substrate and could additionally hamper catalyst recovery.

Therefore, continuous gas-phase dehydration of glycerol seems to be more promising in this context. Shiju et al. reported acrolein yields of up to 70 % (at full substrate conversion) by the use of Nb2O5/SiO2 catalysts.[27] The experiments were conducted at

320 °C and ambient pressure in a continuously operating flow reactor (tr = 10 hours).

The study revealed a dependency of the dehydration performance on the overall acidity of the system. Even though the catalyst deactivated over the course of the reaction, recovery in acidity and activity was enabled by post-treatment with oxygen at elevated temperatures. The last mentionable approach towards a green production of acrolein is an indirect process via the intermediate 3-hydroxypropionaldehyde

(3-HPA). Toraya et al. demonstrated 97 % yield of acrolein from 3-HPA at room temperature and within one hour using a 0.2 M 3-HPA solution, which was adjusted to a pH of 2 with hydrochloric acid (HCl, 35 %).[28] Bio-based 3-HPA can be obtained via enzymatic conversion of glycerol with yields of up to 87 %. Detailed information concerning different reaction conditions and investigated bacteria is given in the relevant literature.[29] However, this method is not yet commercialized and current 3-HPA production is still based on petrochemical feedstock.[4d] It should be mentioned that the direct oxidation of 3-HPA to AA is also an attractive and bio-based alternative with regard to bio-AA production, but no (bio)chemical transformations are known at present.[29-30]

2.1.3.3 AA production from glycerol

In addition to being a suitable substrate for bio-acrolein, glycerol might also be a viable substrate for the direct production of green AA by novel alternative production routes. Moreover, AA production from glycerol can either be processed by a single catalyst as well as by combinations of multiple catalysts in e.g. one-pot processes or consecutive beds. The direct conversion of glycerol to AA was described by Chieregato et al. by the use of V-W-Nb-based catalytic systems.[31] The experiments were conducted in a fixed-bed tubular gas-phase reactor operated at 280-370 °C. They used highly diluted feed streams composed of 2 mol% glycerol, 4 mol% oxygen, 40 mol% 14 water, and 54 mol% helium. Complete conversion was observed with the main products being AA (34 %) and acrolein (17 %). Further by-products were e.g. CO, CO2,

AcH and AcOH. It is worth mentioning that after 100 h time on stream, yield of AA dropped from 34 to 31 %, while acrolein formation was increased (21 %). However, combined yield of AA and acrolein for the best catalyst was limited to approx. 51 % with a very low space-time-yield (STY) resulting from the strongly diluted substrate concentration in the feed. In 2011, Witsthammakul et al. reported a process for the conversion of glycerol to AA via an integrated dehydration-oxidation bed system.[32] In this case, AA is produced in a single reactor via subsequent oxidation of glycerol dehydration products in a consecutive reactor bed. By the combination of medium pore zeolites (HZSM-5) and low substrate concentrations (10-30 wt%), complete glycerol conversion was achieved at 300 °C, with high acrolein selectivity (81 %). Subse- quently, a separated catalyst bed was used to convert glycerol-dehydrated products to AA. While catalysts with high V-content promoted total oxidation towards CO, highly dispersed V-Mo-O phases afforded 48 % acrolein conversion with high selectivity towards AA (98 %). The combined catalytic reaction system gave an overall AA yield of 38 %. Dubois et al. described a two-bed oxydehydration reaction from glycerol to AA in their patent application in 2012.[33] In the presence of molecular oxygen, the system gave an overall AA yield of 75 % at 280 °C using a catalyst combination of

ZrO2 (first bed) and multi-metallic catalyst (second bed).[2a, 33-34] Despite the impressive results, no information with regard to catalyst stability and re-usability were disclosed. In contrast to glycerol to AA conversion via acrolein, which mostly suffered from fast catalyst deactivation, Li et al. presented a novel two-step protocol for AA production from glycerol. [35] The process consists of glycerol deoxydehydration to allyl alcohol, followed by oxidation to acrylic acid. In the first step, glycerol was converted to allyl alcohol by formic acid in either a batch or a continuous flow system at

235 °C. The reported yield of allyl alcohol was 99 %. The second step, namely oxidation of allyl alcohol to AA, was investigated in a quartz fixed-bed microreactor with a continuous substrate flow of 20 mL/min (20 wt% allyl alcohol) in a temperature range of 280-400 °C. Here, supported and unsupported multiple-metal oxides were used as catalyst. Under optimal reaction conditions (340 °C, SBA-15 supported

M-V-W-O catalyst), 90 % overall yield of AA were obtained in the gas-phase reaction. 2 Fundamentals 15

In combination with a catalyst being on stream for over 100 h, this study provides a novel and promising process for the large scale production of AA from bio-renewable glycerol.

2.1.3.4 AA production from acrylonitrile

Acrylonitrile is a bio-renewable platform chemical, entailing potential for being used as intermediate in bio-AA production. Currently, acrylonitrile is predominately produced by the SOHIO process, where propene is converted to acrylonitrile in the presence of air, ammonia and a mixed-metal oxide catalyst at temperatures of

400-500 °C.[36] However, a renewable production of acrylonitrile would enable bio-AA formation via hydrolysis, which is, apart from that, one of the conventional and out- dated AA production technologies.[4d, 37] Furthermore, the biotransformation of acrylonitrile has been reported.[38] In 2010, purified nitrilase brought a yield of 92 % mol % using Rhodococus ruber bacteria under optimized conditions.[39] Ammoxidation of glycerol to acrylonitrile was, for instance, described by Bañares and Guerrero-Pérez in

2008.[40] Glycerol conversion of 83 % with 58 % selectivity towards acrylonitrile was achieved by a V-S-Nb/Al2O3 catalyst at 400 °C. An alternative and solvent-free reaction using microwave irradiation at 100 °C gave 47 % conversion with 80 % selectivity towards acrylonitrile.[41] Moreover, Le Nôtre et al. reported that acrylonitrile can be obtained from glutamic acid.[42] The latter is an industrial waste-product generated in the production of e.g. bioethanol and can further be derived via fermentation.[43] The described process consists of two steps where the intermediate 3-cyanopropanoic acid (3-CNPA) is derived by oxidative decarboxylation of glutamic acid (70 % yield within one hour). 3-CNPA is further transformed to acrylonitrile (17 % yield within 18 hours) in a second step of decarbonylation/elimination reaction. The low yields obtained in this study were attributed to reactant degradation and polymerization.

2.1.3.5 AA production from 3-hydroxypropionic acid

3-hydroxypropionic acid (3-HP), the β-isomer of lactic acid, is one candidate most likely to contribute efficiently to a bio-renewable AA production. The essential reason for the high potential is the outstanding dehydration performance of 3-HP for various 16 conditions and catalysts. Craciun et al. demonstrated and patented a dehydration process yielding 97-98 % AA, even in concentrated 3-HP solutions (20-80 %).[44] Best results were obtained with silica or high surface area γ-alumina as catalysts at temperatures of 250 °C. In 2014, Chwae et al. described 93 % AA yield in a tubular reactor filled with glass beads and heated to 190 °C.[45] Dehydration was achieved by adding inorganic acids to the feedstock (10 wt% 3-HP). The flow rate was 0.25 mL/min. The difference in dehydration selectivity between 3-HP and LA is attributed to the respective elimination mechanism.[2a] However, the bottleneck of this route towards bio-AA is the bio-based production of the substrate. 3-HP yields from glucose are currently too low for a commercialization.[46] This problem might be overcome by coupling fermentation with co-reactions in the future.[47] Alternatively to fermentation of sugars,

Kim et al. demonstrated the direct biotransformation of glycerol to 3-HP using

Klebsiella pneumoniae.[48] Nonetheless, selectivity of 3-HP reached only 11 % at full substrate conversion.

Nevertheless, the high potential ability of 3-HP en route to bio-based AA is reflected in the consortia of well-known companies, following this topic in the last years. OPX

Biotechnologies and Dow Chemical collaborate since 2011 to develop an industrial scale process for the production of AA from bio-renewable feedstock and reported the successful fermentation of 3-HP in 2013.[49] In the same year, a consortium of BASF, Cargill and Novozymes independently demonstrated the sustainable production of 3-HP in pilot scale and announced the successful production of superabsorbent polymers from AA, derived from 3-HP (2014).[50] With BASF, one of the largest AA producers worldwide, backing out of the mentioned collaboration project in 2015, the “leftover” remained committed to bringing the technology to market.[51]

2.1.3.6 Further routes to bio-AA

The possibilities for novel and sustainable routes towards AA are not limited to the above mentioned approaches. Here, a few less prominent technologies for the production of bio-AA need to be mentioned briefly. In 2012, Lejkowski et al. reported the first catalytic synthesis of an acrylate from CO2 and ethene using a homogeneous nickel-based catalyst.[52] The direct synthesis of acrylates from alkenes and CO2 is one of the most economically attractive and most challenging “dream reaction” at once. 2 Fundamentals 17

The development of an efficient AA production route based on ethene (e.g. from bio- ethanol) and CO2 would, for instance, utilize millions of tons of CO2 along a novel value-added chain. Following up on the latter work, Huguet et al. described the development of an efficient one-pot catalyst system with different sets of ligands, forming α,β-unsaturated carboxylic acid salts (e.g acrylates and methacrylates) in high yields and selectivity for the linear[53] product. This method might pave the way for a novel and sustainable production of AA in the future. Moreover, bio-catalytic dehydration of LA was investigated but the enzymatic dehydration of LA and its derivatives was strongly equilibrium limited at ambient conditions.[54] Furthermore, the feasibility of AA production from fermentation is shortly reviewed by Straathof et al., revealing various attractive metabolic pathways for the conversion of sugars to acrylate.[55] However, fermentative AA production mainly suffers from AA toxicity for most organisms and therefore still needs to be further evaluated to achieve acceptable acrylate yields.

2.2 Lactic acid as substrate for bio-AA production

The last presented substrate for a sustainable production of acrylic acid is lactic acid

(LA). LA and 3-HP (both hydroxypropionic acid isomers) are currently expected of bearing most potential for a large-scale and commercial application due to promising dehydration activity and selectivity.[3d, 56] Dehydration of LA to AA is the followed and investigated route of this thesis and therefore introduced in more detail in the further course of this work. First of all, an overview of LA applications and general properties as well as an introduction in the commercial production of LA is given.

After defining important LA derivatives and their potential products, the dehydration of LA to AA is reviewed in detail.

Lactic acid or 2-hydroxypropionic acid is the simplest α-hydroxy acid and furthermore the most widely occurring hydroxy acid in nature. In the main and historic application area of LA, the food sector, LA is used for the production of fermented milk products (e.g yoghurt) and lacto-fermented vegetables (e.g. sauerkraut). It is also known as food additive E 270 and used as e.g. acidifying or emulsifying agent. More- over, LA bi-functionality (alcohol and carboxylic acid) and the associated reactivity open up further application opportunities.[5b, 5c] In recent years, esters of low molecular alcohols are used more and more often as solvents due to the increasing demand of environmentally-friendly and green alternatives. For instance, Corbion has developed and commercialized different specialized applications of lactate esters in elec- tronics and precision cleaning.[5b] Another rapidly growing and important market for LA is the use for the production of the biodegradable polymer polylactic acid (PLA).

PLA is a type of bioplastics, a new generation of plastics offering several benefits, like e.g. increasing resource efficiency, reduction of carbon footprint and (GHG) emis- sions and additional end-of-life solutions (composting). NatureWorks, for instance, has developed a commercial production of PLA and provides Inego®, a biogenic, highly resistant, low-flammable and transparent polymer for e.g. packaging manufacturing since 2002.[57] It is worth mentioning that the PLA market still grows between 10-24 % per year and is expected to garner $5.2 billion by 2020.[5c, 58] 2 Fundamentals 19

Due to the variety of applications and their still growing market, LA became an essential platform chemical in the bio-based economy. The production scale of LA has exceeded 300 ktpa and is estimated to surpass 600 ktpa by 2020. The current market price for LA ranges from $1300-$1600 per ton, depending on the respective product purity grade (50-88 wt%). LA costs are anticipated to drop, resulting from an expected increase in production capacity and improved manufacturing technology, making LA an even more attractive substrate in the future.[5c, 5d, 59] Major commercial LA suppliers are Corbion, NatureWorks, Cargill, and Galactic.[58b, 60]

Pure lactic acid is a white and crystalline solid. However, LA is typically produced and handled as aqueous solution in different concentrations. LA solutions are colorless, hygroscopic and mildly acidic (pKa = 3.79-3.86) liquids, miscible with water and most polar solvents. The viscosity, density, Tm and Tb strongly depend on the respective LA concentration (LA commercially available in solutions from 20-90 wt%).[61]

Due to its two functional groups, LA incorporates a high chemical reactivity and e.g. tends to self-esterification. Hence, commercial LA solutions in equilibrium contain a fraction of LA oligomers, depending on the total concentration of lactoyl units. A commonly produced and applied solution of LA is 90 wt% in water, which only contains about 66 % free acid units. A straightforward example for linear intermolecular esterification of LA is the dimer lactoyl lactate (L2A), typically present in concentrated LA solutions (see Scheme 4). Dimer formation may be followed by further condensation to the trimer (L3A) and water, and so on. However, esterification reversibility enables oligomer hydrolysis to LA.[5c, 62]

Scheme 4: Reversible dimer (L2A) formation from two LA molecules via intermolecular esterification (i) and formation and hydrolysis of LA oligomers (ii); adapted from Dusselier et al.[5c]

The structure of LA is further characterized by a chiral center in C2 position. LA therefore exists as two optically active stereoisomers, L-(+)- and D-(-)-lactic acid (see Figure

2). An equimolar mixture of both stereoisomers is called racemic DL lactic acid. Sub- strate stereochemistry is especially important for PLA chemistry, as optical purity of

LA is crucial for the physical properties of PLA.[5c, 61b, 62-63]

Figure 2: Structural formula of L-(+)- and D-(-)-lactic acid enantiomers.

The LA substrate used in this work was PURAC® FCC 88, supplied by Corbion, which is naturally produced by fermentation of sugar (see chapter 2.2.1). PURAC® is an

88 wt% solution of LA in water and further characterized by stereochemical purity of

≥ 97 % in L-(+)-lactic acid and an initial oligomer content of 15-20 wt%.[64]

2.2.1 State-of-the-art lactic acid production

In principle, there are several routes to LA proceeding on the basis of petro- and bio- based starting materials (see Scheme 5). Chemical synthesis of LA can be achieved by either high pressure hydrocarboxylation of acetaldehyde or hydrolysis of lactonitrile.

The latter is generated by the addition reaction of AcH and hydrogen cyanide and e.g. produced as a by-product of the industrial acrylonitrile production (Sohio-process).[62, 65] Moreover, LA may also be synthesized chemically from bio-based feedstock like glycerol.[66] However, over 90 % of the current LA production is based on bacterial fermentation of carbohydrates, using homolactic organisms (genius Lactoba- cillus). Hence, biotechnological LA production almost covers today’s total LA demand.[5b, 67] Compared to the chemical synthesis, biotechnological production process of LA offers several advantages like low substrate costs, reduced process temperature and less energy consumption.[62, 68] 2 Fundamentals 21

+ CO, + H2O, e.g. Ni(II)iodide on silica, 230 °C, 350 °C, 3h Y = 44 % (iii) O N + acetaldehyde + HCN + H3O

OH lactonitrile

O OH OH HO bacterial fermentation (i) HO O O OH Y > 90 % O O OH OH OH n lactic acid starch

OH Cu O, NaOH, H O, 200 °C, 14 bar N , 6 h (ii) HO OH 2 2 2 Y = 76 % glycerol

Scheme 5: Overview of different LA production methods: Commercial and biotechnological process (i) and chemical synthesis from bio-based (ii) and petro-based (iii) feedstocks; adapted from Beerthuis et al.[2a]

In general, starch is used as feedstock for industrial fermentation processes and the yield of LA (in relation to used glucose) is greater than 90 %. Besides, the use of cheap and abundant biogenic resources and non-food materials (e.g. lignocellular biomass feedstock) for LA production made excellent progress over the past few years.[59a, 67a]

Polymeric carbohydrate substrates are typically pre-treated to release included sugars of the used biomass. Acid hydrolysis is commonly used to break chemical bonds and open the three-dimensional structure releasing glucose molecules. Furthermore, various other processes can be used for biomass pre-treatment (e.g. mechanical com- minution, steam explosion, alkaline hydrolysis or pyrolysis). The bioconversion of carbohydrate materials can further be optimized by direct hydrolysis and fermentation in one step. In simultaneous saccharification and fermentation (SSAF), enzymatic hydrolysis of carbohydrates and subsequent microbial fermentation of the derived sugars are coupled, leading to an increased saccharification rate and productivity.

Additionally, reactor volume as well as capital costs are reduced by this approach.[68-69] 22

Different microorganisms can be used to convert glucose to LA. Mostly, homo-fermentative lactic acid bacteria (LAB) are used for the commercial production of LA, as

CO2 and ethanol are produced as by-products in the fermentation with heterofer- mentative LABs. Anaerobic LA-producing organisms utilize pyruvic acid, the end product of the Embden-Meyerhof pathway (see Scheme 6). Under aerobic conditions

CO2 and H2O formation is energetically favored. In the first step, glucose is converted to pyruvate via glycolysis that is further reduced to lactate with NADPH in the second step.[61a, 62, 68]

2 ADP + 2 P 2 ATP HO HO 2 NAD+ 2 NADH + O 2 NADH 2 NAD O O HO 2 O 2 ñ Oñ- O - - 2 H2O HO HO 2 HO

glucose pyruvate lactate

Scheme 6: Simplified metabolic pathway from glucose to lactate; according to Auras et al.[61a]

Even optically pure L(+)- and D(-)-lactic acid can be obtained by fermentation and appropriate microorganism selection (in contrast to chemical synthesis of LA). Most widely used LABs are Lactobacillus and Streptococcus species. Organisms that predominately produce the L(+)-isomer are L. amylophilus, L. bavaricus and L. casei, while strains as L. delbruecki, L. jensenii or L. acidophilus produce either the

D(-)-isomer of racemic mixtures.[62, 68]

As in many biotechnological processes, product recovery from fermentation broth is the main challenge of the technology. The conventional process for the recovery of LA includes the separation of micro-organisms, product precipitation in form of calcium lactate and subsequent conversion to LA by the addition of sulfuric acid, followed by product purification. To be precise, separation and product purification account for up to 50 % of LA production costs and are characterized by low reactor productivity and atom efficiency: One ton of calcium sulfate (“waste by-product”) is produced per ton of LA. Therefore, further optimization and process improvement of 2 Fundamentals 23

the recovery technology is needed in the future. Promising approaches like membrane-based separation and purification techniques (micro- and ultrafiltration, electro dialysis or ion exchange) or reactive distillation have recently been demonstrated.[5b, 67-68]

2.2.2 Lactic acid reactivity and potential products

Chemically seen, LA is a highly oxidized and reactive molecule with low energy density, making it excellent as a precursor for different products, like e.g. AA. Especially the presence of two functionalities within the LA molecule and the resulting reactivity gives LA a central role as bio-renewable platform chemical for various reactions and the synthesis of chemical intermediates. Scheme 7 shows the variety of potential LA-derived products.

Scheme 7: The central role of LA (cyan) as platform chemical for the synthesis of LA derivatives and PLA (orange) or LA-derived platform chemicals (blue).

In addition to the targeted dehydration of LA to AA (Scheme 7, pathway 1), several

LA-based compounds (orange) and platform chemicals (blue) can be derived from

LA. The dehydration of LA to AA is described in chapter 2.2.3. The focus of this chapter is on the related chemistry of LA, which defines the basis of this thesis. The special character of both functional groups within the α-hydroxy acid lead to diverse acid catalyzed reactions like self-esterification, oligomerization and even polymerization.

Moreover, several reactions of LA have been observed, most of which are catalyzed by nucleophilic (Nu) and electrophilic (El) centers. On the one hand, this makes LA a versatile and bio-renewable feedstock for different platform chemicals. On the other hand, selective dehydration to AA is complicated by the high reactivity of LA. In the following section, the most important LA derivatives are introduced, which can, beyond that, even be seen as possible LA substrates for various reactions. Furthermore, relevant reaction pathways of LA, which are competing with LA dehydration, are briefly described.

Lactide (3,6-dimethyl-1,4-dioxan-2,5-dion) is the cyclic dimer of lactic acid, which is formed by intermolecular esterification, more specifically selective cyclization

(Scheme 7, pathway 2). Due to the ester groups, lactide hydrolyzes in acidic and aqueous conditions and consequently forms two molecules of LA. Due to this condensation / hydrolysis equilibrium, lactide can mainly be found and handled in dry conditions. Lactide is produced in medium-scale and its most prominent application is the production of high molecular weight PLA. A low-cost and continuous process starting from LA has been developed by Cargill Dow LLC. Aqueous LA solution is self- esterified to low molecular weight LA oligomers by in-situ vacuum distillation (water removal) at 200-250 °C and 10-100 mbar. The oligomer intermediate is subsequently cracked into the cyclic dimer lactide via the so-called “backbiting mechanism”. The reaction is catalyzed by homogeneous tin-catalysts and conducted at a reaction temperature above 200 °C and a reduced pressure of 1-80 mbar. The obtained lactide mixture can further be purified by vacuum distillation or crystallization and subsequently polymerized top PLA via ring-opening polymerization (ROP).[70]

2 Fundamentals 25

In addition to specific polymerization, which is of interest for industry, unwanted and spontaneous polymerization of LA is known (Scheme 7, pathway 3). As typical for esterification reactions, LA oligomerization / polymerization is accelerated by the presence of Brønsted acidity. This is also the reason for autocatalytic self-esterification of aqueous LA solutions to oligomers (see chapter 2.2).[71]

Scheme 8: LA oligomers and lactide as self-esterified LA intermediates en route to PLA.

The salts and esters of LA are called lactates. Sodium lactate, for instance, is the most frequently occurring lactate in the human organism. Further prominent representa- tives of metals forming lactates are potassium, magnesium and calcium. Lactates are typically formed by neutralization of LA with e.g. sodium hydroxide (Scheme 7, pathway 4) Furthermore, LA can be transformed into so-called alkyl lactates by esterification with various alcohols (Scheme 7, pathway 4). Due to the reduced reactivity resulting from protecting one of LA’s reactive sites, inter- and intramolecular reactions are slightly limited in these esters. The most prominent ester of LA is lactic acid ethyl ester (ethyl lactate), which is conventionally produced from LA and ethanol using acidic catalysts. Ethyl lactate is an environmentally benign solvent and further used as alternative LA substrate (also methyl lactate) for various applications like e.g. dehydration to AA.[72] The α-hydroxyl functionality can even be esterified or substituted to give e.g. α-alkyloxy derivatives of LA or lactates (Scheme 7, pathway 5). 26

2-acetoxypropionic acid (2-APA), for instance, can be synthesized via acetoxylation of

LA with AcOH and is also an attractive substrate for bio-renewable AA production

(see chapter 2.2.3).[73] Instead of esterifying the hydroxyl functionality, substitution reactions are possible. In the context of this thesis, α- and β-brominated LA-species play a crucial role. This class of LA derivatives is therefore separately highlighted in chapter 2.3. Even higher functionalization of LA is accessible by both esterifying the hydroxyl and carboxyl functionality. 2-acetoxypropionate (EAPA) can either be synthesized from 2-APA and ethanol or from ethyl lactate and acetic acid and represents a fully-protected LA skeleton, which is also used as substrate for e.g. bio-AA production.[7] All these LA derivatives are products of esterification reactions of LA with various substrates, typically performed in acidic environment or are formed by neutralization with bases. Next to the diversity of the LA molecule and its functionalization, this underlines the high and versatile reactivity of LA. It is worth mentioning that most esters of LA possess reduced boiling temperatures compared to “free” LA, which is due to reduced ability of hydrogen bond formation. An overview of the above mentioned lactates and LA derivatives is given in Scheme 9.

Scheme 9: LA derivatives formed in the presence of acids and alcohols or bases.

2 Fundamentals 27

In addition to the above described LA derivatives, LA can also be converted to several other platform chemicals or intermediates (see Scheme 7). A very important reaction pathway starting from LA, especially observed as undesired side reaction in LA dehydration, is C3-decomposition to AcH via either decarbonylation or decarboxylation

(Scheme 7, pathway 6).[5c, 9b] AcH formation via decarbonylation (CO and H2O as coupling products) is triggered by activation of the carboxylic group via protonation (see

Scheme 10, i).[9b] This phenomenon was observed in gas-phase dehydration of LA in the presence of Brønsted acidic catalysts and even autocatalytic in liquid-phase and high temperature processes.[8, 74] In most cases, decarbonylation could be avoided by reducing the acidity of the catalytic system.[5c] LA, like many other carboxylic acids, tends to decarboxylate to AcH, CO2 and H2 in the presence of redox active metals like iron or nickel.[8-9, 75] Moreover, catalytic decarboxylation was reported in the presence of high alkaline oxides, similarly to the behavior of other carboxylic acids (e.g. MgO,

BaO)[76] and further received increased importance in basic media.[9b] Decarboxylation is initiated by a nucleophile-activated carboxylic functionality (see Scheme 10, ii).

Furthermore, decarboxylation of β-hydroxy acids is well known and, for instance, a frequently observed side reaction in the Perkin reaction.[9b]

Scheme 10: Mechanisms of LA decarbonylation (i) and decarboxylation (ii) triggered by activation of carboxylic group via protonation (i) or nucleophilic attack (ii).

The above described reactions, especially self-esterification and C3-decomposition of

LA via decarboxylation and decarbonylation, are the most relevant side reactions and derivatizations (lactates and LA derivatives) of LA, especially in the context of this thesis and in the liquid phase. Nevertheless, LA is a suitable starting material for various products and reactions. Worth mentioning is the condensation/dehydration of

LA to 2,3-pentanedione (Scheme 7, pathway 7), the reduction to 1,2-propanediol

(Scheme 7, pathway 8) or the oxidation to pyruvic acid (Scheme 7, pathway 9). More- over, as highly reactive and oxidized hydrocarbon (C:O atomic ration of 1), catalytic upgrading or reforming of LA is described in the literature. However, these consecutive reactions of LA were not observed within this thesis and are therefore not further highlighted at this point. Detailed overviews of further reactions using LA as substrate and the respective reaction conditions and application areas are given by

Dusselier et al.[5c] and Mäki-Arvela et al.[5d]

2.2.3 Dehydration of lactic acid to acrylic acid

Since the first patent describing the feasibility of LA dehydration in 1958[6], several groups have approached and investigated this subject. The dehydration, which is nominal the elimination of one water molecule, proceeds via a carbocation transition state at the α-carbonyl position. For this reason, the most common and competing reaction pathway is the decarbonylation to AcH. Furthermore, decarboxylation as well as substrate oligomerization need to be inhibited to achieve high product selectivity. Particularly, gas-phase dehydration of LA to AA and the involved catalytic system and process operation has been described in detail in literature. Another examined approach was the dehydration of LA to AA in near- and supercritical water.

In contrast, liquid-phase dehydration of LA to AA has so far scarcely been described.

LA dehydration has not yet been commercialized and to use the technology on an industrial scale, still further improvements are necessary. This chapter describes the development of the state-of-the-art LA dehydration and, in this context, also the use of LA derivatives as substrate for bio-AA production (see chapter 2.2.2).

2 Fundamentals 29

2.2.3.1 Gas-phase dehydration of lactic acid

The first systematic approach towards gas-phase dehydration of LA towards AA was reported by Holmen and patented by the Minnesota Mining & Manufacturing Company in 1958 (see Scheme 11).[6] The conversion of LA to AA was described by the use of silica-aluminum salts (sulfate, nitrate, phosphate, etc.) and mixtures of other inorganic salts. The reaction was conducted at reaction temperatures between 200-600 °C and ambient pressure. The best results were obtained at a reaction temperature of

400 °C and over a NaSO4/CaSO4 catalyst with a theoretical AA yield of 68 % (no information on AA selectivity was provided).

Scheme 11: Dehydration of LA to AA over NaSO4/CaSO4 catalyst at 400 °C according to Holmen (1958).[6]

The gaseous and heterogeneously catalyzed selective dehydration of LA to AA has been investigated since then and particular progress has been made in the last few years. The most relevant work is described in the following.

It is well known that Y-type zeolite can be used as dehydration catalyst and Wang et al.[77] (2008) and Sun et al.[78] (2010) described AA yields of 50-56 % by the modification of NaY with KNO3 and lanthanum oxide. In 2011, Zhang et al. modified fully Na+-exchanged Y zeolite with alkaline phosphates for the dehydration of LA.[79] The experiments were conducted in a fixed-bed tubular reactor at ambient pressure, 350 °C and with a LHSV of 2.3 h-1. The used feedstock was a 34 wt% aqueous LA solution. The obtained yield of AA was 58 % with 72 % selectivity in AA. Observed by-products were CO, AcH, propionic acid, acetic acid and 2,3-pentanedione. Interestingly, the high selectivity towards AA was attributed to the formation of alkali lactates, which were in-situ generated by proton transfer to the basic NaY support. Furthermore, the phosphate seems to stabilize the carboxyl functionality of both LA and AA. However, the used catalyst suffered from severe catalyst deactivation at longer times on stream. 30

Due to accumulation of LA oligomers on the catalyst surface, selectivity to AA dropped to 40 % after 28 h. The used catalyst was reactivated at 500 °C in air and showed a better performance in terms of a reduced decline of AA selectivity with time on stream.

Yan el al. reported LA dehydration over hydroxyapatite catalysts in 2014.[4f] The experiments were carried out at 360 °C and ambient pressure in a fixed-bed tubular reactor and further characterized by LA feed concentration of 10 or 35.7 wt%. The used carrier gas was nitrogen and the weight hourly space velocity (WHSV) of LA was

2.1 h-1. The highest observed yield of AA was 62 % with a selectivity of 74 %. Detected by-products were mainly CO and AcH resulting from decarbonylation. Additionally, PA, AcOH, 2,3-pentanedione, and 1-hydroxyacetone were observed. The authors reported a strong correlation between surface acidity/basicity ratio and catalyst performance. The catalyst suffered from deactivation over 8 hours and 12 hours were required for full catalyst regeneration (360 °C). In 2016, the same group employed Rb+ and Cs+-exchanged beta zeolites for the sustainable production of AA from LA. The experimental setup and reaction parameters were retained and the beta zeolite exchange degrees were varied (RbxNa1-xβ and CsxNa1-xβ). The best performing catalysts were Rb0.95Na0.05β and Cs0.081-0.90Na0.19-0.10β with AA selectivity of approximately 70 % and yields from 60-65 %. The authors claimed that both weakly acidic and weakly basic surface sites (suitably balanced acidity/basicity ratio) are crucial for highly selective dehydration catalysts. Most competing observed side reactions were acid-catalyzed decarbonylation and decarboxylation as well as base-catalyzed condensation of LA. Recently, the group reported the use of K+-exchanged zeolites for AA production by gas-phase dehydration of LA.[80] ZSM-5 and β zeolite were found to be more efficient than the rest of the investigated structures. Especially K0.97Na0.03ZSM-5 showed exceptional dehydration performance with very high AA selectivity (80-81 %) and yield (75-78 %). The experiments were performed at 360 °C and under a WHSV of LA of 2.1 h-1. Furthermore, this catalyst showed remarkable stability for longer than 80 h. Again, NH3- and CO2-TPD measurements revealed the essential role of a balanced surface acidity/basicity ratio.

2 Fundamentals 31

In 2014, Ghantani et al. reported their investigations on non-stoichiometric calcium pyrophosphate catalysts for the dehydration of LA to AA.[81] The experiments were conducted in a fixed-bed tubular reactor in a temperature range from 325-400 °C. The feed concentration of LA ranged from 25-80 wt % (aqueous) and the WHSV was varied from 1 to 4.5 h-1. The best performing catalyst with a Ca/P ratio of 0.76 achieved an AA yield of 78 % at full substrate conversion. The best reaction parameters were found to be 375 °C, WHSV of 3.0 h-1 and feedstock concentration of 25 wt% of LA in water. At higher Ca/P ratios and feedstock concentrations, selectivity to AA decreased. The main side products were CO, CO2, AcH, PA, and 2,3-pentanedione. Alt- hough high AA yield and selectivity were obtained, the authors did not provide additional information about catalyst stability.

The so far highest yields of AA obtained from gas-phase dehydration of LA and its derivatives were reported by Godlewski et al. (2013) and Velasquez et al. (2015 & 2017) and patented by The Procter & Gamble Company.[7] LA dehydration was conducted in a glass-lined stainless steel tubular reactor and the used catalysts were condensed phosphates with cations of various metals. The experiments were further performed at temperatures of 300-425 °C and pressures of approximately 25 bar. The best results were obtained by separate gas and liquid feeds, mixed before reaching the catalyst bed with a temperature of 375 °C, whereby the liquid stream was a 20 wt% aqueous LA solution with a feed rate of 0.045 mL/min. The gaseous carrier gas stream was molecular nitrogen with a feed rate of 45 mL/min. The reported AA yields reached

88±2 % with selectivity towards AA of 90±1 %. Observed by-products in decreasing order were AcH, 2,3-pentanedione, AcOH, PA, DiAA, and 1-hydroxyacetone.

A worth mentioning example of lactate dehydration to AA was published by Blanco et al. in 2016.[72a] Alkaline earth phosphates, which were previously shown to be active catalysts for the dehydration of LA, were investigated in the gas-phase dehydration of ethyl lactate to AA and ethyl acrylate. The experiments were conducted in a fixed-bed tubular reactor at atmospheric pressure and temperatures from 350-390°C.

The used substrate was either pure or a 20 wt% aqueous solution. High selectivity towards AA and ethyl acrylate was observed and attributed to the inhibition of de- 32 carbonylation and decarboxylation. In the best performing system configuration, dehydration products reached a selectivity of 87 %. However, ethyl lactate conversion was decreased compared to LA, because of its higher stability and polymerization of ethyl acrylate on the catalyst surface. The latter strongly deactivated the catalysts, which could be hampered by large amounts of water added to the feed. Nevertheless, this resulted in an increased share of decarbonylation and decarboxylation.

In addition to the direct dehydration of LA or lactates to AA, some studies were reported using an “indirect” gas-phase conversion of LA to AA via LA intermediates.

The pyrolysis of LA derivatives towards AA was first reported by Burns et al. in

1935.[82] 2-APA, for instance, can be synthesized from LA or lactide and acetic acid or acetic anhydride by acetoxylation in yields over 90 %. Here, “dehydration” is performed in a previous and separated reaction step. Subsequent pyrolysis of the LA ester forms the unsaturated product (AA) and a carboxylic acid as by-product

(AcOH). Pyrolysis of 2-APA or other LA derivatives like methyl 2-acetoxypropionate (MAPA) is straightforward and more selective in AA than the direct dehydration of

LA (yields up to 95 %). 2-APA pyrolysis proceeds via a circular six-center transition state (concerted Ei mechanism) and does not involve a carbocation transition state.[2a, 4e, 73, 83]

2.2.3.2 Dehydration of LA in near- and supercritical water

In the context of developing environmentally benign chemical processes, dehydration of various biogenic compounds was investigated in near- and supercritical conditions.[84] Water near or above its critical point (220.9 bar and 374 °C) possesses interesting properties for the production of bio-AA from LA dehydration. In addition to being an excellent solvent for organic compounds, near- and supercritical water gen- erates a high proton concentration, capable of acid-catalysis without adding any acid.

Moreover, in contrast to dehydration in the gas phase, energy-intensive steps like evaporation of low-concentrated and aqueous substrate solutions or purification may not be required.[85]

In 1989, Mok et al. investigated the dehydration of LA in supercritical water.[9b] The experiments were conducted in a Hastelloy-C tube at reaction temperatures between 2 Fundamentals 33

325-400 °C and pressures between 207-345 bar. Highest yields of AA (14 %) were reported when low-concentrated LA solutions were treated in basic reaction milieu.

Moreover, acidic conditions increased the share of substrate decarbonylation. Based on these findings, Lira and McCrackin investigated the conversion of LA to AA in near-critical water (1993).[8] The experiments were conducted in a Hastelloy C-276 an- nular reactor at a pressure of 310 bar and temperatures of 320-400 °C. In their work, the focus was strongly on the influence of various catalysts and their influence on the dehydration performance (e.g. phosphoric acid, NaOH and Na2HPO4). At an optimized reaction temperature of 360 °C and by addition of Na2HPO4, AA yield was increased to 58 %. However, LA share in the feed (4 wt%) and substrate conversion

(approx. 15 %) were very low. Finally, Aida et al. investigated the reaction of LA with a flow apparatus in water at high temperatures (450 °C) and high pressures

(400-1000 bar). The study focused on the influence of water density on the dehydration performance without the addition of an additional catalyst. The maximum achieved selectivity to AA was 44 % at 23 % LA conversion (450 °C, 1000 bar). As in the case of the two previous studies, a compromise between high conversion and simultaneous high selectivity towards AA was not achieved.

2.2.3.3 Liquid-phase dehydration of lactic acid

The third and by far least investigated possibility for bio-AA production from LA is the dehydration in the liquid-phase. All previously presented processes for the production of AA from LA were accompanied by several drawbacks. Resulting from the harsh reaction conditions of LA dehydration in the gas phase or in near- and supercritical water, side product formation is induced, limiting AA selectivity. Especially, decarbonylation, decarboxylation and oligomerization strongly compete with the desired dehydration pathway. Moreover, PA is formed in many cases, which is particularly unwanted. PA in the product stream significantly hinders product purification due to the very similar boiling temperature (Tb = 141 °C) to AA and should therefore be avoided. A further considerable aspect is that due to high substrate dilution within the feed stream, STYs are strongly limited. In many cases, an increased LA concentration resulted in decreased dehydration performance. In addition, almost all previously described and heterogeneously-catalyzed processes suffer from severe catalyst 34 deactivation. Last but not least, the energy demand of high temperature and/or high pressure operation is cost-intensive. Furthermore, LA as well as AA and various by-products are highly corrosive under these conditions, leading to additional investment cost for the dehydration reactor and setup (material compatibility).

In the light of the above mentioned disadvantages, LA dehydration in the liquid phase seems favorable. In general, a liquid-phase process is characterized by significantly milder reaction conditions, compared to e.g. gas-phase operation (decreased reaction temperature and pressure). In addition to cost savings due to reduced heat input, working at e.g. ambient pressure allows for the use of cheap and inert glassware, reducing the investment costs of the dehydration reactor and setup. Further- more, the formation of undesired by-products may be reduced at lower reaction temperatures, increasing product selectivity. Besides, liquid-phase processes often enable to choose a simple reactor design (e.g. CSTR). Moreover, liquid-phase dehydration is further motivated by the use of homogeneous catalysts and reaction media, involving several advantages for the dehydration process: Prevention of transport limitations by diffusion, prevention of catalyst deactivation and specifically tailoring the reaction medium to the needs of the process (e.g. by the use of ILs, see chapter 2.4).

However, up until now, only two liquid-phase processes for the production of bio-AA from LA or LA derivatives were described in the literature, which demon- strates the challenging task of dehydrating LA or LA derivatives in the liquid phase.

In 2014, Kuppinger et al. reported a process for the dehydration of hydroxycarboxylic acids in a liquid-phase process, patented by Evonik Industries.[10a] The process is characterized by a simple dehydration setup. A diluted substrate solution (LA or 3-hydroxypropionic acid; 60-90 wt% in water) was fed into a round bottom flask and introduced below the surface of the reaction mixture with the help of a glass tube. The reaction mixture was continuously stirred and the sump temperature was kept at temperatures between 160-300 °C. The liquid reaction mixture consisted of mixtures of metal lactates, metal carbonates, and metal phosphates of varying composition.

Preferably, alkaline or earth alkaline metals were used as cation and the most appropriate salts were K2CO3, KH2PO4, K2HPO4, Na2HPO4, BaHPO4. Moreover, water and polymerization inhibitor (e.g. phenothiazine or 4-methoxyphenol, 0.05-2.5 wt%) were 2 Fundamentals 35

previously added to the molten salt reaction mixture. Reaction products were continuously removed from the hot reaction zone at reduced pressure (preferably

50-100 bar), condensed and collected. Maximum YAA obtained from LA was 1.3 % at a sump temperature of approx. 300 °C, a reaction time of 5.3 h and with K2HPO4 as salt additive. Even though the used salt additives were similar to active phosphate-based catalysts used in gas-phase dehydration of LA and its derivatives, the reported yields of AA were strongly limited in the liquid phase. Moreover, the reported yields of AA (< 1.5 %) may result from thermally-induced substrate decomposition, especially at high reaction temperatures (≥ 300 °C).

While the work of this PhD thesis was on the way, Terrade et al. have reported a novel method for the production of bio-AA from fermentatively produced lactide/LA derivatives.[10b] A patent application of this process has been filed by Purac Biochem BV in 2015.[10c] Terrade et al. described a catalytic cracking of lactide in bromide salts (e.g. tetraphenylphosphonium bromide, PPh4Br) in the presence of a strong acid (e.g. methanesulfonic acid, MSA). This novel process is very different from the previously described processes to bio-AA from LA feedstocks, as the described reaction is a rearrangement of lactide and not a dehydration of LA. Moreover, the reaction is performed at significantly lower temperatures (130 – 175°C) compared to the typical process conditions (300 – 500°C).[81, 86] The reactions were conducted batchwise in 9 mL autoclaves (glass inset) at a pressure of 50 bar (N2), at autogenic pressure in closed autoclaves or at ambient pressure (Schlenk glassware). After cooling to room temperature, the reaction mixture was analyzed by 1H NMR. Alternative feedstocks for this liquid-phase process for bio-AA production include LA oligomers and LA polymers, but no monomer solution of LA was mentioned. The proposed mechanism of lactide cracking proceeds via brominated LA-species and involves a SN2 / elimination sequence, forming AA from 2-BrPA, for example (see Scheme 12). The cyclic and water-free substrate lactide (a form of “dehydrated LA”) is first opened by in-situ formed HBr to give (2-bromopropanoyloxy)propionic acid (1) via nucleophilic substitution.

Compound 1 can either react with a second HBr equivalent to yield two equivalents of 2-bromopropanoic acid (2-BrPA) or eliminate HBr to give 2-(acryloyloxy)propionic acid (2). Compound 2 can react with HBr in an identical manner to give AA and 36

2-BrPA. The formed 2-BrPA eliminates HBr and consequently forms AA. Moreover, the authors claim that traces of 3-bromopropionic acid (3-BrPA) result from the reversible addition of HBr to the vinyl functionality of AA.

Scheme 12: Proposed mechanism for the catalytic cracking of lactide to AA by MSA in PPh4Br; according to Terrade et al.[10b]

At optimized reaction conditions, AA yields of up to 58% obtained after 10 h at 175°C were reported at complete lactide conversion (SAA = 58 %). The reaction mixture included 0.69 mmol lactide (substrate), 0.83 eq. of anhydrous methanesulfonic (strong acid), 5 eq. of tetraphenylphosphonium bromide (bromide salt), and 5 eq. of sulfolane

(solvent). Water-free conditions were claimed to be crucial to achieve high yields of AA.

While this is a remarkable and highly interesting result, the economics of such an alternative process for bio-AA production from biomass-derived lactide would largely improve if higher AA yields could be achieved. It should be mentioned that the above described process was developed in parallel and independent to the technology developed at the Institute of Chemical Reaction Engineering (CRT) of the Frie- drich-Alexander-University Erlangen-Nürnberg (FAU) that forms the basis of this thesis.

However, both processes enable a liquid-phase dehydration of LA or LA derivatives towards bio-acrylic acid under comparable conditions.

2 Fundamentals 37

2.3 Bromopropionic acids

Bromo-organic compounds are used in diverse chemical reactions like bromination[87], oxidations[88], hydrolysis[89], substitutions[90], or catalysis[91] and are therefore involved in the synthesis of complex organic molecules, e.g. in the polymer, pharma- ceutical, and agrochemical industries.[92] Due to their acidity, reactivity, and distinct biological properties[93], halogenated carboxylic acids (HCAs) are typically used as alkylation and acylation reagents in the chemical industry[94]. 2-bromopropionic acid, a bromo-derivative of the above mentioned HCA-group, is used for example in the synthesis of triazole derivatives[95], which have attracted tremendous interest among organic and medical chemists owing to their remarkable and wide range of biological activities.[96] In addition, a recent publication by Terrade et al.[10b] as well as contributions from Wasserscheid et al.[97] have added a new and highly relevant field to the group of technically interesting applications of bromopropionic acids: 2-BrPA as well as 3-BrPA were identified as intermediates en route to bio-AA production from fermentatively-derived LA.

In the context of this thesis, brominated LA derivatives, namely 2-BrPA and 3-BrPA, are of particular importance for the liquid-phase dehydration of LA to bio-AA. More- over, the developed and presented multi-step dehydration process from LA to AA proceeds via 2-BrPA and 3-BrPA and therefore has taken up the challenge to integrate bromopropionic acids into a renewable production route to AA, or even use them as bio-based feedstock for AA production. That is precisely the reason why bromopropionic acids, their respective production methods and especially the related chemistry in terms of bio-AA production are highlighted in the following chapter.

The chemical structures of both 2-BrPA (also α-bromopropionic acid) and its constitutional β-isomer 3-BrPA are shown in Scheme 13. In this work, the transformation of 2- to 3-BrPA is referred to as “isomerization”. 38

Scheme 13: Structural formula of 2- (left) and 3-bromopropionic acid (right) isomers.

2.3.1 Conventional production of 2-bromopropionic acid

Currently, 2-BrPA is produced via the Hell-Vollhard-Zelinsky reaction from mainly crude-oil-based PA (see Scheme 14). In this reaction, PA is processed with catalytical amounts of phosphorous tribromide (PBr3) and elemental bromine at elevated temperatures forming considerable amounts of waste by-products.[11a-c]

Scheme 14: Mechanism of the Hell-Volhardt-Zelinsky reaction for the production of 2-BrPA from petroleum-based PA.

The first step of the above shown reaction sequence is the formation of an acyl bromide (1) from PA and PBr3, which is correlated to the formation of one equivalent of phosphonic dibromide (PBr2OH). The formed acyl bromide tautomerizes to an enolic form (2), which can readily undergo bromination at the α-carbon to give α-bromo 2 Fundamentals 39

acyl bromide (3). In this step, one equivalent HBr is eliminated. The α-bromo acyl bromide intermediate is subsequently quenched by another equivalent of PA, forming an anhydride (4) and thereby releasing one bromide. The latter attacks the anhydride to cleave the C-O bond. As a result, one acyl bromide and one α-bromo carboxylate (5) is formed. After hydrolysis of the α-bromo carboxylate, the targeted molecule, namely 2-BrPA, is obtained. The coupling product of the last reaction step, namely acyl bromide, can be incorporated in the reaction sequence.[11a-c]

By all accounts, this process does not include green and sustainable process conditions. In addition to the reliance on fossil-based PA as feedstock, the use of PBr3, which is finally converted to phosphorous acid (H3PO3), leads to considerable amounts of waste by-products. Moreover, the use of elemental bromine (Br2) is accompanied by high environmental risks, if released. In terms of establishing a sustainable route for the production of bio-AA, conventionally produced 2-BrPA is therefore a non-viable option.

A much more suitable substrate for a sustainable production of 2-BrPA is LA, which is commercially produced by fermentation of sugar (see chapter 2.2.1) and already used as safe and bio-based chemical in the food industry and for the production of the biodegradable polymer PLA. LA can theoretically be converted to 2-BrPA by bromination (substitution of its hydroxyl group by a bromine atom).

2.3.2 Bromination of alcohol functionalities

The following chapter describes the fundamentals of the bromination of alcohol functionalities. First of all, mechanisms of the nucleophilic substitution of alcohols are described. Hereafter, different synthetic methods for the bromination of alcohols are presented.

2.3.2.1 Basics of nucleophilic substitution of alcohols

LA bromination to 2-BrPA is a typical example of a nucleophilic substitution. The

α-carbon attached to the oxygen of the alcohol functionality is characterized by a partial positive charge. The nucleophilic bromide anion, which bears a free electron pair, 40 can attack at the α-carbon and thereby forms a covalent bond. Nucleophilic substitutions have intensively been studied in literature and Hughes and Ingold postulated the concept of two clearly distinct mechanisms for this reaction, namely the SN1 and SN2 mechanism (see Scheme 15). The SN1 mechanism is characterized by two discrete steps, where the first step is a heterolytic bond scission, forming a positively charged intermediate, called carbocation. The product is consequently formed by combination of the formed carbocation with the attacking nucleophile. In the SN1 sequence, the nucleophilic attack leads to a loss of the substrate’s stereochemistry, resulting in a racemic mixture of the product (recombination can take place from both sides of the carbocation). In contrast, the SN2 mechanism involves a Walden inversion transition state. The leaving group is displaced by the nucleophile via a backside attack in a single concerted reaction. This leads to an inversion of the configuration of the stereochemical center.[98]

The steric and electronic environment of the leaving group and the related carbon atom influences and defines the present mechanism (SN1 or SN2). At tertiary substituted carbon atoms, mostly the SN1 mechanism takes place due to stabilization of the charged transition state by +I-effect of the surrounding substituents and the sterically hindering for a single concerted substitution mechanism. In contrast, primary carbon atoms mostly underlie a SN2 mechanism. The single substituent does mostly not bear enough electron-releasing character to sufficiently stabilize the carbocation within a

SN1 sequence. Moreover, primary carbons are less sterically hindered, allowing for inversion of the transition state by a back side attack of the nucleophile.[99] It is more difficult to assign the distinct mechanism within the substitution at secondary carbon atoms (like α-carbon in LA). Due to the strong dependency on the size and electronic character of the nucleophile and the leaving group to be substituted, it is suggested that at most secondary carbons both SN1 and SN2 take place simultaneously.[100]

In most cases, an activation of the OH-group is necessary beforehand, due to the bad leaving group capability of hydroxyl groups. This can either be achieved through activation by a Lewis acid or protonation of the leaving group and the resulting transformation of OH into a good leaving group, namely water.[101] In the case of LA bromination to 2-BrPA, both described mechanisms can be expected as the hydroxyl 2 Fundamentals 41

group is located at the secondary α-carbon. Moreover, it has been observed that the intermediate carbocation can readily undergo decomposition via e.g. decarbonylation to AcH and CO. The latter can either be triggered in acidic environment or even be thermally-induced.[9a, 102] The mechanism for the bromination of LA to 2-BrPA using

HBr is shown in Scheme 15.

Scheme 15: Bromination of LA with HBr including both mechanistic opportunities SN1 (top) and SN2 (bottom).

2.3.2.2 Synthetic methods for bromination of LA

On the one hand, bromination of LA requires the activation of the “bad” leaving group (OH). On the other hand, a bromination agent is required to enable 2-BrPA formation. In the following, different bromination methods are briefly introduced and evaluated in the context of a green and sustainable production of 2-BrPA from

LA.

The use of aqueous and concentrated hydrobromic acid (48 wt% in water, HBr(aq)) is obvious and commonly applied in organic synthesis. The used reaction conditions vary strongly and need to be adjusted to the specific substrate. Typically, HBr(aq) is used in an excess of up to 6 equivalents. Bromination reaction temperature strongly depends on the specific substrate and ranges from 0 to approximately 122 °C (reflux).

Short reaction times (> 12 h) may be used to suppress undesired side reactions. Fre- quently observed side reactions are e.g. etherification and dehydration. However, bromination by-products also strongly depend on the applied substrate and especially on its functionality. As described in chapter 2.2, LA is a highly functional and 42 reactive molecule. Undesired side reactions can therefore be expected in LA bromination. In some cases, addition of sulfuric acid is used to increase bromination selectivity. In industry, HBr(aq) is used as bromination agent for the production of simple alkyl bromides.[103] Moreover, the bromination agent HBr can also be provided from two different molecules, one carrying an acidic proton and the other one carrying a nucleophilic bromide. In this context, ionic liquids are frequently used as part of the bromination reagent and medium. The most important applications are addressed and described in a subsequent part of this thesis (see chapter 2.4.2).

Indeed, the formation of 2-BrPA from LA and HBr(aq) has been first described in 1864 by Kekulé.[104] In the described bromination experiment, a bomb tube was charged with LA and HBr(aq) and subsequently sealed and heated for 48-72 hours. Although no information was provided on bromination selectivity towards 2-BrPA, the formation of gaseous by-products was reported.

In addition to the above described bromination with HBr(aq), further synthetic methods for the bromination of alcohols are known using various bromination reagents.

The following methods differ with respect to the activation step of the leaving group (previously accomplished by an acid) and the nature of the bromination agent itself.

Frequently used substances are phosphorous-based compounds, but sulfur and boron compounds are also mentioned.

The first group of bromination reactions presented here uses phosphorous compounds for the halogenation/bromination of alcohols. The driving force of this concept is the formation of the thermodynamically favored phosphine oxide double bond.[105] Typical phosphorous-based bromination agents are PBr3, PBr5[103b],

PPh3/Br2[106], PPh3/N-bromosuccinimide[107] and PR3/imidazole/Br2[108]. The most common and worth mentioning halogenation/bromination method using phosphorous compounds is the Appel reaction, where triphenyphosphine (PPh3) and a carbon tet- rahalide (CHal4) are used as halogenation/bromination reagents.[109] The bromination is initiated by the in-situ formation of the active halophosphonium salt [Hal-PPh3]+.

By means of the Appel reaction, organohalides are formed in good yields and HCHal3 and triphenylphosphine oxide (OPPh3) are formed as by-products. Due to the wide range of potential substrates and the mild reaction conditions, this method is one of 2 Fundamentals 43

the most successfully applied reactions for the preparation of organohalides in organic synthesis.[110] However, the Appel reaction implies several disadvantages.

Firstly, the use of carbon tetrahalides is correlated with their heptatotoxicity and environmental concerns.[111] A variety of alternative bromide sources for the use in the

Appel reaction have been developed.[110] Moreover, formed by-products need to be removed within the product purification step, which is mostly connected with difficult and time-consuming chromatography methods. The use of polymer supported phosphorous might overcome this limitation.[112] Last but not least, the use of equimolar amounts of PPh3 used in the halogenation/bromination reaction leads to large amounts of waste by-products (triphenylphosphine oxide) and results in low atom efficiency of the process.[113] Therefore, recent progress has been made in the development of catalytic Appel reactions. However, such processes still make use of expensive and hazardous reduction agents (e.g. Ph2SiH2 or COCl2).[112a, 114]

A further method using a sulfur-containing bromination agent for the halogenation of alcohols was published by Pouliot et al.[115] In this study, a diethylaminodifluorosulfinium tetrafluoroborate ([Et2NSF2]BF4) salt was used for alcohol activation and tetra- ethyl ammonium halides ([Et4N]Hal) were used as source of nucleophile. Mostly, benzylic alcohols were used for bromination with the [Et2NSF2]BF4 and ([Et4N]Br system, which was used in an excess of 1.5 equivalents. The reactions were performed in dichloromethane and at ambient conditions for 12 hours. The bromination experiments were characterized by high activity and good product yields (67-92 % isolated yield) with high product purity (only trace amounts of fluorinated by-products mentioned). Based on their mechanistic investigations, the authors described the formation of the intermediate Et2NSF2Br after nucleophile addition. The substrate (alcohol) attacks the sulfur, which subsequently leads to the formation of a S-O bond. In this step, a proton and a bromide is eliminated. The bromide cleaves the C-O bond in the substrate to give the desired product. However, this reaction produces one equivalent of Et2NSO2H, HF and [Et4N]BF4 as by-products per molecule of organohalide and therefore displays low atom economy.

In 1992, Mas and Metivier described another method using a S-O bond for OH-activation. The investigated substrates were tetrahydropyran-2-methanol, pentan-2-ol and 44

D-methyl lactate. Thionyl chloride and gaseous HBr (HBr(g)) were used as bromination reagents. For the production of methyl 2-bromopropionate, methyl lactate was added to an excess of thionyl chloride (1.5 eq.) over one hour at ambient conditions.

After addition of a base (pyridine), HBr(g) is bubbled through the reaction mixture for

55 min before the system is heated to 80 °C for two hours. The reported yield of methyl 2-bromopropionate was 84 %. The only reported by-product of the reaction was chloropropionate (3 %). Over the course of the reaction, an intermediate (ROSOCl) is formed by HCl elimination. The addition of HBr replaces the second chloride to give

ROSOBr under formation of another HCl molecule. After base addition, SO2 is eliminated yielding the desired brominated product. The fact that all formed by-products are gaseous compounds simplifies the product isolation. The authors further claim this method to be applicable for large-scale application, although equimolar use of thionyl chloride and downstream of gaseous and toxic by-products seems cost-intensive and complex.

Poirier et al. investigated the influence of boron tribromide (BBr3) on primary, secondary and tertiary alcohols.[116] The experiments were conducted in dry methanol at a temperature of 0 °C and with an excess of the bromination agent of 1.2-3.2 equivalents. It has been found that BBr3 was a highly active and selective bromination agent for tertiary alcohols, with full conversion and 100 % product yield within 15 min. In contrast, this method showed to be not suitable for primary alcohols. The bromination of secondary alcohols suffered from reduced bromination selectivity. Moreover, it should be noted that the use of BBr3 as bromination agent leads to the formation of boric acid (H3BO3) as waste product and therefore displays low atom economy.

Hence, the method may be suitable in fine chemical synthesis but not in the sustainable production strain of platform chemicals.

Although different methods for the targeted bromination of LA are known in principle and have been reviewed above, most of the described methods do not operate in the framework of a sustainable production strain for 2-BrPA: Mild and liquid-phase reaction conditions, use of cheap and, as far as possible, non-toxic bromination reagents, and the bromination process being atom economical as well as applicable in large-scale. HBr as bromination reagent bears the highest potential to fulfill the above 2 Fundamentals 45

stated requirements. HBr is a cheap and readily available substrate and bromination reactions using HBr are further characterized by H2O being the sole by-product. Alt- hough the reaction of LA and HBr was described by Kekulé over 150 years ago, no further information was provided about bromination performance, optimum bromination conditions and related limitations.

2.3.3 The role of 3-bromopropionic acid en route to bio-AA

Up until now, 3-BrPA has not been given so much coverage in the literature. In principle, some quite historical methods for the production of 3-BrPA are known and have been described. 3-BrPA can, for instance, be synthesized from 3-hydroxypropaneni- trile (ethylene cyanohydrin) by reflux with HBr.[11e] The substrate 3-hydroxypropane- nitrile can be obtained from 2-chlorothanol, which is produced from ethylene and hypochlorous acid, and sodium cyanide.[11f] Further starting materials for 3-BrPA production are e.g. 3-bromopropanal and 3-iodopropionic acid.[11g] Moreover, 3-BrPA can be produced form AA and HBr, which plays a crucial role for the present thesis.[11d]

Obviously, 3-BrPA produced from AA and HBr is not useful in the context of the development of a bio-based AA production strain. However, 3-BrPA formation can be expected to result from the addition of HBr to the vinyl functionality of AA (see

Scheme 16). In particular, HBr-enriched reaction media may cause a consecutive reaction of AA with HBr. Hydrobromination of AA proceeds in anti-Markovnikov addition fashion, which is mainly due to the strong inductive effect of the carboxyl group.[117]

Scheme 16: Hydrobromination of AA via addition of HBr.

3-BrPA formed from AA and HBr can further be seen as another intermediate en route to bio-AA, as it could theoretically be re-converted to AA by dehydrobromination.

Dehydrobromination is a type of dehydrohalogenation reaction, involving the elimination of a hydrogen halide (in this case HBr) from a substrate (3-BrPA), which is associated with e.g. the formation of a double bond-containing product (AA). The

β-elimination of a hydrogen halide is a base-assisted reaction, usually promoted by strong bases like potassium hydroxide or potassium tert-butoxide. [117b] The β-elimination of HBr from 3-BrPA is simplified shown in Scheme 17.

Scheme 17: Base-promoted dehydrobromination of 3-BrPA to AA.

As can be seen in Scheme 17, the used base neutralizes 3-BrPA in a first step, leading to the formation of 3-bromopropionate. Elimination of a bromide anion results in the formation of AA. In principle, two mechanisms of elimination are known, namely E1 and E2. In the E1 elimination mechanism (unimolecular), the elimination of HBr would proceed in two steps: In the first “ionization” step, the C-Br bond breaks and a carbocation intermediate is formed. The carbocation is deprotonated in the second step. Due to the share of a common carbocationic intermediate, E1 competes with SN1 reactions. In contrast, E2 elimination (bimolecular) is a one-step elimination mechanism with a single transition state, in which C-H and C-Br bonds concertedly break to form the double bond (AA). Here, it is reasonable to assume that 3-BrPA elimination follows the E2 mechanism, due to the relatively acidic α-protons of 3-BrPA (activated by carboxylic group) and the reaction conditions applied in this work (see chapter 3.1.5 and 4.2.3).[117a, 118]

On industrial scale, base-assisted dehydrobromination of 3-BrPA as well as base-assisted dehydrohalogenation in general, is unwanted and problematic. The main reason for this is the disposal of large amounts of e.g. alkali halides involved in the base- assisted dehydrohalogenation reaction. In most processes, recovery of the formed salt 2 Fundamentals 47

is restricted, meaning that the halogen value cannot be recovered in a usable form and therefore strongly limits the atom efficiency of such processes. One possibility to overcome this limitation is the thermal cracking of the substrate, which is for instance applied in the production of vinyl chloride from ethylene dichloride. In this example, the formed HCl can be reused to produce more ethylene dichloride from ethylene via oxychlorination.[119]

For potential use in a sustainable production of bio-AA, dehydrobromination of

3-BrPA to AA needs to meet some prerequisites: In addition to being efficient and selective, dehydrobromination of 3-BrPA needs to be atom efficient. In this context, recyclability of the HBr-elimination promoting reaction medium needs to be man- aged. This goes hand in hand with the recovery of HBr. Moreover, safe-handling of HBr must be given.

2.4 Ionic liquids and their role in bio-AA production from LA

Ionic liquids (ILs) are a class of compounds completely composed of ions and further characterized by low melting points (< 100 °C), which distinguishes these substances from classical molten salts.[120] ILs consist of an organic cation and either an organic or inorganic anion, resulting in numerous possible combinations, most of which are characterized by unique physicochemical properties.[120] Hence, skillful combination of different cations and anions can influence IL’s properties such as e.g. viscosity or solubility, making them adaptable for different applications (ILs are so-called de- signer solvents). One noteworthy property ILs have in common is the negligibly small vapor pressure (approx. 10-12 mbar at room temperature)[120]. Further interesting general properties of ILs are non-flammability, high thermal as well as chemical stability and excellent dissolving power for many organic and inorganic compounds. Today, more than a century after the first discovery of ILs (ethanolammonium nitrate, Gabriel and Weiner)[121] and room temperature ILs (ethylammonium nitrate, Walden)[122], many of these compounds are commercially available. A selection of important cations and anions in terms of IL chemistry are shown in Figure 3. Moreover, ILs have recently attracted the interest of the scientific community due to their well-known and interesting properties, offering immense opportunities as, for example, catalysts and novel solvents for chemical transformations.[12b, 12c, 123] Further and much more extensive reviews on ILs in general, their properties and applications are, for instance, given by

Hallett and Welton[12a, 12b]. or Wasserscheid and Keim[12c]. 2 Fundamentals 49

R1 R1 R1 N+ + + P N N+ (i) R1 R2 R2 N + R4 R4 N R R3 R3 2 R1 R2 tetraalkylphosphonium tetraalkylammonium imidazolium pyridinium pyrrolidinium

F F O O – F F N – – – – B– P (ii) Cl Br I S S F F F F F3C CF 3 O O F F halides bis(trifluorosulfonyl)imide tetrafluoroborate hexafluorophosphate

Figure 3: Selection of common cations (i) and anions (ii) in terms of IL chemistry (Rn represent alkyl residues, which can be of same or different chain length).

In the context of this thesis, especially bromide ILs (ILs with a bromide anion) were used as reaction medium and/or reactant for several chemical reactions, like dehydration of LA to AA, bromination of LA to 2-BrPA or isomerization of 2-BrPA to 3-BrPA.

Therefore, the focus of the following part is on selected examples of bromide ILs.

Firstly, the frequently used IL tetrabutylphosphonium bromide ([PBu4]Br) is introduced in terms of its properties, production strain and so far known applications.

Subsequently, the use of IL in the field of bromination reactions is reviewed. Special attention is devoted to a HBr-carrying IL system.

2.4.1 ILs as reaction medium for liquid-phase dehydration of LA: Tetrabutylphosphonium bromide

In the development of a liquid-phase process for the production of bio-AA presented here, ILs are used as reaction medium for e.g. dehydration of LA. In this context, one takes advantage of the above described and promising properties of ILs. ILs with a bromide anion are of particular relevance within the developed production route of bio-AA from LA.

The most relevant and representative example of the used and investigated ILs, namely tetrabutylphosphonium bromide ([PBu4]Br), is a hygroscopic and crystalline substance at ambient conditions. The symmetrical phosphonium cation PR4+ with its four butyl substituents is characterized by weak electrostatic interactions with anions 50 compared to other prominent cations, typically resulting in low melting points of phoshonium-based ILs. The lipophilicity of the cation results from non-delocalized and sterically hindered positive charge of the PR4+ cation. However, the accompany- ing bromide anion, in contrast, introduces high and unprotected charge to the IL, compromising the lipophilicity of the phosphonium cation. Moreover, due to the small size of the bromide anion, melting points of bromide ILs are comparably high.

In this specific case, the combination of the PR4+ cation and the bromide anion results in a melting point of 100 °C of [PBu4]Br. Thermal stability of [PBu4]Br has been proven up to 200 °C, which goes hand in hand with the reported thermal stability of the chloride equivalent tetrabutylphosphonium chloride ([PBu4]Cl) of up to 240 °C. All in all,

[PBu4]Br was reported to be liquid and stable between 100 and 200 °C, which is a typical temperature range for liquid-phase processes. [PBu4]Br is commercially available (e.g. Sigma-Aldrich Chemie GmbH).[120, 124]

[PBu4]Br is typically prepared by quaternization of tributylphosphine using 1-bromobutane, conducted at a reaction temperature of approximately 150 °C (see Scheme

18). Stalpaert et al. reported contaminations of HBr and 1-butene in commercially available [PBu4]Br, resulting from side-reactions of the production route. These contaminations lead to a distinct acidity of [PBu4]Br (due to free HBr), if the IL is not further neutralized by e.g. treatment with tetrabutylphosphonium hydroxide

([PBu4]OH).[124c, 125]

Scheme 18: Quaternization reaction of tributylphosphine and 1-bromobutane to [PBu4]Br; adapted from Stalpaert et al. and Bradaric et al. [124c, 125]

[PBu4]Br is, for instance, investigated and used in extraction tasks (e.g. liquid-liquid extraction) and has been studied with respect to the related phase behavior and equilibria studies. Kareem et al. reported its application for the removal of aromatic hydro- carbons from naphtha.[126] In their study, they introduced [PBu4]Br-based deep eutectic solvents (DESs) to the model system of toluene/heptane (aromatics/aliphatics). The 2 Fundamentals 51

DESs consisted of [PBu4]Br as salt and either ethylene glycol or sulfolane as hydrogen bond donor (HBD). The novel DES showed high selectivity and distribution coefficients at low toluene concentrations, giving the possibility for applying the

[PBu4]Br-based DES in industry to remove traces of aromatics from e.g. naphtha prior to thermal cracking. Moreover, the presented DES were easily prepared, reducing the synthesis and purification costs of commercially used extracting agents. In addition,

[PBu4]Br is occasionally mentioned in literature in the context of ionic and mixed sem- iclathrate hydrates. [127]

Moreover, there are a few examples described in literature where [PBu4]Br is used in synthesis and catalysis. Knifton investigated the preparation of ethylene glycol directly form synthesis gas via ruthenium “melt” catalysis.[128] A catalyst system was described where the ruthenium source (e.g. Ru(V)oxide) was dispersed in a quater- nary phosphonium or ammonium salt, such as [PBu4]Br. Here, the used salt provided a polar and fluid medium for solubilization of the active catalyst and therefore ef- fected the desired conversion of synthesis gas to ethylene glycol. This phenomenon may overcome the limitations of the usually applied, high pressure and homogene- ously catalyzed reactions. In 2017, Stalpaert et al. investigated [PBu4]Br as solvent and catalyst for the dehydration of diols to conjugated dienes.[124c] The experiments were conducted in sealed glass vials at a reaction temperature of 200 °C. In addition to the investigated substrate, [PBu4]Br, HBr and other additives, an internal standard (e.g. mesitylene)was added prior to the start of the reaction for quantification purposes.

The authors postulated and experimentally corroborated a reaction mechanism for the model compound 1,2-hexanediol, allowing for the selective transformation of diols to conjugated dienes. In contrast, selective formation of conjugated dienes was not achieved with purely acidic catalysts. Furthermore, Stalpaert et al. assessed the bio-based production of 1,3-butadiene from 1,4-butanediol as first application of the process. In this example, 94 % product yield (1,3-butadiene) were obtained at full substrate conversion. 52

2.4.2 Ionic liquids as reaction medium for bromination of LA

Among the reported ionic liquid-based innovations, new methods for the halogenation of alcohols and diols appear particularly promising because halide ILs have been reported as excellent halogenating agents acting at the same time as solvents and vapor pressure reducing agents for the applied or formed hydrogen halide acids.[129] These methods can overcome some disadvantages of traditional bromination methods for alcohols, including corrosive and toxic halogenation agents, difficult product separation, and low atom economy.[109, 115-116, 130] For instance, BASF[131] (chlorination of butane-1,4-diol) and SOLVAY[132] (hydrochlorination of acetylene) already apply ILs in industrial halogenation processes.

As briefly mentioned in chapter 2.3.2, HBr, which was determined to be the most practicable bromination reagent for LA within a green production strain of AA, can be provided from two different sources. One molecule carries an acidic proton, the other provides the nucleophilic bromide. Ren et al. described the use of ILs, namely

1-n-buty-3-methylimidazolium halides ([BMIM]Hal) for the halogenation of different alcohols.[129e] Butanol and octanol were used as substrates. The IL served as solvent and was simultaneously used as a source of halide. Additionally, different (in)organic acids were used (HCl, H2SO4, CH3SO3H). The reactions were conducted at room temperature for 24 hours. Depending on the respective halide anion and acid employed, obtained product yields varied widely. However, substitutions with H2SO4 and bromide gave almost quantitative product yields (> 95 %). However, the use of

[BMIM]Hal together with the investigated acids (HA) was accompanied by the formation of [BMIM]A and no information on the recyclability of this species was provided. Leadbeater et al. studied the influence of microwave radiation on comparable bromination agents.[129b] They combined para-toluenesulfonic acid or H2SO4 with

1-methyl-3-alkylimidazolium halides. The used bromination agent was 1-methyl-3-isopropylimidazolium bromide. Different alcoholic substrates were investigated, including diols, olefinic-, allylic- and benzylic aclohols. The experiments were conducted under microwave irradiation at 200 °C and short reaction times (up to

10 min). When linear alcohols were used as substrates, product yields of over 95 % were observed. Higher functionalized alcohols gave noticeably less product yields or 2 Fundamentals 53

decomposed under the applied reaction conditions. In contrast to the method described by Ren et al., microwave irradiation was not applicable to secondary or tertiary alcohols (substrate decomposition was observed).

In 2009, Gupta et al. investigated the bromination of alcohols with ILs.[129a] In contrast to previously presented studies, the bromination was performed using an acidic IL

([BMIM]HSO4) as protonation reagent and sodium halide (e.g. NaBr) as halogen source. Mostly benzylic alcohols were used as substrates for the reaction, which was conducted at 80-90 C. Moreover, they compared a conventional heating method to microwave irradiation. Conventionally heated experiments (tr = 15-20 hours) showed product yields of 55-78 %, while microwave experiments resulted in product yields of 70-90 %. After the experiment, the reaction mixture contained the reaction products of the ILs [BMIM]SO4 and [BMIM]OH. Although the authors claimed successful regeneration of the IL by addition of NaHSO4 and subsequent extraction with dichloromethane, the obtained yield with a re-used IL dropped to 40 %.

Another method using ILs and HBr for the bromination of alcohols was reported by

Ranu et al. in 2005.[129d] Various substrates were investigated in the presence of alkyl halides under sonification at room temperature (1-2 hours). The used bromination agent was tert-butyl bromide. The reactions were conducted using 1-n-pentyl-3-methylimidazolium bromide ([PMIM]Br) as solvent. The authors demonstrated that the used alkyl halide eliminated one molecule of HBr under sonification conditions, serving as bromination reagent. The in-situ generation of HBr worked best when the IL bromination was performed in an IL compared to pure mixtures of alcoholic substrate and tert-butyl bromide. In most organic solvents, no reaction was observed. The reported yields were in the range of 78 to 95 %, even fore sensitive substrates like (thi- ophen-2-yl) methanol. No further information was provided on product selectivity.

Although providing high product yields, this method seems to be limited to small-scale preparation of high-value products, due to the relatively high price of tert-butyl halides.

In 2011, Li et al.[133] reported the use of a zwitterionic HBr-carrier system as bromination medium for 1,7-heptanediol. In the related work, the bromide salt resulting from 54 the addition of HBr to the zwitterion acts as catalyst, brominating agent, and solvent.[133] More importantly, the applied HBr is chemically bound to the zwitterion which greatly reduces HBr vapor pressure and makes the system relatively easy to handle. Zwitterions are neutral molecules bearing both a positive and a negative charge locally separated from each other. The zwitterions used in this study had a positively charged imidazolium backbone and a negative charge of a sulfonyl group, connected to the backbone by a C3 or C4 alkyl chain. Upon addition of HBr(aq), the sulfonyl group is protonated, leaving a positive charge on the imidazolium ring. The bromide serves as counter ion. In contrast to all previously described bromination methods, the formed IL provides both bromination functionalities within a single molecule: The sulfonic acid serves as proton source, acidic enough to protonate the alcohol group of the substrate. The bromide displayes nucleophilic character and sub- stitutes the activated alcohol. The reactions were conducted under inert gas and for

1-2 hours at a reaction temperature of 100 °C. Moreover, the IL serves as solvent in addition to being the brominating agent and was used in an excess of up to 3 equivalents. Reported yield of 1,7-dibromoheptane was 95 %. After the bromination reaction, HBr-loaded and un-loaded zwitterion as well as the bromination product and water (by-product) were observed. Furthermore, IL regeneration by addition of

HBr(aq) was demonstrated. After extraction of the HBr-loaded zwitterion with e.g. dichloromethane and re-use of the IL in the bromination of 1,7-heptanediol, a constant product yield of 94 % was observed over the first five cycles. High product yield and selectivity, as well as possible recycling of the HBr-carrying zwitterion and the reduced HBr vapor pressure make this approach an elegant and effective method for the bromination of alcohols. The above described bromination method with HBr-carrying zwitterions is illustrated in Scheme 19.

Scheme 19: Alcohol bromination using HBr-carrying zwitterions, adapted from Li et al.[133] R = H, alkyl, aryl, etc; n = 3,4.

2 Fundamentals 55

2.5 Objective of this thesis

Basic aim of this work was the development of a liquid-phase dehydration process for the production of bio-AA from fermentatively derived LA to provide a technical as well as economical feasible alternative to the incumbent petro-based route. AA production from biogenic LA feedstock is expected to be safer than the petro-based production route (high exothermicity of gas-phase oxidation of propene, flammability of feed stream that contains propene and air, acrolein being a carcinogenic intermediate). Furthermore, an alternative and sustainable production from LA represents a process that produces AA with better environmental attributes, such as lower non- renewable energy use (NREU) and CO2 footprint. However, it becomes obvious from the described state-of-the-art technologies that the dehydration of LA to bio-AA has been an elusive and 60-year old problem. Several groups have attempted to develop efficient dehydration catalysts and process conditions for the conversion of LA to bio-AA, especially in the gas phase. However, most attempts suffered either from low yield and selectivity towards bio-AA or short catalyst lifetime.

The renewable production of AA from LA in the liquid phase is expected to offer economic and LCA advantages over the gas-phase dehydration of LA because of ben- eficial process conditions (e.g. lower temperature and pressure). Moreover, the significantly lower reaction temperature may reduce the corrosion potential and the number of by-products, which would enormously simplify the downstream processing.

On the basis of the developed liquid-phase technology for the dehydration of LA and

LA derivatives towards bio-AA (The Erlanger “NADA” process), LA should first be evaluated as substrate for the dehydration process, especially with regard to dehydration selectivity. Based on the obtained results and the postulated reaction mechanism, the process was split into a spatially separated reaction sequence of the

“NADA” process. The individual reactions, namely bromination of LA, isomerization of 2-BrPA to 3-BrPA and dehydrobromination of 3-BrPA to AA should be developed, respectively, and the corresponding reaction media and conditions should be optimized. After realizing a separated and multi-step approach from LA towards bio-AA, 56 whose development involved several insights into the functionality and chemical requirements for an efficient liquid-phase dehydration of LA to bio-AA, possibilities for potential process shortcuts should finally be examined.

3 Experimental section 57

3 Experimental section

In the following, used experimental procedures and corresponding setups as well as characterization techniques are described in detail and in chronological order. In addition, commonly employed analytical methods and respective calculations are presented.

Bio-based feedstock material produced naturally by fermenting carbohydrates, namely PURAC® (L-lactic acid, 88wt% in water) and PURALACT® (L-lactide, polymer grade), were purchased from Corbion. All further used chemicals were purchased from commercial suppliers and used without further purification. Not commercially available substances were synthesized from purchased compounds within this thesis (see given synthesis instructions in Appendix, chapter 7.1.2).

Due to the (highly) corrosive nature of most investigated reaction mixtures, thermal control was largely ensured via a RS 206-3738 digital thermometer, connected to a

Ni/Cr thermal element, enclosed by a small glass tube/PTFE coating and inserted into the reaction medium through a silicone/PTFE plug.

3.1 Experimental procedures

3.1.1 Benchmark one-step dehydration of LA in the NADA molten salt reaction matrix

Dehydration of LA was conducted in a 100 mL three-necked glass reactor. Tetrabu- tylphosphonium bromide ([PBu4]Br, 100 mmol, 98%; Sigma-Aldrich Chemie GmbH,

Taufkirchen, Germany; catalog # 189138) was mixed with lactic acid (88wt% LA in water, 50 mmol; Corbion Purac Co., Lenexa, KS) and 2-BrPA (99%, 5 mmol; Sigma-Al- drich Chemie GmbH, Taufkirchen, Germany; catalog # B78300) at room temperature and atmospheric conditions. The reaction mixture was heated to 150 °C under continuous stirring with a magnetic stirring bar at 500 rpm. The reaction mixture was batchwise refluxed and gaseous by-products were collected in a gas sampling bag for qualitative analysis, collected in a hydrostatic column for quantification, or routed to 58 the off-gas. After an overall reaction time of 168 h, the hot reaction mixture was allowed to cool down to room temperature and was analyzed via off-line 1H NMR for quantification.

3.1.2 Bromination reactions in aqueous reaction medium

3.1.2.1 Bromination with HBr(aq)

The bromination of lactide or lactic acid in aqueous reaction medium was conducted in a 100 mL three-necked glass reactor. Lactide (L- lactide; polymer grade; Corbion Purac Co., Lenexa, KS) or lactic acid (88wt% LA in water; Corbion Purac Co., Lenexa,

KS) was mixed with aqueous hydrobromic acid (HBr(aq); 48 %; Sigma-Aldrich, Chemie

GmbH, Taufkirchen, Germany; catalog # 244260) in the desired molar ratio (molar ratios for LA:HBr ranged from 6:1 to 1:12). The aqueous reaction mixture was heated to the desired reaction temperature under continuous stirring with a magnetic stirring bar at 500 rpm. The reaction mixture was batchwise refluxed and gaseous by-products were collected in a gas sampling bag for qualitative analysis, collected in a hydrostatic column for quantification, or routed to the off-gas. After an overall reaction time of 1 to 72 h, the hot solution was allowed to cool down to room temperature.

Precipitated polymer was filtered off. The aqueous reaction phase was analyzed via off-line 1H NMR for quantification.

3.1.2.2 Extractor screening for aqueous bromination medium

To investigate liquid-liquid extraction capabilities of different organic solvents, an ex-situ screening of six potential extractors was conducted at room temperature. After determination of the partition coefficient, the four most promising extractors were tested in-situ under aqueous reaction conditions. The ex-situ LLE screening was conducted in 15 mL glass vials. 2.0 g (approximately 1.35 mL) of a simulated product solution containing 3.60 g lactide (25 mmol) and 7.65 g 2-BrPA (50 mmol), dissolved in 50.57 g of aqueous HBr (300 mmol, 48 wt%) were mixed with 2 mL of an organic solvent. The vials were sealed and vigorously stirred with a magnetic stirring bar for

24 h. The phases were separated with a glass pipette and an internal 1H NMR standard was added to both phases. Acetic acid (CH3COOH; 100 %; Merck Schuchardt OHG, 3 Experimental section 59

Hohenbrunn, Germany; catalog # 100063) served as standard for the aqueous phase, mesitylene (C6H3(CH3)3; 98%; Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany; catalog # M7200) for the organic phase. The aqueous as well as the organic phase were both analyzed via off-line 1H NMR for quantification.

3.1.2.3 Bromination in biphasic reaction medium / In-situ LLE

The bromination of lactide or lactic acid in biphasic reaction medium with an organic extractor was conducted in a 100 mL three-necked glass reactor. Lactide (L, L- lactide; polymer grade; Corbion Purac Co., Lenexa, KS) or lactic acid (88wt% LA in water; Cor- bion Purac Co., Lenexa, KS) was mixed with aqueous hydrobromic acid (HBr(aq); 48 %;

Sigma-Aldrich, Chemie GmbH, Taufkirchen, Germany; catalog # 244260) and a suitable organic solvent (equimolar amount). The biphasic reaction mixture was heated to the desired reaction temperature under continuous stirring with a magnetic stirring bar at 500 rpm. The reaction mixture was batchwise refluxed and gaseous by-products were collected in a gas sampling bag for qualitative analysis, collected in a hydrostatic column for quantification, or routed to the off-gas. After an overall reaction time of 1 to 72 h, the hot solution was allowed to cool down to room temperature and transferred into a separation funnel. The phases were separated and approximately 1 mL of an internal 1H-NMR standard was added. Acetic acid (CH3COOH; 100 %; Merck

Schuchardt OHG, Hohenbrunn, Germany; catalog # 100063) served as standard for the aqueous phase, mesitylene (C6H3(CH3)3; 98%; Sigma-Aldrich Chemie GmbH, Taufkir- chen, Germany; catalog # M7200) for the organic phase. The aqueous as well as the organic phase were both analyzed via off-line 1H NMR for quantification.

3.1.2.4 Bromination in aqueous reaction medium with ionic liquid additive

The bromination of lactide or lactic acid in aqueous reaction medium with an ionic liquid additive (tetrabutylphosphonium bromide, [PBu4]Br) was conducted in a

100 mL three-necked glass reactor. Lactide (L- lactide; polymer grade; Corbion Purac Co., Lenexa, KS) or lactic acid (88wt% LA in water; Corbion Purac Co., Lenexa, KS) was mixed with aqueous hydrobromic acid (HBr(aq); 48 %; Sigma-Aldrich, Chemie GmbH,

Taufkirchen, Germany; catalog # 244260) and tetrabutylphosphonium bromide 60

([PBu4]Br; 98%; Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany; catalog # 189138) in the desired molar ratio (molar ratios for LA:HBr:[PBu4]Br ranged from 1:1:1 to

1:3:10). The aqueous IL-containing reaction mixture was heated to the desired reaction temperature under continuous stirring with a magnetic stirring bar at 500 rpm.

The reaction mixture was batchwise refluxed and gaseous by-products were collected in a gas sampling bag for qualitative analysis, collected in a hydrostatic column for quantification, or routed to the off-gas. After an overall reaction time of 1 to 72 h, the hot solution was allowed to cool down to room temperature. The molten salt phase was analyzed via off-line 1H NMR for quantification.

3.1.3 Bromination with zwitterionic HBr carriers

3.1.3.1 Preparation of zwitterions

All zwitterions were prepared via altered synthetic routes based on protocols by Yo- shizawa et al.[134] and Cole et al.[135]. For zwitterion preparation, equimolar amounts of the desired Lewis-basic backbone (imidazole, pyridine, phosphine) and the desired alkylation agent (1,4-butanesultone or 1,3-propanesultone) were dissolved in acetonitrile and refluxed for 48 h. The product precipitated over course of the reaction. After cooling to room temperature, the residue was thoroughly washed and dried in vacuum. Products were analyzed via 1H and 13C NMR spectroscopy. All products were obtained as white powders in very good yield and excellent purity. Details can be found in the Appendix (chapter 7.1.2).

3.1.3.2 Loading of zwitterions with HBr

The ionic liquid bromination medium was prepared by first mixing equimolar amounts of the desired zwitterion (75 mmol) and 48 wt.% hydrobromic acid

(HBr(aq), 75 mmol) in a 100 mL three-necked glass reactor. To facilitate stirring and ensure homogeneity of the reaction mixture, two additional milliliters of water were added. After vigorous stirring for 15 minutes, a clear and colorless solution was obtained. 50 mL of cyclohexane were added at room temperature and atmospheric conditions. The biphasic reaction mixture was heated to a temperature of approx. 70 °C under continuous stirring with a magnetic stirring bar at 600 rpm for 16 to 24 hours. 3 Experimental section 61

The protic and HBr-loaded ionic liquid was obtained by removing the water using a

Dean-Stark-apparatus and finally decanting the cyclohexane phase off, after the reactor was cooled down to room temperature. All active HBr-loaded ionic liquids were obtained as highly viscous and hygroscopic oils.

3.1.3.3 Bromination with HBr-loaded ionic liquids

The bromination of lactide or lactic acid with HBr-carrying zwitterions was conducted in a 100 mL three-necked glass reactor (see Figure 4). Lactide (L-lactide, polymer grade, Corbion Purac Co., Lenexa, KS) or lactic acid (88wt% LA in water, Corbion

Purac Co., Lenexa, KS) was added to the HBr-loaded zwitterion in the desired molar ratio. The amount of HBr can be readjusted (ntitrated NaOH) by addition of aqueous hydrobromic acid (HBr, 48 wt%, Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany, catalog # 244260). The reaction mixture was heated to the desired reaction temperature under continuous stirring with a mechanical overhead stirrer at 600 rpm. The reaction mixture was batchwise refluxed and gaseous by-products were collected in a gas sampling bag for qualitative analysis, collected in a hydrostatic column for quantification, or routed to the off-gas (regulated with valve V-1). After an overall process time of 1 to 24 h, the reaction mixture was allowed to cool down, quenched with methanol (CH3OH, 99.8 %, anhydrous, Sigma-Aldrich, Chemie GmbH, Taufkir- chen, Germany; catalog # 322415) and analyzed via off-line 1H NMR for quantification. 62

off-gas

offline V-1 condenser GC analysis

gas sampling bag

MIC mechanical stirrer

hydrostatic column

TIC

bromination reactor

Figure 4: Flowsheet of batch reactor for bromination of LA and lactide.

3.1.3.4 LLE screening experiments for recycling of zwitterions and 2-BrPA isolation

The ex-situ LLE screening was conducted in 15 mL glass vials. 1.0 mL of a quenched product solution (75 mmol [MIMBS]Br, 12.5 mmol lactide, 120°C, 5 h) was mixed with 1.0 mL of an organic solvent. The vials were sealed and vigorously stirred with a magnetic stirring bar for 24 h at room temperature. The phases were separated with a glass pipette and an internal 1H NMR standard was added to both phases. Acetic 3 Experimental section 63

acid (CH3COOH; 100%; Merck Schuchardt OHG, Hohenbrunn, Germany; catalog # 100063) served as standard for the aqueous phase, mesitylene (C6H3(CH3)3; 98%;

Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany; catalog # M7200) for the organic phase. Both phases were analyzed via off-line 1H NMR for quantification.

3.1.3.5 Recycling of zwitterions

After completion of the bromination reaction, the reaction mixture was quenched with 10 mL of water. When cooled to room temperature, the quenched reaction medium is extracted (3x20 mL) with an organic solvent to remove the product from the aqueous phase. The combined organic phases as well as the aqueous phase are quantitatively analyzed by 1H NMR spectroscopy to give an isolated product yield (YAA, isolated) and to determine the amount of non-isolated C3-compounds. Finally, the aqueous IL phase is reused in the HBr loading experiment to start a further bromination cycle.

3.1.4 Isomerization reaction in ionic liquid reaction matrix

The isomerization of 2-BrPA in an ionic liquid reaction matrix was conducted in a

100 mL three-necked glass reactor. The IL, e.g. tetrabutylphosphonium bromide ([PBu4]Br, 98%; Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany; catalog #

189138), was mixed with 2-BrPA (99%; Sigma-Aldrich Chemie GmbH, Taufkirchen, Ger- many; catalog # B78300) at room temperature and atmospheric conditions. The isomerization mixture was heated to the desired reaction temperature under continuous stirring with a magnetic stirring bar at 400-500 rpm. The isomerization mixture was batchwise refluxed and gaseous by-products were collected in a gas sampling bag for qualitative analysis, collected in a hydrostatic column for quantification, or routed to the off-gas. After an overall reaction time of 1 to 72 h, the hot reaction mixture was allowed to cool down to room temperature and was analyzed via off-line 1H NMR for quantification.

Optionally, after the above described isomerization experiment, the reaction mixture was allowed to cool down to 100 °C and was analyzed via off-line 1H NMR. Subse- 64 quently, the adapted isolation unit was connected to the above described experimental setup using valve V-1 (see Figure 5). After the reaction mixture reached a constant temperature of 100 °C, the vacuum distillation experiment was started at reduced pressure of 2-10 mbar. The reaction mixture was stepwise heated to a reaction temperature of 160 °C (temperature ramp with 20 °C steps held for 30 min respectively) under continuous stirring with a magnetic stirrer at 800 rpm. Liquid products removed from the reaction mixture were condensed in a distillation flask (0 to-50°C) and gaseous by-products were condensed in a cooling trap (-197 C). After an overall distillation time of 2 hours the collected distillate as well as the purified reaction matrix were analyzed via off-line 1H NMR.

batchwise isomerization operation off-gas product isolation operation

V-1 offline V-2 GC analysis

gas sampling bag

condenser hydrostatic column

off-gas

TIC membrane pump distillation flask cooling trap 1 - 1000 mbar 0 to -50 °C -197 °C product heating mantle collection

magnetic stirrer

isomerization reactor

Figure 5: Flowsheet of batch reactor for isomerization of 2-BrPA and subsequent product isolation via distillation at reduced pressure.

3 Experimental section 65

3.1.5 3-BrPA conversion in TOA

The conversion of 3-BrPA to AA in presence of base was conducted in a 100 mL three- necked glass reactor. Trioctylamine ([CH3(CH2)7]3N, 98 %; Sigma-Aldrich Chemie

GmbH, Taufkirchen, Germany; catalog # T81000) was mixed with 3-BrPA (97 %;

Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany; catalog # 101281) at room temperature and atmospheric conditions. The reaction mixture was heated to the desired reaction temperature under continuous stirring with a magnetic stirring bar at 400-

500 rpm. The reaction mixture was batchwise refluxed and gaseous by-products were collected in a gas sampling bag for qualitative analysis, collected in a hydrostatic column for quantification, or routed to the off-gas. Optionally, products were semi- batchwise removed under reduced pressure (90-100 mbar) or by Ar strip gas

(50-1000 mL/min). Liquid products were condensed and collected in a cooled flask whereas gaseous by-products were collected in a gas sampling bag for qualitative analysis, collected in a hydrostatic column for quantification, or routed to the off-gas. After an overall reaction time of 0.5 to 5 h, the reaction mixture and the collected distillate were analyzed via off-line 1H NMR for quantification.

3.1.6 Thermally-induced HBr recovery from [TOAH]Br

Thermally-induced HBr recovery from [TOAH]Br was conducted in a 100 mL three- necked glass reactor. [TOAH]Br (100 mmol) was slowly heated to 100 °C under atmospheric pressure. The molten salt was heated progressively to higher reaction temperatures (150-200 °C) under continuous stirring with a magnetic stirring bar at

500 rpm and gaseous HBr was semi-batchwise removed with a flow of argon strip gas (50-1000 mL/min). HBr recovery was determined by pH monitoring of a downstream washing flask. The gas stream is passed through water via a frit and formed

HBr is absorbed in the liquid. Subsequent acid-base titration of the HBr-enriched aqueous solution (50 mL H2O) was used to verify the trapped HBr amount. Aqueous

NaOH solution (1 M) was dropwise added until the acidic solution was neutralized. This was followed by color-change of the previously added ethanolic phenolphthalein indicator solution (0.1 mL). 66

3.1.7 HBr recovery from [TOAH]Br via extraction

HBr recovery from [TOAH]Br via extraction was performed in a 100 mL three-necked glass reactor. [TOAH]Br (25 mmol) was mixed with 50 mL water. The biphasic reaction mixture was heated to the desired reaction temperature (25-100 °C) under continuous stirring with a magnetic stirring bar at 500 rpm. The reaction mixture was batchwise refluxed and gaseous by-products were collected in a gas sampling bag for qualitative analysis, collected in a hydrostatic column for quantification, or routed to the off-gas. After an overall reaction time of 24 h, the (cooled-down) reaction solution was transferred into a separation funnel. The phases were separated and subsequently both analyzed via off-line 1H NMR. Acid-base titration of the aqueous phase was used to determine the extracted amount of HBr. Aqueous NaOH solution (1 M) was dropwise added until the acidic solution was neutralized. This was followed by color-change of the previously added ethanolic phenolphthalein indicator solution (0.1 mL).

3.1.8 Semi-batchwise dehydrobromination of 2-BrPA to AA

The elimination of HBr from 2-BrPA in an ionic liquid reaction matrix was conducted in a 100 mL three-necked glass reactor (see Figure 6). Tetrabutylphosphonium bromide ([PBu4]Br, 100 mmol, 98 %; Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany; catalog # 189138) was mixed with 2-BrPA (100 mmol, 99 %; Sigma-Aldrich Chemie

GmbH, Taufkirchen, Germany; catalog # B78300) at room temperature and atmospheric conditions. The reaction mixture was heated to the desired reaction temperature (160-200 °C) under continuous stirring with a magnetic stirring bar at 500 rpm. The reaction products were semi-batchwise removed using a flow of argon strip gas

(500 mLn/min) or reduced pressure (5-25 mbar). Removed liquid compounds were optionally separated in a vigreux column, condensed and collected in a cooling trap

(0 to -60 °C; acetone-nitrogen mixture). Removed gaseous products (especially HBr) were either trapped in a second cooling trap (-196 °C; liquid N2) or absorbed in a pH monitored downstream scrubber unit. After an overall reaction time of 3 h, the hot reaction mixture was allowed to cool down to room temperature. Mesitylene

(C6H3(CH3)3; 98%; Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany; 3 Experimental section 67

catalog # M7200) and methanesulfonic acid (CH3SO3H; ≥99 %; Sigma-Aldrich Chemie

GmbH, Taufkirchen, Germany; catalog # 471356) were added to the collected distillate as internal standard and the collected distillate, as well as the reaction mixture, were both quantitatively analyzed via off-line 1H NMR (JEOL ECX 400 MHz). Collected gaseous products (HBr) were evaporated by progressive removal of the cooling unit.

Vaporized products were collected in a gas sampling bag for qualitative analysis via off-line GC, passed through a downstream washing flask including pH monitoring unit for quantification, or routed to the off-gas.

68 S t r i p

g N a 2 s /

A /

r p u s v r t a g r e c i p

u M

u g m a F C s

o & p

p e F d r u c v a e r o i t g g h l i e u r o e y

m n o u d m p h n x r e e

r o a r g t a e n b i n t e a i g r t o i c

o c m n

t s m a o t i n r r t r i l e n e r a t i o s n e p a r a t T i I o C n c o

u n d n e i n t s e r l i q c o u 0 o i

d t l i o n

g - r 6

a t 0 r p c

a ° p p p C o

i 1 r n l l o g e d c u t i c o g c t a o n

o s -

l 1 t i n r 9 g a 6

p t

° r p C a i p n

2 g m e 1 m H

– H 2 b

O 1 B r

0 w r a o 0

s n a 0 f c s f e

h m - r

g u i p n b a b u g a s b

r m f l e a p r s k H p B H I r g

o a r f s e f

- s c g a o a m s v p e l i r n y g

b a g G C o

a f f n l i a n l y e s

i s

Figure 6: Flowsheet of the experimental rig for semi-batchwise dehydrobromination of 2-BrPA. 3 Experimental section 69

3.2 Analysis, quantification and calculations

3.2.1 Characterization and quantification of liquid phase samples with NMR spectroscopy

All qualitative and quantitative NMR spectra were recorded on a JEOL ECX 400 MHz instrument with a sample temperature of approximately 20 °C. Chemical shifts were reported relative to the peak of SiME4 using 1H-(residual) chemical shifts of the deu- terated solvent as a secondary standard and are reported in ppm. 1H scans were conducted with an excitation frequency of 399.72 MHz and 8 scans. 13C NMR spectra were recorded with an excitation frequency of 100.51 MHz and 250 scans.

Conversion (�), yield (�) and selectivity (�) of all performed experiments were determined via the integrals of the corresponding peaks in the 1H NMR spectrum, using the following equations (with � = reactant, � = product, � = stoichiometric coefficient).

��,� − ��,� �� = ��,� (1)

�� − ��,� |��| � = ∙ � (2) ��,� ��

�� − ��,� |��| �� = ∙ = � (3) ��,� − ��,� ��

Quantification with 1H NMR was enabled by the use of an internal standard, which was added after completion of the reaction. Depending on the performed reaction and resulting product 1H NMR spectrum, different IS were necessary to ensure straight and unrestricted integration of internal standard proton signal(s). Typical compounds used in this work were acetic acid (AcOH), mesitylene and methanesulfonic acid (MSA). Additionally, proton signals of thermally stable ILs were used for quantification purposes (e.g. [PBu4]+ and MIMBS). 70

3.2.2 Characterization of biphasic reaction systems

The concentration ratio of a species i in the organic phase and the aqueous phase correspond to the partition coefficient Ki. The concentration ci can be obtained from the molar amount ni and the volume of the respective phase, whereby ni was obtained from 1H NMR spectroscopy using an internal standard (IS) integral for quantification purposes. The selectivity of the extracting agent can be obtained by the ratio of the respective partition coefficients. The following equations were used to determine the partition coefficient of the substrate LA (KLA), the product 2-BrPA (K2-BrPA), and the extraction selectivity towards 2-BrPA (SLLE,2-BrPA).

��(��) �� = �� (��) (4)

��(��) �� = (5) �� (��)

�� ,�� = (6) ��

3.2.3. Characterization of HBr-loaded ionic liquids

The HBr-loading of the ionic liquid was investigated by acid-base titration of the HBr- enriched water that was removed during the drying procedure. Aqueous NaOH solution (1 M) was added dropwise until the acidic solution was neutralized. This was followed by color-change of the previously added ethanolic phenolphthalein indicator solution (0.1 mL). As no HBr-loss into gas-phase was observed, acid-base titration of the trapped water allows stating the HBr-loading grade of the prepared ionic liquid:

�� = ��,�� − �� (7)

A reference experiment of direct HBr determination by acid-base titration of the solved and loaded IL resulted in a fully closed HBr balance (n0,HBr = nloaded,HBr + ntitrated,NaOH) and therefore confirmed the developed method. 3 Experimental section 71

The residual water content of the IL was analyzed via Karl-Fischer titration (Metrohm

756 KF Coulometer, the single-component solution used for the analysis was apura®

CombiCoulomat fritless Karl Fischer reagent for coulometric water determination,

Merck) in dry methanol. The standard loading procedure yielded ILs with water contents ≤ 3 wt.-%.

3.2.4 Calculation of HBr amount from measured pH value

In some experiments, (continuous) pH monitoring of the downstream scrubber unit

(washing flask) was used to determine the amount of trapped HBr using the following equations:

�� = − ��[�] = − ��[��] (8) �(��) = �(��) × �(�� ) (9) �(��) = �(��) × �(��) (10)

3.2.5 Off-line analysis of gaseous products

In some experiments, gaseous by-products were collected in a gas sampling bag and subsequently analyzed via gas chromatography (GC). Gaseous probes were qualitatively analyzed with a Varian GC 450 (ShinCarbon ST100/120 column, 2 m length,

0.75 mm inner diameter). A thermal conductivity detector (TCD) was used for CO and CO2 detection. The method was characterized by an injection time of 20 seconds and a temperature range from 40-200 °C (10 K min-1). The start temperature (40 °C) was held for 2.5 min and the target temperature (200 °C) was held for 7.5 min Argon was used as a carrier gas with a column pressure of 4.82 bar. Detected CO (coupling product of AcH) and CO2 (coupling product of ethylene) were used to prove the presence of undesired and competing side reactions, namely decarbonylation and decarboxylation.

4 Results and Discussion

The results presented in this thesis are grouped into three parts. In the first part, the newly developed technology for a liquid-phase production of bio-AA is introduced and evaluated for lactic acid as feedstock. Furthermore, the results of the development and optimization of a multi-step processing of the NADA concept are presented in the main part of this thesis. Finally, possibilities for process shortcuts are discussed.

4.1 The Erlanger “NADA” process: A Nucleophile assisted dehydration to acrylates

The novel concept of “Nucleophile Assisted Dehydration to Acrylates” (NADA) enables a liquid-phase dehydration of lactic acid towards bio-acrylic acid. This innovative technology was developed within this work in close collaboration with the industrial partner P&G within the research project “Liquid-phase dehydration of lactic acid obtained via fermentation for the production of bio-acrylic acid” (from 2014 to 2017) at the Institute of Chemical Reaction Engineering (CRT) of the Friedrich-Alexander-

University Erlangen-Nürnberg (FAU). The following chapter introduces the fundamentals of the here-developed Erlanger “NADA” process and highlights all associated components and their specific role within the “NADA” technology. Moreover, mechanistic fundamentals are presented. The chapter ends with details on current limitations of the NADA process and underlines the improvement which could be made by multi-step processing.

At the beginning of this project, several possibilities of LA functionalization were investigated to overcome the (liquid-phase) dehydration limitations resulting from feedstock sensitivity and bad leaving group capability of hydroxide (OH-).[A,B] Espe- cially the use of bromide (Br-) as nucleophilic substitute showed promising results and was therefore further investigated regarding its ability to enable AA formation from LA in the liquid phase. Hence, the eponymous origin of the “Nucleophile As- sisted Dehydration to Acrylates” technology is the crucial role of the used nucleophile, namely bromide. The development and optimization of a one-step liquid-phase 4 Results and Discussion 73

process for the production of bio-AA from bio-renewable feedstock is entirely described by Nagengast.

Here, the focus lies on the fundamental concept of the “NADA” technology, which defines the conceptual task of this PhD thesis: Development of an optimized, multi- step liquid-phase LA dehydration process.

4.1.1 The NADA technology

In principle, the NADA concept can be described in simplified terms as an HBr-triggered dehydration of LA to AA at reaction temperatures from 120-220 °C (see Scheme 20).

Scheme 20: The NADA concept: An HBr-triggered dehydration of LA to bio-AA.

The substrate, converted to AA in the NADA process, can be chosen from a broad range of different LA derivatives: LA, lactide, lactates, LA oligomers and polymers, or functionalized LA derivatives like e.g. 2-APA. However, LA is the most privileged bio-renewable feedstock, as illustrated in chapter 2.2 and is therefore in the focus of this study. The role of the leaving group is essential to achieve a liquid-phase dehydration of LA. Mechanistic investigations revealed that HBr serves as acidic catalyst and bromination agent for the in-situ bromination of LA via nucleophilic substitution of OH- and hence, introduces a good leaving group (Br-). LA bromination and dehydrobromination of bromopropionic acids to AA proceed via reactive transition states, allowing for competing courses of the reaction, e.g. decarbonylation to AcH and CO. 74

Hence, a charge-stabilizing reaction matrix is used to quench the readily decomposing carbocation transition state and therefore enables the formation of AA. Ionic liquids containing a bromide anion have shown to be the most suitable reaction matrices for the NADA process, even acting as thermally-stable solvent and bromide source.

Currently, the best performing IL is tetrabutylphosphonium bromide ([PBu4]Br), which is additionally characterized by high thermal stability (up to 220 °C) and recyclability (demonstrated for four cycles) within the one-step NADA process. Nonethe- less, the choice of the NADA reaction matrix is by no means limited to [PBu4]Br, as other bromide ILs and bromide-enriched melts were demonstrated to be active in the production of AA from LA derivatives. Moreover, it was shown that the source of acidity can be alternative acidic compounds like e.g. HBr, 2-BrPA, or H4P2O7. In the case of using an organically bound form of HBr, e.g. 2-bromopropionic acid, HBr is formed in-situ by decomposition of the compound (AA and HBr). When HBr-free acids are used, HBr is generated in-situ with the help of the IL (exemplary shown for

[PBu4]Br and H4P2O7 in Scheme 21). [97b-d]

Scheme 21: In-situ HBr formation within the [PBu4]Br reaction matrix.

Together, acid and reaction matrix form the so-called “NADA molten salt reaction matrix”. The state-of-the-art combination of the “NADA molten salt reaction matrix” is 2-BrPA and [PBu4]Br (acid:IL molar ratio of 1:20). However, molten salt combinations are not strictly limited to the ones disclosed herein. Quite in the contrary, various other combinations have successfully been demonstrated by Nagengast. After bromination of LA, 2-BrPA is dehydrobrominated in the molten salt catalyst and forms AA. A consecutive addition of HBr to the vinyl group of AA can lead to the formation of 3-BrPA. The hydrobromination/dehydrobromination equilibrium between AA and

3-BrPA strongly depends on the HBr concentration and residence time. The observed intermediates found over the course of LA dehydration in the “NADA molten salt reaction matrix” were used to devise a mechanism for LA dehydration. As shown in 4 Results and Discussion 75

Scheme 22, the proposed mechanism is a sequential conversion of LA to AA proceeding via brominated LA intermediates.[97b-d]

To sum up, the Erlanger NADA process can be described as an HBr triggered dehydration of commercial bio-renewable LA (and LA derivatives) feedstock in a specific acidic and bromide-enriched molten salts matrix proceeding via brominated LA intermediates to form bio-acrylic acid (acrylates).

Scheme 22: Postulated mechanism for LA dehydration to AA triggered by HBr in the NADA molten salt reaction matrix.

It has been shown that short reaction times enabled by in-situ product removal are essential to achieve high yields of AA in good selectivity. The best results were obtained by catalytic cracking of lactide in the “NADA molten salt reaction matrix” with outstanding YAA of over 80 %. Moreover, a continuous operation of the NADA process has successfully been developed and demonstrated in kg-scale by Kadar. [97b-d]

4.1.2 Benchmarking of the one-step NADA process with LA feedstock

At the beginning of the studies described here, an exemplary LA dehydration experiment in the one-step NADA process was performed. The experiment provided the benchmark result for upcoming experiments and optimization steps. Moreover, the experiment was used to highlight the limitations of the state-of-the-art one-step

NADA process in the case of bio-based LA feedstock (PURAC®, 88 wt% LA in water).

The LA dehydration experiment was monitored over 168 h at 150°C at a LA:[PBu4]Br molar ratio of 1:2 and the used acid was 2-BrPA. The time-resolved course of the reaction is shown in Figure 7 and the calculated and related quantities, namely LA conversion (XLA), AA yield (YAA) and selectivity (SAA) are depicted in Figure 8. Further details on the experimental procedure are given in the Experimental section.

35 nLA nLA,oligo

30 n2-BrPA

n3-BrPA

25 nAA

20 mmol 15

0 0 20 40 60 80 100 120 140 160 Reaction time / h

Figure 7: Time-resolved profile of the NADA process: Lactic acid conversion to acrylic acid in [PBu4]Br.

Conditions: 50 mmol LA (PURAC®, 88 wt% LA in water), 100 mmol [PBu4]Br , 5 mmol 2-BrPA, 168 h, 150 °C, ambient pressure. Lines are only given as a guide to the eye.

4 Results and Discussion 77

Under the applied reaction conditions, LA and LA oligomers are slowly converted over the course of the reaction. XLA reaches 90 % after 168 h. The share of LA oligomers, present from the start of the reaction, originate from the used feedstock PU-

RAC®, which contains an initial oligomer content of approximately 15-20 wt%.[64]

2-BrPA, which was previously added as source of HBr (5 mmol), is also slowly decomposed but in detectable quantities (approximately 1 mmol) until the end of the experiment. 3-BrPA is observed over the whole reaction time of the experiment and in constant range of two to three mmoles. The formation of 3-BrPA is assumed to result from hydrobromination of AA. Simultaneously and most importantly, the amount of formed AA increases over the first 96 h of the experiment and reaches a maximum of approximately 16 mmol (YAA = 30 %), which was determined to be the dehydration equilibrium of the one-step batch operation at 150 °C and ambient pressure. The formed stoichiometric amount of H2O, which is the coupling product of LA dehydration to AA, as well as the high water content of the substrate itself, considerably contribute to the equilibrium limitation. Note that 88 wt% LA in water (PU-

RAC®) correspond to a LA:H2O molar ratio of approximately 1:0.7. Dehydration selectivity is clearly limited over the course of the the reaction with selectivity in AA of

28 to 37 %. The selectivity gap is best observable when the equilibrium-limited YAA is reached after 96 h and LA is still converted in considerable amounts (12 %). The mainly observed and competing side-reaction is decarbonylation of the C3-skeleton to AcH and CO, which proceeds via a carbocation transition state and is known to be present at elevated temperatures and acidic conditions.[72b, 136] 78

100 XLA

YAA

80 SAA / % /

AA & S &

AA 40

, Y ,

LA X 20

0 0 20 40 60 80 100 120 140 160 180 Reaction time / h

Figure 8: Time-dependent LA conversion (XLA), AA yield (YAA) and selectivity (SAA) of AA production from LA with [PBu4]Br and 2-BrPA. Conditions: 50 mmol LA (PURAC®, 88 wt% LA in water), 100 mmol

[PBu4]Br , 5 mmol 2-BrPA, 168 h, 150 °C, ambient pressure.

The above described experiment is one example for LA dehydration in the “NADA molten salt reaction matrix” and used to illustrate the typical course of the batchwise

LA dehydration and the related limitations of the process. Moreover, due to the high temperature sensitivity of the reaction system, a comparably low reaction temperature was chosen to minimize the share of decarbonylation. Note that the share of decarbonylation is still tremendous, with selectivity towards AcH/CO of approximately

60 %. Finally, it should be noted that 30 % yield of AA at 150 °C in the “NADA molten salt reaction matrix” is the so far highest YAA ever reported in a liquid-phase process based on commercial and aqueous LA (PURAC®, 88 wt% LA in water).

4 Results and Discussion 79

4.1.3 Motivation for a multi-step NADA process

The described one-step NADA process offers an attractive and novel method to convert LA and derivatives thereof to bio-AA in a liquid-phase process. Especially, low reaction temperature (120-220 °C), a simple reactor design (STR/CSTR) as well as the homogenously HBr-catalyzed dehydration system differentiate this technology for liquid-phase LA dehydration from other known processes.[78-79, 81, 136a] However, this novel technology suffers from limitations, in some instances even arising from its component properties. This chapter illustrates the current drawbacks of the NADA technology using bio-LA feedstock (PURAC®) and proposes ways to overcome the presented barriers.

First and foremost, it should be noted that bio-renewable LA feedstock, which is an

88 wt% LA solution in water, contains substantial amounts of water, a fact that limits dehydration per se. Furthermore, highly concentrated LA solutions are characterized by non-negligible self-esterification behavior.[3a, 5c, 64b] Moreover, further side reactions of LA, like e.g. decarbonylation[3a, 5c, 136b, 137], decarboxylation[3a, 5c, 136b, 137] or condensation[74, 138], are known to be competing at elevated temperatures, limiting the selectivity in AA of the NADA process. The most prominent observed side-reaction, which results from the readily decomposing carbocation transition state, is C3-decomposition via decarbonylation to AcH and CO.[3a, 5c]

Bearing in mind the postulated mechanism of LA dehydration in the “NADA molten salt reaction matrix” (see Scheme 22), a one-pot LA conversion via brominated species requires conflicting conditions: While the first reaction step, namely the bromination of LA to 2-BrPA, is an acid catalyzed nucleophilic substitution and requires high concentrations of HBr/bromide, HBr is eliminated in the subsequent dehydrobromination of bromopropionic acids to AA. Here, HBr removal or at least minimized HBr concentrations are essential. High HBr/bromide concentrations at the end of the reaction network may lead to enhanced 3-BrPA formation by HBr addition to the vinyl group of AA, Moreover, decomposition via competing pathways like e.g. decarbonylation of the C3-skeleton is strongly favored in acidic environment, limiting SAA of the one-step process. In addition, it is quite likely that HBr elimination from the β-substituted propionic acid (3-BrPA) enables a simplified and more selective formation of 80 the desired product, namely AA. The state-of-the-art one-step NADA process has a limited degree of freedom for potential optimizations of the individual reaction steps.

Dehydrating LA using our NADA method in a single reactor will therefore always be a compromise solution of reaction conditions.

To overcome the described limitations and thus to achieve a selective and efficient production of bio-AA from LA, a multi-step NADA process was conceptualized based on the postulated mechanism of the NADA concept. A schematic overview of the multi-step NADA process is shown in Scheme 23.

Multi-step NADA process

I. Bromination II. Isomerization III. Dehydrobromination of LA to 2-BrPA from 2- to 3-BrPA of 3-BrPA to AA

LA AA

acidic bromide IL basic environment environment environment

I II III

Scheme 23: Overview of the multi-step NADA process.

The first reaction of the three-step NADA process operation is the bromination of bio- based LA to 2-BrPA (I). The bromination of LA is a nucleophilic substitution, typically performed in acidic environment. The introduction of bromide in this first step of the reaction sequence is expected to overcome the limitations of LA dehydration resulting from the bad leaving group capacity of the hydroxide. Moreover, water is removed in a separated reaction step compared to the one-step dehydration. In the second step of the reaction sequence, 2-BrPA is transformed to 3-BrPA, which is formally 4 Results and Discussion 81

an isomerization reaction (II) from �- to �-position of the bromopropionic acid. How- ever, 3-BrPA formation is assumed to result from dehydrobromination of 2-BrPA and subsequent HBr addition to the double-bond of AA. It is suggested that high HBr/bromide concentrations and even charge stabilization of the transition states are key factors for high isomerization selectivity in 3-BrPA. It should be noted here that direct synthesis of AA from 2-BrPA is highly desired but isomerization of 2-BrPA to 3-BrPA and subsequent conversion to AA is considered to be more appropriate, circumvent- ing the challenging dehydrobromination of 2-BrPA. Hence, the third and last step of the multi-step NADA process is the dehydrobromination of 3-BrPA. The elimination of HBr from 3-BrPA is typically performed in basic environment and is expected to form AA in high selectivity.

The decoupling of acid- and base-catalyzed reaction steps allows for increased degrees of freedom for optimized reaction conditions of each individual reaction and its environment. Furthermore, the inclusion of 3-BrPA is assumed to enable simplified dehydrobromination. The only coupling product of the process is H2O, making the process an atom efficient and sustainable alternative for the production of bio-AA.

However, high selectivity of the multi-step process and efficient HBr recycling are crucial for meeting the requirements of an efficient and environmentally friendly process.

In the following chapters, the development of the multi-step NADA process is described. After investigation of the feasibility of the individual reaction steps, the multi-step process was developed and optimized. Finally, the multi-step NADA process was evaluated and opportunities for process shortcuts are given. New insights of the individual reaction steps of the NADA reaction sequence were used to gain a better understanding of the one-step process.

4.2 Development and optimization of a multi-step liquid-phase NADA process for the production of bio-acrylic acid

In the following, the development of the multi-step NADA process is presented. In the first part of the process development, LA bromination to 2-BrPA is in the focus.

Large parts of this section (4.2.1) were developed within the master thesis of

J. Mehler.[C] Furthermore, the results of 2-BrPA isomerization to 3-BrPA in bromide

ILs are presented. The third part of the multi-step NADA process development deals with the work on base-assisted HBr elimination from 3-BrPA to AA.

4.2.1 Synthesis of 2-bromopropionic acid from lactic acid and lactide

In principle, 2-BrPA is accessible from LA via replacement of the hydroxyl group by bromide. A typical protocol would include refluxing the alcoholic substrate with aqueous hydrobromic acid (HBr(aq)).[103b] Indeed, the formation of 2-BrPA from LA and

HBr(aq) is known for a long time and has been first described by Kekulé in 1864.[104] Therefore, at the beginning of the studies described here, the bromination of lactic acid and lactide in aqueous reaction medium was investigated. In this context, a commercial hydrobromic acid (HBr(aq); 48 %) from Sigma-Aldrich Chemie GmbH was used.

4 Results and Discussion 83

4.2.1.1 Lactic acid bromination with aqueous hydrobromic acid

In the first experiments, the formation of 2-BrPA from LA and lactide in the presence of HBr(aq) at elevated temperatures was confirmed (see Scheme 24).

Scheme 24: Observed reaction mechanism of lactide bromination in HBr(aq) at elevated temperatures.

If lactide was used, substrate hydrolysis was already observed after addition of HBr(aq) at room temperature. When heated to elevated temperatures (e.g. 120 °C) for 5 hours, formation of 2-BrPA was proven via 1H NMR spectroscopy

(see Scheme 25).

Scheme 25: Partial 1H NMR spectra of an aqueous bromination experiment with lactide (0.05 mol) and HBr(aq) (0.3 mol). 1H NMR of pure lactide (top, **traces of water) and 1H NMRs of samples taken at a reaction time of 0 (mid) and 5 h (bottom). Experiment conducted at 120 °C and ambient pressure. *DMSO-d6 as solvent. 84

1H NMR analysis of the aqueous reaction solution revealed instant hydrolysis of lactide upon addition of HBr(aq). After diluting lactide with HBr(aq), an aqueous solution containing monomeric LA (3.9 ppm, q, 1H, 4; 1.1 ppm, d, 3H, 3) was formed. Heating the reaction solution to 120 °C for 5h led to the formation of 2-BrPA (4.4 ppm, q, 1H,

6; 1.5 ppm, d, 3H, 5). However, the selective production of 2-BrPA from LA and

HBr(aq) is very challenging. Two side-reactions were observed, which will be discussed in the following.

First investigations have identified that in aqueous and acidic reaction medium

(HBr(aq) as bromination agent for LA) acid-catalyzed substrate decarbonylation and self-oligomerization / polymerization reactions predominate. The postulated reaction network of the aqueous bromination system is shown in Scheme 26.

Scheme 26: Postulated reaction network in aqueous reaction medium with HBr(aq) as bromination agent showing the desired bromination (mid) and both side-reactions, namely substrate decarbonylation (top) and polymerization (bottom).

Enormous mass-loss over longer reaction times and significant evolution of gas above reaction temperatures of approximately 100 °C resulted from the strongly competing decarbonylation pathway, which is a well-known side-reaction in the field of LA chemistry.[2a, 4f, 5d, 28, 81, 86c] Gaseous by-products, namely AcH and CO, were mainly lost in the open setup and were detected by off-gas analysis. 1H NMR analysis of liquid samples taken from a heavy water (D2O) filled gas bubbler (connected to the off-gas 4 Results and Discussion 85

stream of the experimental setup) qualitatively revealed the formation of acetaldehyde. CO was qualitatively proven with a gas detector (Dräger PAC 3500). The presence of quenched acetaldehyde in the gas bubbler combined with a CO-rich off-gas stream is a strong evidence for decarbonylation. Quantification of non-water soluble and gaseous by-products was ensured by a water-filled hydrostatic column connected to the off-gas stream of the experimental setup (see Experimental section).

Moreover, polymerization has been verified by precipitation of black and amorphous solid from temperatures above 80 °C. The intramolecular and acid-catalyzed

(poly)condensation of LA to oligomers/polymers of LA showed to be strongly temperature and concentration-dependent. High reaction temperatures and high LA concentrations were correlated to a high degree of polymerization. The non-water soluble polymer was separated by filtration.

It should be noted that the aqueous bromination system is quite restricted. Tr is limited by the boiling temperature of the reaction mixture, which was at approximately

120 °C. This goes hand in hand with high HBr volatility under standard bromination conditions. Additionally, the desired bromination of LA to 2-BrPA as well as the observed side-reactions require elevated temperatures and acidic reaction conditions.

Note, that both decarbonylation as well as (poly)condensation produce stoichiometric amounts of water adding to the water in the aqueous bromination agent (HBr(aq), 48 wt% in water) limiting the equilibrium conversion of LA to 2-BrPA, where water is again formed as coupling product. To reduce the amount of by-products formed by the addition of water (shift the equilibrium of the reaction towards the substrate side) was therefore no possibility to close the selectivity gap.

Obviously, LA polymerization is unwanted. However, the depolymerization of polylactic acid back to LA is described in the literature[70a, 139] and theoretically allows recycling of polymerized substrate. For instance, this method is applied by Cargill[70a] in the large-scale production process of lactide and lactide polymers. Nevertheless, this is related to a further process step and therefore an undesired opportunity which should be prevented. In contrast, decarbonylation is a challenge for the entire process and poses a risk for a potential application. Acetaldehyde and CO are health hazard- 86 ous components, which require special handling and additional downstream processing. Furthermore, every molecule of lactic acid that undergoes decomposition via decarbonylation is lost for the overall process selectivity and consequently lowers the economic viability of the technology.

To investigate and evaluate the potential of the aqueous reaction medium for the bromination of LA, we screened different LA:HBr molar ratios at different reaction temperatures and reaction times. The results are summarized in Table A 1 (see Appen- dix). For experiments with an excess of substrate (e.g. LA:HBr of 2:1), quantification via 1H NMR was not possible due to high degrees of oligomerization, resulting in an overlap of 2-BrPA and LA oligomer proton signals.

Due to direct hydrolysis of lactide to LA in the HBr(aq) containing reaction medium, the cyclic dimer of lactic acid can be seen as LA feedstock (2 equivalents of LA) with reduced initial water content (compared to e.g. PURAC®). To keep the initial water content as low as possible, lactide is used as starting material for all upcoming bromination reactions. As far as not otherwise stated, this is the case for all following LA bromination experiments.

4.2.1.2 Parameter variation of aqueous bromination medium

As the bromination of LA with HBr is a stoichiometric reaction, a LA:HBr molar ratio of 1:1 would be most favorable from an economic point of view. However, reaction mixtures with high LA concentrations tend to polymerize, leading to significant amounts of substrate-based oligomers/polymers.[71c, 140] This was visually followed by precipitation of amorphous solid turning reaction mixtures black and highly viscous.

In 1H NMR analysis, LA oligomers and polymers were observed as multiplets in the chemical shift (δ) range of 1 to 2 ppm and 4 to 5 ppm.[141]

In a first set of experiments, a variation of the molar ratio of LA to HBr(aq) from 1:1 to

1:12 was performed. Experiments were conducted for 72 h at 100 °C to minimize the share of decarbonylation, which is strongly enhanced at higher reaction temperatures. The results are shown in Figure 9. 4 Results and Discussion 87

XLA

Y2-BrPA

60 S2-BrPA

50 / % /

40 2-BrPA

& S & 30 2-BrPA

, Y , 20

LA X 10

0 1:1 1:3 1:6 1:9 1:12 LA:HBr molar ratio

Fig- ure 9: Concentration-dependent 2-BrPA synthesis with HBr(aq). Conditions: 0.05 mol lactide, 0.05-0.60 mol HBr(aq), 72 h, 100 °C, ambient pressure.[C]

As can be observed in Figure 9, a LA:HBr molar ratio of 1:1 results in the lowest Y2-BrPA

(15 %) and S2-BrPA (37 %). This is mainly due to the low water content of the reaction mixture. In highly concentrated solutions LA tends to polymerize leading to a high degree of substrate oligomerization. After a reaction time of 72 h, the aqueous mixture was of tar-like consistency and black colored. Increasing the amount of HBr(aq) leads to an increased Y2-BrPA of 20 (1:3) or even 24 % (1:6). Additionally, S2-BrPA was also increased to a molar ratio of 1:9 (54 %). Further addition of HBr(aq) does not lead to a further increase in S2-BrPA and Y2-BrPA. Obviously, the addition of HBr(aq) not only increases the amount of the bromination agent but also the amount of the bromination coupling product water (48 wt% HBr in water). Due to the equilibrium nature of the bromination it can therefore be claimed that Y2-BrPA does not increase over a certain level by further addition of HBr(aq). Even within shorter reaction times (5 h) the results of three experiments performed at 120 °C showed that increasing the amount of

HBr(aq) only leads to a slightly increased Y2-BrPA (from 17 to 21 %). However, this observation goes hand in hand with a significant drop in S2-BrPA (from 64 to 48 %) This 88

can be explained by the determined loss of C3-species, which is assumed to arise from decarbonylation at 120 °C (see Table A 1, Appendix).

Therefore, it can be stated that the ideal molar ratio of LA:HBr for LA bromination with HBr(aq) is higher than 1:1 and lower than or equal to 1:6. It should be noted that an increase in amount of HBr is also accompanied by a larger reactor volume and therefore lower space-time-yield. A LA:HBr molar ratio of 1:3 seemed to be a compromise between Y2-BrPA and reactor volume and was therefore chosen for future investigations of the aqueous bromination system.

Additionally, we investigated the influence of substrate concentration on the performance of the aqueous bromination system. The experiments were performed with a

LA:HBr molar ratio of 1:3 and the reaction time was 72 h. The reaction temperature was 105 °C as this value is accessible with all of the used aqueous LA solutions. The results are summarized in Table A 2. Increasing the substrate concentration stepwise from 20 to 100 wt% (lactide) led to enhanced XLA and Y2-BrPA after 72 h. While no formation of 2-BrPA was observed when a 20 wt% LA solution was used, Y2-BrPA continuously increased to 26.1 % with lactide as substrate. Moreover, a notable example is a Y2-BrPA of 22.6 % when PURAC® (88 wt%) is used as substrate. However, with increasing substrate concentration, XLA and Y2-BrPA, the selectivity in 2-BrPA decreased stepwise from 62.5 % (40 wt%) to 39 % (100 wt%). It had to be concluded that increasing substrate concentration led to increasing LA oligomerization/polymerization, indicated by precipitation of black and solid material. Nonetheless, high substrate concentrations were found to be essential to achieve 2-BrPA yields over 20 %, as the additional water in low concentrated LA solutions limits the equilibrium conversion of the bromination of LA with HBr(aq).

A third parameter to be investigated was the reaction temperature (Tr) of the bromination. The first investigations showed that elevated temperatures are needed to form

2-BrPA from LA and HBr(aq). Table 1 shows that first traces of 2-BrPA were observed at temperatures of 80 °C within 72 h of reaction time. A further increase of Tr to 100 °C resulted in 19.6 % Y2-BrPA at LA conversion of approximately 50.0 %. The reaction is characterized by mass loss through the gas-phase (CO from decarbonylation) and precipitation of black solid (oligomers from LA esterification). When the bromination 4 Results and Discussion 89

is performed at 120 °C for 72 h, a LA conversion of 90.0 % and 2-BrPA yield of 12.1 % was determined. The decreased S2-BrPA (from 37.7 to 13.4 %) results from increased arising decarbonylation at elevated Tr, which resulted in 78 % C3-loss. Assuming that the bromination is in equilibrium at long reaction times (72 h), precipitation of LA oligomers over longer reaction times shifts the reaction towards the substrate side, which additionally lowers Y2-BrPA and S2-BrPA.

To gain further insights into the temperature influence, a second set of experiments was performed at shortened reaction time (24 h) within the relevant temperature range (90-120 °C). Shortened reaction times could reduce the amount of formed polymer and substrate decomposition and thus increase Y2-BrPA. In the following experiments a LA:HBr molar ratio of 1:3 was used. The determined temperature-dependency of the LA bromination with HBr(aq) is shown in Figure 10.

Table 1: Variation of the reaction temperature of 2-BrPA synthesis with HBr(aq).[a]

Reaction temperature / °C XLA / % Y2-BrPA / % S2-BrPA / % Loss / mol%

25 0 0 - -

60 0 0 - -

80 < 2 traces - -

100 52.0 19.6 37.7 31

120 90.0 12.1 13.4 78

[a] 0.05 mol lactide, 0.30 mol HBr(aq), 72 h, ambient pressure, quantified via 1H NMR using AcOH as internal standard.[C] 90

XLA

Y2-BrPA

70 S2-BrPA

60 / % /

50 2-BrPA

40 & S &

2-BrPA , Y ,

LA 20 X

0 90 100 110 120 Reaction temperature / °C

Figure 10: Temperature-dependent 2-BrPA synthesis with HBr(aq). Conditions: 0.05 mol lactide, 0.30 mol

HBr(aq), 24 h, ambient pressure.[C]

These results confirm the initial assumption that a decrease in reaction time increases the S2-BrPA due to lower degree of polymerization and substrate decomposition. Com- pared to the bromination experiment with a reaction time of 72 h (Y2-BrPA = 12.1 %;

S2-BrPA = 13.4 %), the Y2-BrPA (26.6 %) and S2-BrPA (40.0 %) was significantly increased at a reaction time of 24 h. Moreover, whereas the S2-BrPA was constantly enhanced by reduced reaction temperatures, the observed trend for Y2-BrPA is working in opposite direction (see Figure 10). For instance, at a reaction temperature of 100 °C the S2-BrPA is increased to 52.9 % and Y2-BrPA is reduced to 12.7 %. The selectivity of the LA bromination reaches a maximum at 90°C with 71.4 %. Unfortunately, Y2-BrPA is significantly decreased to 5.0 % at this reaction temperature. This is due to low reaction rates at mild temperatures. Thus, it can be concluded that a reaction temperature of at least

110 °C (Y2-BrPA = 25.0 %; S2-BrPA = 49.3 %) is crucial to perform the reaction in an economically relevant timeframe. To still achieve acceptable Y2-BrPA, the reaction time for the LA bromination at 110 and 120 °C was further reduced to 5 h (see Table A 3 in

Appendix). Again, a further reduction of the reaction time increased the S2-BrPA even 4 Results and Discussion 91

more. After five hours, Y2-BrPA of 20 % (110 °C) was observed with a reasonable selectivity of 50 %. In contrast, a reaction temperature of 110 °C showed higher selectivity

(62.5 %) but significantly reduced Y2-BrPA (10 %), which is the reason why this temperature did not display a viable alternative to 120 °C. In both experiments, decarbonylation accounted for approximately 75 % of the undesired side-reactions, which was followed by gas evolution and the related C3-loss.

An overview of the narrow temperature and time frame accessible with the aqueous bromination system is presented in form of a 3D plot in Figure 11. Here, all performed experiments with a LA:HBr molar ratio of 1:3 are used to show the limitation of LA bromination with HBr(aq). As can be seen in time- and temperature-dependent Y2-BrPA plot (top), highest yields were accessible with a reaction time of 24 h. High reaction temperatures (110-120 °C) are essential to achieve acceptable yields of 2-BrPA, which were in the maximum range of 25 %. Unfortunately, the reverse behavior is true for the determined S2-BrPA pattern. Short reaction times (≤ 24 h) and low reaction temperatures are necessary to achieve high S2-BrPA. The highest selectivity was found after a reaction time of 5 h at a reaction temperature of 90 °C. However, the related Y2-BrPA was very small (5 %). A clear trend was observed for longer reaction times (> 24 h), characterized by decreasing Y2-BrPA and S2-BrPA for all investigated reaction temperatures. Hence, to minimize the share of side-reactions, namely oligomerization and decarbonylation, short reaction times are crucial.

To conclude, the synthesis of 2-BrPA from LA and HBr(aq) is possible but severely limited with regard to product yield and selectivity. Under optimized reaction conditions for the aqueous bromination system, namely 5 h, 120 °C, and a LA:HBr molar ratio of 1:3, selectivity towards 2-BrPA (S2-BrPA) was limited to approximately 50 %.

The obtained yield of 2-BrPA (Y2BrPA) was approximately 20 %. Even at longer reaction times (24 h) the highest determined Y2-BrPA was approximately 25 % (110 °C) with a corresponding S2-BrPA of 49 %. Moreover, the performance of the aqueous bromination system was limited in its composition itself. Decarbonylation as well as (poly)condensation of substrate both produce stoichiometric amounts of water, which strongly limited the equilibrium conversion of LA in addition to the high water content of the bromination agent itself (HBr(aq), 48 wt.% in water). This was most likely the reason 92

for the limited yields of 2-BrPA when HBr(aq) was used as bromination agent. Another characteristic of the aqueous reaction medium was its boiling temperature of approximately 120 °C, which restricted the reaction temperature for the bromination of LA.

Shorter reaction times positively affected the product selectivity. However, a further reduction of tr at even higher Tr was inaccessible. Moreover, the high acidity of the reaction medium, which is mainly due to the bromination agent HBr(aq) (pKa = -

9.00[142]) was also assumed to enhance the reaction rate of the acid-catalyzed side-reactions, especially substrate decarbonylation.

To increase Y2-BrPA it was necessary to either remove the product from the reaction mixture or to increase the bromide concentration in the reaction mixture (without increasing proton concentration and water content), following Le Chatelier’s principle.

The bromination experiment at 120 °C with a LA:HBr molar ratio of 1:3 was chosen to be the reference for new concepts and further optimizations of LA bromination as it combines maximized Y2-BrPA (20 %) and S2-BrPA (50 %) with acceptable reaction time (5 h).

4 Results and Discussion 93

Figure 11: 3D-plot of temperature- and time-dependent Y2-BrPA (top) and S2-BrPA (bottom) in aqueous reaction medium. Conditions: 0.05 mol lactide, 0.30 mol HBr(aq), ambient pressure, quantified via 1H NMR using AcOH as internal standard.[C] 94

4.2.1.2 Concepts to increase the performance of the aqueous bromination medium

The bromination of LA with HBr(aq) was characterized to be strongly limited, especially in Y2-BrPA and S2-BrPA. In this chapter, two concepts are followed to overcome the current limitations of LA bromination with HBr(aq). On one hand, liquid-liquid extraction (LLE) was investigated for ex- and in-situ product removal approaching enhanced bromination performance. On the other hand, ionic liquids (ILs) were tested as additive to increase the bromide concentration of the reaction medium and therefore shift the bromination equilibrium towards the product side. Moreover, IL additives enable a higher Tr for the bromination of LA at ambient pressure as they increase the boiling temperature of the reaction mixture.

Liquid-liquid extraction of 2-BrPA

Liquid-liquid extraction (LLE) is an efficient and economical method for the separation of organic compounds from aqueous environment. The extraction procedure is based on two immiscible phases and one or more solute with different solubility in each phase. LLE describes the separation process which is based on the distribution of a component i between two liquid phases. One effective method for quantitative description of the extraction is the partition coefficient Ki, which is defined by the concentration ratio of the organic and the aqueous phase.[143] A detailed description of the experimental procedure and evaluation of the LLE experiments is given in the

Experimental section.

In-situ LLE was used to remove the solute 2-BrPA from the aqueous reaction medium during the bromination of LA. Assuming that the investigated bromination of LA is in equilibrium, constant extraction of 2-BrPA into a second organic layer could increase Y2-BrPA. Further product isolation and purification would be accessible by distillation of the formed product/extractor solution.

The following chapter summarizes the results of ex- and in-situ LLE applied to the synthesis of 2-BrPA from LA in aqueous bromination medium (HBr(aq)).

4 Results and Discussion 95

Prior to the extraction study of 2-BrPA, a screening of suitable extractors was performed. A suitable organic solvent for in-situ extraction of 2-BrPA from the aqueous reaction medium needs to fulfill the following requirements:

§ High K2-BrPA

§ High extraction selectivity towards 2-BrPA

§ Large miscibility gap with HBr(aq)

§ Boiling temperature Tb between H2O and 2-BrPA to ensure product isolation

and purification

§ Chemical and thermal stability vs. the aqueous bromination system

Based on the requirements shown above, the following organic solvents were investigated in a room temperature ex-situ LLE screening: n-decane, cyclooctane, di-n-butyl ether, toluene, chlorobenzene and bromobenzene. The results of the study are summarized in Table 2.

Table 2: Results of ex-situ room temperature LLE screening.[a] Partition coefficient Extracting agent Tb / °C K2-BrPA[b]

n-decane 174 0.23

Cyclooctane 151 0.45

di-n-butyl ether 141 3.54

Toluene 111 1.43

Chlorobenzene 132 1.58

Bromobenzene 156 1.58

[a] 1.5 mL of model solution (25 mmol lactide, 50 mmol 2-BrPA, 300 mmol HBr(aq)) mixed with [b] 1 1.5 mL of extracting agent for 24 h, K2-BrPA = corg/caq, obtained via quantitative H NMR measurements using AcOH (aq. phase) and mesitylene (org. phase) as internal standards.[C]

All investigated extracting agents showed high selectivity towards 2-BrPA, with no

LA observable in 1H NMR spectra of the organic phases. In addition, the used solvents were stable under room temperature extraction conditions. No visual or spec- troscopic verifiable changes were found and the phases clearly separated within one minute. The by far highest partition coefficient was found for di-n-butyl ether (3.45).

The aromatics showed comparable partition coefficients for 2-BrPA in the range of

1.5. Partition coefficients of cyclooctane and n-decane were comparably low, which could mainly be due to their low polarity. Consequently, they were not considered for future LLE experiments.

The most promising organic solvents for 2-BrPA extraction were investigated in in-situ LLE of 2-BrPA to check if constant product removal positively influences the bromination equilibrium following Le Chatelier’s principle. The experiments were conducted using a standard aqueous reaction mixture with a LA:HBr molar ration of

1:3. Additionally, an equimolar amount of the corresponding organic solvent is added prior to the LA bromination experiment.

First of all, three experiments with toluene as extractor and varying reaction time (5, 24 and 72 h) were performed to gain first insights on the influence of an extraction medium present in the reaction system. Due to the formation of a low boiling azeotrope (H2O-toluene), the reaction temperature was limited to approximately 100 °C.

While decarbonylation was found again, no polymerization was observed during the experiment. The results of this first LLE test are presented in Figure 12. 4 Results and Discussion 97

XLA Y 70 2-BrPA, aq. phase Y2-BrPA, org. phase S 60 2-BrPA

50 / % /

40 2-BrPA

& S & 30 2-BrPA

, Y , LA

X 10

5 (no LLE) 5 24 72 Reaction time / h

Figure 12: Time-dependent 2-BrPA synthesis in aqueous reaction medium using toluene for in-situ product removal compared to a standard experiment (120 °C) without LLE.0.10 mol PURAC®, 0.30 mol

HBr(aq), 34 mL extracting agent, 100 °C, ambient pressure.[C]

As can be seen in Figure 12, 2-BrPA selectivity is widely spread, ranging from approximately 40 to 65 %. Due to the high dilution of 2-BrPA in the organic phase, analysis with 1H NMR spectroscopy was inaccurate and susceptible to integration errors, complicating the calculation of S2-BrPA after 5 h. However, a clear trend was observed in these experiments. The results in Figure 12 show a significant increase of Y2-BrPA in both the reaction and extraction phase with longer reaction times. The share of 2-BrPA in the toluene phase is higher than in the aqueous phase, which goes hand in hand with the previously determined partition coefficient. After 24 h a combined Y2-BrPA of approximately 13 % was achieved with a remarkable S2-BrPA of 64 %. After 72 h the combined Y2-BrPA can be further increased to 26 %. Unfortunately, S2-BrPA is decreased to 41 %. Assuming that the bromination is in equilibrium but decarbonylation contin- ues, LA conversion still increases while bromination of LA has already achieved a maximum value. This leads to a drop in S2-BrPA at longer reaction times and has also been observed in the standard aqueous bromination system. Comparing the results 98 to the aqueous benchmark, the decreased reaction rate after 5 h is remarkable, which might also be due to the limited reaction temperature of 100 °C. Compared to a standard aqueous experiment conducted for 24 h at 100 °C (Y2-BrPA = 12.7, S2-BrPA = 52.9 %),

Y2-BrPA is almost identical for both bromination concepts. Hence, the presented results indicate that Y2-BrPA is unaffected by the presence of an extracting agent, whereas S2-BrPA is slightly increased, which may be due to hampered polymerization. For further insights on the influence of LLE on LA bromination with HBr(aq), a second experimental study including the remaining extractors of the previously performed screening was carried out. Due to the results of the LLE experiment with toluene, the reaction time was elongated to 24 h for the further LLE experiments. When chlorobenzene was used as extractor, the reaction temperature was again limited to 113 °C by the formation of a low boiling azeotrope (H2O-chlorobenzene). The reaction mixture with bromobenzene reached the desired reaction temperature of 120 °C. Unfortunately, di-n-butyl ether (K2-BrPA = 3.54) was not stable to in-situ reaction conditions, forming the hydrolysis product n-butanol. Instability was confirmed via 1H NMR spectroscopy. The corresponding results are presented in Figure 13 and are compared to the aqueous benchmark experiment (LA:HBr molar ratio 1:3, 5 h, 120 °C).

After a reaction time of 24 h, both chlorobenzene (27.3 %) and bromobenzene (29 %) showed an increased combined Y2-BrPA compared to the bromination experiment with toluene. Again, the amount of 2-BrPA in the aromatic phases is higher than in the aqueous phases, which is in accordance with the previously determined partition coefficients. Additionally, Y2-BrPA and S2-BrPA are in the range of a standard aqueous bromination experiment at a temperature of 110 °C and a reaction time of 24 h

(Y2-BrPA = 25.0 %, S2-BrPA = 49.3 %). Although the combined yields of 2-BrPA are slightly increased compared to the aqueous benchmark, the anticipated improvement through the use of LLE was not achieved.

4 Results and Discussion 99

XLA Y 80 2-BrPA, aq. phase Y2-BrPA, org. phase

70 S2-BrPA

60 / % /

50 2-BrPA

40 & S &

30 2-BrPA

, Y , 20

LA X 10

[-] @ 120 °C [toluene] @ 100[Cl-benzene] °C @ 113[Br-benzene] °C @ 120 °C

Extracting agent and reaction temperature / °C

Figure 13: Extractor-screening with promising organic solvents for in-situ product removal compared to a standard experiment without LLE. 100 mmol PURAC®, 300 mmol HBr(aq), 24 h, ambient pressure.[C]

To sum up, in-situ LLE of 2-BrPA from the aqueous bromination medium was possible and could be used for simultaneously 2-BrPA production and isolation. The addition of the investigated organic solvents to the reaction medium did limit the reaction temperature due to the formation of low boiling azeotropes in specific cases. Even though slightly enhanced yields for 2-BrPA were achieved and the selectivity was partially increased probably due to hampered substrate polymerization, significantly longer reaction times were needed to achieve the quantities of the aqueous benchmark experiment. LLE did not bring the anticipated enhancement for the bromination of LA with HBr(aq). Nevertheless, ex-situ product removal was successfully demonstrated with acceptable partition coefficients, which could be used for 2-BrPA isolation and subsequent purification. 100

[PBu4]Br as ionic liquid additive in the aqueous bromination medium

This chapter illustrates a second attempt to increase the performance of the aqueous bromination medium, namely the use of a bromide IL additive. According to Le Cha- teliers’s principle, an increase in bromide concentration should shift the reaction equilibrium towards the product side (compare Scheme 26). Additionally, ionic liquids and their charge-stabilizing character[12b] might also suppress decarbonylation by fast quenching the positively charged carbocation transition state of the bromination of

LA. In contrast to inorganic metal bromide salts, ionic liquids were highly soluble in the aqueous bromination medium and additionally of liquid state at typical reaction temperatures due to their relatively low melting temperature (Tm). Moreover, the addition of an IL formally reduces the LA concentration in the reaction mixture and functions as a kind of diluter. As known from previous studies, a decreased share of polymerization was expected when the LA concentration is reduced. Finally, the use of ILs might also increase the accessible reaction temperature by increasing the boiling temperature of the reaction mixture. [PBu4]Br was chosen as the bromide IL of choice, as it was already successfully employed in the one-step NADA process. Ad- ditionally, [PBu4]Br is characterized by good thermal and chemical stability.

In a first experiment, we investigated the bromination of LA with HBr(aq) with

[PBu4]Br as ionic liquid additive at the standard reaction temperature of 120 °C. A

LA:HBr:IL molar ratio of 1:3:5 was chosen. The bromination results within a reaction time of 72 h are shown in Figure 14. 4 Results and Discussion 101

XLA

Y2-BrPA

70 S2-BrPA

60 / % /

50 2-BrPA

40 & S &

2-BrPA , Y ,

LA 20 X

0 5 24 48 72 Reaction time / h

Figure 14: Time-dependent 2-BrPA synthesis in aqueous reaction medium with a bromide ionic liquid additive. 12.5 mmol lactide, 75 mmol HBr(aq), 125 mmol [PBu4]Br, 120 °C, ambient pressure.[C]

As can be seen in Figure 14, XLA increases with longer reaction times and reaches a value of 69 % after 72 h. The maximum Y2-BrPA (32 %) was determined after 48 h, indicating that the equilibrium is already reached. After 24 h, S2-BrPA continuously decreases from 59 % to 46 % (72 h), indicating an arising share of decarbonylation, as no polymerization was observed in the bromination experiments with [PBu4]Br. There- fore, addition of [PBu4]Br and the related substrate dilution suppressed the polymerization of LA. Furthermore, it can clearly be seen that the reaction rate is decreased when [PBu4]Br is used (Y2-BrPA of 13 %) compared to the aqueous benchmark experiment (120 °C, 5 h, LA:HBr molar ratio of 1:3, Y2-BrPA of 20 %). Compared to the highly polar reaction medium of the aqueous reaction system (H2O), the addition of the IL decreased the polarity of reaction medium. According to the Hughes-Ingold rules, decreasing polarity slows down the rate of a reaction where a charged transition state is evolved. Obviously, the SN mechanism of the bromination of LA is accompanied by the formation of a charged transition state (e.g. carbocation). Other reasons for the decelerated reaction rate might be the availability of the bromide ions, which are 102

strongly bound to the [PBu4]+ cation, resulting in reduced nucleophilic character and higher substrate dilution. Nevertheless, it can be noted that addition of [PBu4]Br and the resulting higher bromide concentration in the reaction mixture shifts the bromination equilibrium to the product side. An increased overall Y2-BrPA of 32 % was observed, which was the so far highest Y2-BrPA reached in the experiments using HBr(aq).

In a second set of experiments (see Table A 4, Appendix), different LA:HBr:IL molar ratios were tested in their LA bromination performance. Moreover, LA bromination was conducted at 140 °C to achieve higher reaction rates and to consequently enable shorter reaction times (5 h). The most important results are summarized in Figure 15.

XLA(120 °C)

XLA(140 °C) 75 Y2-BrPA(120 °C

Y2-BrPA(140 °C)

S2-BrPA(120 °C 60

S2-BrPA(140 °C) / % /

2-BrPA & S &

2-BrPA

, Y , LA

X 15

0 1:3:0 1:3:1 1:3:5 1:3:5 1:3:10

LA:HBr:[PBu4]Br molar ratio

Figure 15: Influence of [PBu4Br] concentration and reaction temperature Tr on the synthesis of 2-BrPA with ionic liquid additive. 12.5-125 mmol PURAC®, 37.5-75 mmol HBr(aq), 75-125 mmol [PBu4]Br, ambient pressure.[C]

As shown above, addition of small amounts of [PBu4]Br (≤1 eq.) decreases Y2-BrPA to approximately 10 % after 5 h. S2-BrPA (55 %) is slightly increased compared to the aqueous benchmark. The experiment with 5 eq. of [PBu4]Br (compare also Figure 14) leads 4 Results and Discussion 103

to a minor increase of Y2-BrPA (13 %). This behavior confirms again the assumption of a reduced bromination rate initiated by the addition of [PBu4]Br, even if only small amounts of IL are used (see Table A 4, Appendix). Increasing the reaction temperature to 140 °C accelerated the bromination rate and therefore positively influenced

Y2-BrPA and S2-BrPA for a LA:HBr:IL molar ratio of 1:3:5. Compared to the aqueous benchmark, Y2-BrPA (31 %) and S2-BrPA (63 %) are significantly increased within a reaction time of 5 h. The bromination equilibrated within 5 h (compare Figure 14). The increased

S2-BrPA (compared to the aqueous reaction system) could be explained by suppressed polymerization by dilution of the reaction medium and hampered decarbonylation by bromide-induced quenching of the highly reactive and charged transition state. A further increase of [PBu4]Br to 10 eq. led to higher substrate conversion (XLA = 61 %) and reduced Y2-BrPA (15 %) and S2-BrPA (25 %). Under these conditions, the major reaction pathway observed was decarbonylation.

Interestingly, if the reaction time of the new benchmark experiment (LA:HBr:IL molar ratio of 1:3:5, Tr of 140 °C) is elongated to 24 h, no noteworthy increase of Y2-BrPA (33 %) was observed. Though, an arising formation of 3-BrPA (Y3-BrPA = 16 %) was observed hinting for emerging isomerization of the product 2-BrPA (see Table A 4, Appendix).

In another experiment within the performed screening (LA:HBr:IL molar ratio of

1:1:5, 5 h) the reaction temperature was further increased to 160 °C. Here, no more 2-BrPA was found in the reaction medium after 5 h. However, the yield of 3-BrPA was approximately 6 %. In addition, considerable amounts of AA were found

(YAA = 11 %). These observations are in good accordance to the results of the one-step

NADA reaction system (see chapter 4.1.2). Furthermore, the drastically dropped

S2-BrPA (27 %) of the experiment at 140 °C showed that the desired product 2-BrPA is instable under these harsher bromination conditions. Even if those conditions were not desirable for the selective bromination of LA to 2-BrPA and are therefore not further focused in this chapter, the observations open interesting possibilities for the upcoming reactions within the multi-step NADA reaction system (2-BrPA conversion at higher temperatures).

104

To sum up, the use of an ionic liquid additive, namely [PBu4]Br, resulted in an increased Y2-BrPA and S2-BrPA compared to the aqueous benchmark experiment. This was achieved by a shift of the bromination equilibrium initiated by the high excess of bromide present in the reaction medium (Y2-BrPA = 32 %). Even if we observed a decelerated bromination rate for the IL-containing system, the increased boiling temperature of the reaction system enabling higher reaction temperature of 140 °C led to improved results within a reasonable reaction time of 5 h. The S2-BrPA was increased to 63 %, which was mainly due to suppressed polymerization and hampered decarbonylation in the IL-containing reaction medium. As a result, the concept of using [PBu4]Br as additive in the bromination of LA with HBr(aq) could be seen as the new benchmark.

However, the so far reached Y2-BrPA and S2-BrPA are still far away from profitable quantities. The high excess of bromide did not shift the equilibrium over a level of 32 %, which could be explained by the still substantial amount of water, resulting from the bromination agent itself (HBr(aq)). Moreover, the reaction is still characterized by a huge share of decarbonylation, which was assumed to be due to the high acidity of

HBr(aq) and resulted in a maximum reachable S2-BrPA of 63 %.

4 Results and Discussion 105

4.2.1.3 Zwitterionic HBr carriers for the synthesis of 2-BrPA from lactide and LA

To overcome the limitations of LA bromination with HBr(aq), ammonium-based zwitterions as HBr carriers and brominating agents were investigated.[133] The applied zwitterionic carrier molecule is loaded with HBr by protonation of the sulfonate group, thus forming an acidic bromide IL. This IL-based bromination system provides several benefits compared to HBr(aq): The bromide anion of the acidic bromide

IL appears to be more nucleophilic than that in pure HBr[133]. Furthermore, due to the drastically reduced H2O content of the brominating agent, the formation of 2-BrPA that releases water as coupling product should be significantly increased. Another very attractive feature of protonated zwitterion bromides is the low vapor pressure of HBr in these systems allowing for a save handling of HBr at elevated temperatures.

Thus, the zwitterionic HBr carriers combine the roles of solvent, proton source, buffer, and brominating agent within one substance.

The general concept of applying SO3H-functionalized bromide ionic liquids for a con- venient and efficient production of 2-BrPA from LA is illustrated in

Scheme 27. Note that the zwitterion serves as HBr carrier and its chemical structure remains unchanged over the course of the reaction. Therefore, successful reloading and reuse of the salt over several bromination cycles is expected.

Scheme 27: LA bromination cycle using tailor-made zwitterionic HBr carrier.

106

Screening of HBr-carrying zwitterions

To start the experimental work with HBr-loaded zwitterions as LA bromination agents, different zwitterion structures were tested to identify structural requirements for this particular application (see Table 3).

Table 3: LA bromination using different brominating agents: comparison of aqueous HBr (48 wt%) and different SO3H-functionalized bromide ionic liquids.

Entry HBr source XLA / %[a] Y2-BrPA / %[a] S2-BrPA / %[a]

1 HBr(aq) 40 20 50

2 [MIMBS]Br 48 48 > 99

3 [PyrBS]Br 69 54 78

4 [PBu3BS]Br oligomerization 0 0

Reaction Conditions: 0.05 mol lactide, 0.30 mol HBr source, 120°C, 5 h, ambient pressure; [MIMBS]Br = 1-(4- butanesulfonic acid)-3-methyl imidazolium bromide; [PyrBS]Br = 1-(4-butanesulfonic acid) pyridinium bromide;

[PBu3BS]Br = Tri-n-butyl-(4-butanesulfonic acid) phosphonium bromide; [a] conversion, yield and selectivity de- | | | | 1 , , , , termined by H NMR as X = , Y- = ∙ and S- = ∙ = , respectively; nr = , , ,, [C] moles of reactant; np = moles of product.

The zwitterions that contain nitrogen (entries 2, 3) were active in the bromination of

LA towards 2-BrPA. In contrast to HBr(aq), these ionic liquid HBr carriers proved to be highly selective to 2-BrPA. The highest S2-BrPA in this first screening was achieved with

[MIMBS]Br. With this system, no by-products were observed in 1H NMR and the mass balance could be entirely closed. In contrast, phosphonium-based ionic liquids did not show any activity. [PBu3BS]Br was inactive in the bromination under standard conditions (120 °C) and only lactic acid oligomers were found after 5 h in 1H NMR. At bromination temperatures above 120 °C, all systems showed an increasing tendency to promote substrate oligomerization/polymerization, as well as decarbonylation to AcH and CO.

4 Results and Discussion 107

To further reveal influencing effects on the selectivity and yield of the 2-BrPA synthesis, zwitterion-derived bromide salts of different lipophilicity and acidity were investigated. The respective results are shown in Figure 16.

XLA 100 Y2-BrPA S2-BrPA 80

/ % / Increasing lipophilicity 20

2-BrPA 0

[MIMBS]Br [EIMBS]Br [BIMBS]Br [OIMBS]Br & S & XLA 100 Y2-BrPA 2-BrPA S

80 2-BrPA

, Y , LA

X 60

20 Increasing acidity

0 [MIMBS]Br [MIMPS]Br [PyrBS]Br

IL bromination agent

Increasing lipophilicity in the bromide ionic liquid is easily realized by extending the alkyl chain attached to the imidazolium cation (Figure 16, top part). While replacing the methyl group in the [MIMBS]+ cation by an ethyl group does not influence the bromination activity significantly, the use of [BIMBS]Br already shows a clearly decreased Y2-BrPA and S2-BrPA. With the very lipophilic [OIMBS]Br as brominating agent for LA, the activity drops drastically. From these results we conclude that a decreased polarity of the reaction medium slows the rate of the bromination reaction, probably due to less efficient charge stabilization during the course of the bromination reaction. 108

The bottom part of Figure 16 shows the influence of the acidity of the zwitterion- based bromide IL on the bromination selectivity and yield. It is known from literature that the acidity increases in the following order:

[MIMBS]Br < [MIMPS]Br < [PyrBS]Br.[144] A decreasing S2-BrPA was observed as acidity increased from the mildly acidic [MIMBS]Br to the stronger acids [MIMPS]Br and

[PyrBS]Br (see Figure 16). This confirmed the assumption that too high acidity is detrimental to S2-BrPA as substrate decarbonylation and self-oligomerization are favored in strongly acidic media.

The following set of experiments has been carried out using the IL [MIMBS]Br. This selection was made due to the very attractive performance of this IL in the screening experiments (see Table 3 and Figure 16), but also based on economic considerations.

MIMBS is easily produced in large quantities from the reaction of two commercial compounds: 1-methyl imidazole and 1,4-butane sultone.

LA bromination using [MIMBS]Br - temperature variation

In a first set of experiments, the influence of the reaction temperature (Tr) on the selectivity of the LA bromination with [MIMBS]Br was investigated. The results are shown in Figure 17. Starting at the initial screening temperature of 120°C with a substrate to IL ratio of 1:3, an IL residual water content of < 1.5 wt%, and HBr loading of

75 mol%, a Y2-BrPA of 35% and S2-BrPA of approximately 100% was achieved within 5 h.

No by-products were observed by 1H NMR and the mass balance was closed. Work- ing at temperatures lower than 120°C led to a reduced LA conversion (XLA), due to a reduced reaction rate. After LA bromination in [MIMBS]Br 100°C, a Y2-BrPA of only 9% was observed. Interestingly, an increase of Tr to 140°C was not accompanied by a substantial additional increase in Y2-BrPA (37%). Instead, decarbonylation of LA was observed at 140°C leading to a decrease in S2-BrPA. Moreover, the formation of 3-bromopropionic acid (3-BrPA) was observed from an emerging isomerization activity of the system. These studies indicate that the so-formed 3-BrPA originates from a two- step isomerization process including AA formation from 2-BrPA and subsequent addition of HBr to AA to form 3-BrPA. Note that addition of HBr to AA under the applied reaction conditions forms highly selectively 3-BrPA. Moreover, heating of 4 Results and Discussion 109

2-BrPA in bromide salts in the absence of HBr forms AA as major compound as reported by Terrade et al.[10b]. Thus, the combined yield of brominated LA species YX-BrPA

(Y2-BrPA + Y3-BrPA) after 5 h reaction at 140°C increased to 49%, whereas the overall selectivity SX-BrPA decreased to 82%.

From the above shown experiments it was concluded that 120°C is the optimum Tr for the selective synthesis of 2-BrPA from LA in [MIMBS]Br. At this temperature, all side and consecutive reactions are suppressed to a very large extent and still acceptable reaction rates can be realized.

XLA

Y2-BrPA

Y3-BrPA 100

SX-BrPA / % /

80 X-BrPA

& S & 60 3-BrPA

, Y , 40 2-BrPA

, Y , 20

LA X

0 100 120 140 Temperature / °C

Figure 17: Temperature variation in the 2-BrPA synthesis from lactide using [MIMBS]Br as bromination agent; Conditions: 12.5 mmol lactide, 75 mmol IL, 75 mol% HBr loading grade of IL, water content of < 1.5 wt%, 5 h, ambient pressure.[C]

110

LA bromination using [MIMBS]Br - molar ratio variation

In a second set of experiments, the influence of the LA:IL molar ratio on the outcome of the lactide bromination reaction was investigated (see Figure 18). As the ionic liquid reaction matrix was prepared with low initial water content, it was assumed that an increasing amount of bromide IL should allow higher yields in 2-BrPA due to a more favorable position of the bromination/hydrolysis equilibrium. Consequently, LA:IL ratios from 1:1 to 1:6 were tested to confirm this expectation (see Figure 18).

Indeed, an increased Y2-BrPA was found when the LA:IL molar ratio increased from 1:1

(29%) to 1:3 (48%). This increase in yield goes hand in hand with an increase in selectivity to 2-BrPA as substrate decarbonylation and self-oligomerization are effectively suppressed at the 1:3 ratio. However, no further improvement in Y2-BrPA was observed when the LA:IL molar ratio was 1:6. Hence, a LA:IL molar ratio of 1:3 was found ideal for the bromination of lactide with [MIMBS]Br and is used in the following investigations. The decreased S2-BrPA in the bromination experiment with an LA:IL ratio of 1:1 is assumed to result from non-hydrolyzed substrate (lactide) and comparably high substrate concentration. The low amount of water in the reaction medium in this experiment is insufficient for complete lactide hydrolysis and the remaining intermediate may lead to additional bromination and oligomerization by-products, causing a drop in S2-BrPA. 4 Results and Discussion 111

XLA

Y2-BrPA

100 S2-BrPA

/ % / 2-BrPA

60 & S &

2-BrPA 40

, Y ,

LA X 20

0 1:1 1:3 1:6 LA:[MIMBS]Br molar ratio

Figure 18: Concentration-dependent 2-BrPA synthesis with [MIMBS]Br; Conditions: X mmol lactide, 75 mmol IL, ~ 90 mol% HBr loading grade of IL, water content of ~ 3 wt%, 5 h, 120°C, ambient pressure.[C]

Influence of water

The following experiments aim to shed more light on the role of water in the lactide bromination using [MIMBS]Br. In all experiments HBr was added as 48 wt% aqueous solution to the zwitterion followed by removal of water via azeotropic distillation (see

Experimental section for details) to defined water and HBr contents (see below).

To study the influence of water more systematically, two different sets of experiments were performed: First, four samples of pre-dried [MIMBS]Br were prepared (75 mmol MIMBS, HBr loading grade: 75 mol%, ~3 wt% initial water content) and varying amounts of additional H2O were added (0.5 to 4.0 g). The results of these experiments are shown in Figure 19 (top). They clearly confirm that the presence of increasing amounts of H2O negatively influences the bromination reaction. The more H2O added, the lower the Y2-BrPA after 5 h of reaction time was. An addition of 4 g

(17.1 wt%) of H2O, for instance, led to a significant drop of Y2-BrPA to 14%, compared to Y2-BrPA of 37% obtained by a water addition of only 0.5 g (3.8 wt%). In all of these 112

experiments, an exceptional S2-BrPA could be realized with no detectable formation of by-products. Therefore, the presence of large amounts of water does not affect the selectivity towards 2-BrPA.

In the second set of experiments, the drying times of the initial HBr-zwitterion solution was varied to adjust the residual water content of the IL. The residual water content of the brominating reagent was monitored by the amount of trapped water in the

Dean-Stark apparatus and subsequent Karl-Fischer titration of the prepared ILs. The results of the LA bromination with [MIMBS]Br of different H2O content produced by this loading method are shown in Figure 19 (bottom). In contrast to the results from the first set of experiments, no significant decrease of Y2-BrPA was observed, although the residual water content tripled along this series. It is noteworthy that the ILs used in these experiments had a much higher HBr-loading grade of approximately

90 mol% (determined via acid-base titration of collected water) resulting from the shorter drying times. From the results, it was concluded that the HBr loading of the zwitterion has a dominating influence on the resulting Y2-BrPA over the residual water content of the reaction matrix. Hence, to obtain an optimum brominating reagent for

LA, the water content of [MIMBS]Br should be as low as possible to overcome the equilibrium-limiting influence. High loading grade of HBr on the zwitterion is even more important as the detrimental effect of water is much less pronounced for systems with high HBr concentrations. 4 Results and Discussion 113

100 80 75 mol% HBr loading grade of IL 60

/ % / 20 3,8 5,9 10,1 13,7 17,1

2-BrPA 0 XLA 0 5 10 15 20

Y2-BrPA & S & S2-BrPA

100 2-BrPA 80

, Y , 90 mol% HBr loading grade of IL LA

X 60

20 3,5 5,5 7,5 12,5 0 0 5 10 15 20

H2O content of [MIMBS]Br / wt%

Figure 19: Influence of water content on 2-BrPA synthesis with [MIMBS]Br. Varying amounts of H2O added prior to bromination after complete drying procedure (top) and varied residual H2O content by incomplete drying procedure (bottom). Conditions: 12.5 mmol lactide, 75 mmol IL, 5 h, 120°C, ambient pressure.[C]

Then, it was attempted to increase the HBr loading to 100 mol% while keeping the water content of the system to a minimum. In the optimized HBr loading procedure, the MIMBS zwitterion was loaded with HBr(aq) and subsequently dried at 100 °C for

16 h to a residual water content below 1.5 wt%. The resulting HBr loading grade was determined via acid-base titration. HBr loss during the entire drying procedure corresponded to 24 mol% of the added HBr. Compensation of this loss was realized by adding again HBr in the form of its aqueous solution. The resulting IL contained

7.5 wt% of water and a formal HBr loading of 100 mol% (every sulfonic acid group protonated by HBr). By adding more aqueous HBr to the dried system, HBr loadings of 133 mol% were realized resulting in a water content of 14 wt% in this particular case. The results obtained with the [MIMBS]Br with different degrees of HBr loading and water have been compared under standard reaction conditions (120 °C, 5 h) and the corresponding results are shown in Figure 20. 114

XLA

Y2-BrPA

S2-BrPA 100 20 H2O content

80 16

/ % / 2-BrPA

60 12

& & S

2-BrPA O content O wt% /

40 8 2

, Y ,

LA X

20 4

0 0

75 % HBr-loading (1)90 % HBr-loading 100(2) % HBr-loading133 (3) % HBr-loading (4) HBr loading grade of IL bromination agent [MIMBS]Br

Figure 20: Influence of HBr loading grade and water content on 2-BrPA synthesis with [MIMBS]Br. Conditions: 12.5 mmol lactide, 75 mmol IL, 5 h, 120°C, ambient pressure. IL loading grade varied by preparation method (1, 2) and addition of HBr(aq) after drying procedure (3, 4).[C]

Increasing the HBr loading grade from 75 (Figure 20, entry 1) to 90 mol% (entry 2) led to an increase of Y2-BrPA from 35 to 48 % despite the higher H2O content of the higher- loaded reaction matrix. The fully-loaded IL (entry 3) achieved an even higher Y2-BrPA of 60 % with a S2-BrPA of approximately 100 %. No side-products were observed by

1H NMR spectroscopy and the mass balance was closed. However, over-saturation of

[MIMBS]Br (entry 4) did not result in higher 2-BrPA yield and significant lactide decarbonylation to AcH and CO was observed, decreasing S2-BrPA. This observation is attributed the fraction of “free HBr” (resulting from over-saturation of the zwitterion), increasing the acidity of the reaction medium and therefore decreasing S2-BrPA.

4 Results and Discussion 115

Influence of reaction time

As the next step, the optimization of the reaction time for the lactide bromination reaction was of interest. So far, all experiments were conducted for 5 h. For the time variation experiments, fully loaded [MIMBS]Br (100 mol% HBr loading, 7.5 wt% water) was used at 120 °C with a molar ratio of LA:IL of 1:3. The results are shown in

Figure 21.

XLA

Y2-BrPA

Y3-BrPA 100 SX-BrPA

80 / % /

2-BrPA 60 & S &

2-BrPA , Y ,

LA 20 X

1 3 5 8 24 Reaction time / h

Figure 21: Time-dependent 2-BrPA synthesis with [MIMBS]Br. Five independent bromination experiments with different reaction times. Conditions: 12.5 mmol lactide, 75 mmol IL, full HBr loading grade of IL by compensating HBr lost during drying procedure, IL water content of 7.5 wt%, 120°C.[C]

The results indicate that reaction times longer than 5 h do not significantly increase the 2-BrPA yield. Instead, lactide conversion increases only slightly at prolonged reaction time combined with an increasing amount of side products formed. Interest- ingly, after 24 h reaction time also some 3-BrPA is found in the reaction solution (3 %).

This indicates that over long reaction times some of the 2-BrPA product forms AA by dehydrobromination followed by hydrobromination to 3-BrPA even at the mild reaction temperatures of 120 °C. 116

2-BrPA isolation from aqueous IL phase via liquid-liquid extraction

All experiments so far indicated that the 2-BrPA yield in the [MIMBS]Br reaction system is limited to around 60 % but excellent selectivity could be achieved. A logic concept to drive the reaction to full conversion of the lactide substrate and 100 % 2-BrPA yield was therefore to isolate the 2-BrPA formed after 5 h from the reaction mixture, to subsequently regenerate the IL with fresh HBr, followed by further conversion of the residual lactide from the first run. For isolating 2-BrPA from the reaction mixture,

LLE was chosen. In-situ LLE was examined first but led to a decreased bromination performance (compare chapter 4.2.1.2). Therefore, ex-situ LLE was investigated to isolate the formed 2-BrPA from the IL-containing reaction matrix. To ensure a stable liquid-liquid biphasic system for the extraction experiment, the IL reaction matrix was quenched with water after the bromination experiment to form an aqueous IL-containing phase. The added water stayed in the IL phase and was subsequently removed via azeotropic distillation. In the screening experiments for the extraction of 2-BrPA from the aqueous IL phase, various organic solvents were tested. The results are shown in Table 4. Further information on the experimental procedure is given in the Experimental section.

Table 4: Results of ex-situ LLE solvent screening. [a]

Organic solvent K2-BrPA Tb of extractor at 1 atm. / °C

toluene 0.21 111

chloro benzene 0.23 132

bromo benzene 0.17 156

diphenyl ether 0.15 258

chloroform 1.18 61

dibutyl ether (DBE) 1.82 141

1-heptanol 28.1 175 [a] Extraction conditions: 1 mL of quenched product solution, 1 mL of organic solvent, 24 h, room temperature, ambient pressure; Tb = boiling temperature; see Experimental section for details on ex-situ LLE solvent screening. 4 Results and Discussion 117

The determined partition coefficients of 2-BrPA were in the range of 0.15 to 28.1, depending on the nature of the organic extractant. Whereas aromatics and aryl halides had low extraction ability for 2-BrPA, chloroform (1.18) and dibutyl ether (1.82) showed reasonable K2-BrPA and can therefore be used for product isolation. 1-heptanol, however, showed by far the highest K2-BrPA (28.1) of all tested extracting agents. How- ever, for this solvent, additional LA was detected in the extraction phase (KLA of 3.26) resulting in a selectivity of 8.62 in favor of 2-BrPA extraction. The other investigated extractors displayed high selectivity towards 2-BrPA, with no LA detectable in the

1H NMR spectra of the organic layer (details on 1H NMR measurements see Experi- mental section). Hence, 1-heptanol can be used not only for product and substrate isolation, but also for possible clean-up procedures of the IL reaction matrix after the bromination experiment.

Recycling of the zwitterionic reaction matrix

Finally, the recycling of the zwitterionic reaction matrix under optimized conditions was studied. A detailed overview of the recycling procedure is given in the Experi- mental section. In this recycling study, the formed product, 2-BrPA, was isolated after the bromination experiment using the two most attractive extraction solvents from our comparative study (see Table 4), DBE and 1-heptanol. As can be seen from Figure

22, the isolated amount of 2-BrPA (n2-BrPA) is constant over four bromination cycles using 1-heptanol for product isolation (n2-BrPA is within the error of the standard experiment, see Figure A 19, Appendix), indicating that HBr-reloaded zwitterionic- based molten salts maintain good activity during recycling. The same very encourag- ing recycling behavior was found for DBE as an extracting agent. Here, three recycling procedures were demonstrated. Note that in both recycling series, a drop of Y2-

BrPA was observed with increasing number of bromination experiments, due to incomplete substrate (LA) extraction from the IL. Non-extracted LA was included into the calculation of Y2-BrPA of the subsequent experiment. This led to a formally decreasing

Y2-BrPA with increasing number of bromination cycles. However, the constant, isolated amount (moles) of 2-BrPA in the organic phase clearly showed the recyclability of the brominating agent and reaction matrix over the performed cycles without any significant loss in HBr loading capacity. 118

isol. Y2-BrPA with 1-heptanol

isol. Y2-BrPA with DBE 25 70 isol. n2-BrPA with 1-heptanol

isol. n2-BrPA with DBE

20 60 / % /

/ mmol / 50

15 2-BrPA

2-BrPA 40 10

30 IsolatedY

Isolated n Isolated 5

1 2 3 4 Bromination cycle

Figure 22: The effect of recycled [MIMBS]Br with 1-heptanol and DBE on the isolated n2-BrPA and Y2-BrPA of bromination of LA to 2-BrPA. Lines are only given as a guide to the eye.

Comparison of commercial LA species

To keep the water content of the reaction matrix at a minimum, lactide (PURALACT®) was used as substrate in previous bromination experiments. This cyclic di-ester of LA can be seen as a water-free model substrate of LA, which is hydrolyzed by the residual

H2O of the IL, forming two equivalents of LA. Of course, a much cheaper alternative and bio-renewable feedstock for the production of 2-BrPA is PURAC® from Corbion. Two experiments with both starting materials were conducted under optimized reaction conditions (see Figure 23). 4 Results and Discussion 119

XLA

Y2-BrPA S 100 2-BrPA

/ % / 2-BrPA

60 & S &

2-BrPA 40

, Y ,

LA X 20

0 lactide Purac FCC 88 Substrate

Figure 23: Substrate-dependent 2-BrPA synthesis with [MIMBS]Br. Conditions: 25 mmol lactic acid (PU- RAC®) or 12.5 mmol lactide (PURALACT®), 75 mmol IL, full HBr loading grade of IL, water content of < 1.5 wt%, 5 h 120 C, ambient pressure.[C]

In contrast to the above described reaction starting from lactide, PURAC® as substrate showed a reduced Y2-BrPA of 42 %. This confirmed the previous observation of additional H2O present in the reaction matrix limiting the bromination equilibrium. PU-

RAC® contains 12 wt% of H2O that equals 120 g, or 6.67 moles, of H2O per kilogram of PURAC®, compared to 880 g or 9.77 moles of LA. Therefore, PURAC® as a substrate significantly increases the H2O content in the bromination medium, which decreases

Y2-BrPA compared to a bromination using lactide as substrate. However, almost no decrease in S2-BrPA was observed with PURAC® as substrate (S2-BrPA = 95 %). For large- scale applications, recycling of unconverted LA is aspired, and due to lower feedstock costs, PURAC® might therefore be a viable alternative to more expensive lactide. Fur- thermore, PURAC® is an aqueous solution, whereas lactide is a solid (Tm is approx.

95-100 °C), and thus can be pumped at room temperature, in contrast to lactide, which needs to be molten beforehand. The use of PURAC® can therefore save additional costs for pumps, heating of pipes and the general periphery of the plant. An economic 120 assessment of the two possible substrates is still needed, but -at first glance- the use of PURAC® feedstock for a renewable production of 2-BrPA appears to be an attractive alternative to lactide.

Summary of LA bromination with HBr loaded zwitterions

In summary, a novel method for the synthesis of 2-BrPA starting from renewable LA or lactide was developed, using SO3H-functionalized bromide ILs. In this concept, a zwitterionic-type molten salt serves as water-free HBr carrier and acidity buffer.

When the zwitterionic-type molten salt is loaded with HBr, the formed ionic liquid has shown to be a selective and efficient brominating agent and reaction matrix (solvent, proton source and buffer) for the bromination of LA to 2-BrPA. It was deduced that both the buffered acidity and the charge-stabilizing ability of the protic bromide

IL play an essential role to handle the sensitive character of LA. Especially, imidazole- based ILs showed exceptional performance in the synthesis of 2-BrPA. The best results were observed with [MIMBS]Br and [EIMBS]Br. Under optimized conditions for

[MIMBS]Br, we achieved a Y2-BrPA of 60 % with an exceptional selectivity of up to

100 %. This demonstrated a significantly improved performance compared to the use of HBr(aq), where decarbonlyation and self-oligomerization dominated at comparable reaction conditions (120 °C, 5 h, ambient pressure). Moreover, product and substrate isolation as well as brominating agent recycling have successfully been demonstrated.

4 Results and Discussion 121

4.2.2 Isomerization of 2-BrPA to 3-BrPA in bromide ILs

The conversion of 2-BrPA is the second step of the multi-step NADA process. Based upon the aforementioned results of LA bromination with HBr(aq) in [PBu4]Br (compare chapter 4.2.1 ) and the guiding results of the one-step NADA reaction system (compare chapter 4.1.2), 2-BrPA conversion in [PBu4]Br was investigated. On the basis of the postulated mechanism and reaction principle of NADA, the conversion of 2-BrPA in an IL reaction matrix was investigated without the use of an additional source of acidity as no acidic catalyst is needed in this step of the reaction sequence (see Scheme

22).

4.2.2.1 Isomerization of 2-BrPA to 3-BrPA in [PBu4]Br

An emerging isomerization of 2-BrPA to 3-BrPA was observed when the reactant

(2-BrPA) was heated to elevated temperatures in the bromide IL [PBu4]Br (see Scheme 28).

Scheme 28: Observed isomerization reaction of 2-BrPA to 3-BrPA in [PBu4]Br at elevated temperatures.

Due to this interesting finding, the reactivity and isomerization behavior of 2-BrPA was studied and will be discussed in this chapter. The formation of 3-BrPA from

2-BrPA was observed in a [PBu4]Br reaction matrix at 140 °C and confirmed via

1H NMR analysis. Figure 24 shows the partial 1H NMR spectra of 2-BrPA (1) and

[PBu4]Br (2), as well as two partial 1H NMR spectra of samples taken from the isomerization mixture after a reaction time of 3 (3) and 20 h (4). In addition to 2-BrPA

(4.4 ppm, q, 1H, 2; 1.5 ppm, d, 3H, 1) and [PBu4]Br (0.85 ppm, m, 8H, 6; 1.38 ppm, m,

16H, 4 and 5; 2.18 ppm, m, 12H, 3), 3-BrPA is detected (2.78 ppm, t, 3H, 8; 3.53 ppm, t, 3H, 7). After a reaction time of 20 h the determined molar ratio of 2-BrPA:3-BrPA is approximately 0.6. 122

Figure 24: Partial 1H NMR spectra of an 2-BrPA isomerization experiment with [PBu4]Br. 1H NMR of pure 2-BrPA (1), 1H NMR of pure [PBu4]Br (2, **traces of water) and 1H NMRs of samples taken at a reaction time of 3 (3) and 20 h (4). Experiment conducted at 140 °C and ambient pressure with a IL:2-

BrPA molar ratio of 1:1. *DMSO-d6 as solvent.

In the first part of this chapter, the molar C3-composition within the isomerization matrix was studied and used as qualitative approach to compare the isomerization activity at various reaction conditions. To validate the influence of concentrations, isomerization of 2-BrPA to 3-BrPA was carried out at different IL:reactant molar ratios. The screening experiments were conducted at a reaction temperature of 160 °C and the corresponding results are presented in Figure 25. 4 Results and Discussion 123

2-BrPA 100 3-BrPA AA

composition / mol% / composition

3 C

20 Molar Molar

0 0 1:3 1:1 3:1 6:1 9:1 12:1 15:1 IL:reactant molar ratio / -

Figure 25: Concentration-dependent 3-BrPA synthesis with [PBu4]Br; conditions: X mmol 2-BrPA,

133.5 mmol [PBu4]Br, 3 h, 160 °C, ambient pressure.

First of all, thermal stability of pure 2-BrPA in the investigated temperature range

(100-180 °C) was proven. When 2-BrPA is heated for three hours to 100, 120, 140, 160, and 180 °C without the [PBu4]Br melt, respectively, no substrate decomposition is observed. Additionally, not even traces of 3-BrPA are found in 1H NMR analysis (see

Figure A 20, Appendix), which was a first indication for the vital role of the bromide IL in terms of 2-BrPA isomerization. When 2-BrPA is used in triple excess compared to [PBu4]Br (1:3, Figure 25), small amounts of 3-BrPA are observed together with oligomers after 3 h. Substrate oligomerization is observable in the 1H NMR spectrum and probably resulting from self-esterification at high substrate concentration. In contrast, at higher dilution (≥1 eq. IL) no oligomers were found. Interestingly, a decreasing share of 2-BrPA within the molar C3-composition of the reaction mixture is observed with increasing amounts of [PBu4]Br. At an equimolar IL:2-BrPA molar ratio

(1:1, Figure 25), the share of 3-BrPA within the molar C3-composition is approximately 30 % after 3 h (67 % 2-BrPA). Additionally, about 3 % of the molar C3-compo- 124

sition of the reaction mixture is AA. Further excess of [PBu4]Br leads to further conversion of 2-BrPA. For instance, at a IL:2-BrPA molar ratio of 6:1, the share of 2-BrPA accounts for 17 % after 3 h whereas 3-BrPA (46 %) and AA (37 %) are preferably formed (6:1, Figure 25). With a further increase of the IL:2-BrPA molar ratio up to

15:1, the share of AA within the molar C3-composition of the isomerization mixture is enhanced to 60 % while 3-BrPA drops to 26 % and 2-BrPA reaches a constant value of approximately 14 %. However, pretended high AA amounts within the molar

C3-composition are relativized by the absolute YAA (quantitative experiments are shown and discussed below and in chapter 4.3.2.1) and atom efficiency of the reaction

(high IL excess is crucial).

Mainly the formation of the isomerization product 3-BrPA was observed when an equimolar IL:2-BrPA molar ratio was used, whereas an excess of [PBu4]Br led to the formation of AA and 3-BrPA. Based on the postulated mechanism of NADA (see

Scheme 22), AA formation is assumed to result from dehydrobromination of bromopropionic acids and 3-BrPA is assumed to form by HBr addition to the vinyl group of

AA. As the focus of this chapter is on the selective isomerization of 2- to 3-BrPA, an

IL:2-BrPA molar ratio of 1:1 was used for further investigations. Direct formation of

AA from 2-BrPA will be discussed in chapters 4.2.3 and 4.3.

In the following, we investigated the influence of the reaction temperature on the isomerization of 2-BrPA to 3-BrPA in [PBu4]Br at a IL:2-BrPA molar ratio of 1:1. The experiments were conducted for 20 h as the goal was to achieve high isomerization grades and to determine the related isomerization equilibrium (compared to the experiments shown above). The results of the temperature-dependent isomerization of

2- to 3-BrPA are shown in Figure 26.

At a reaction temperature of 100 °C neither isomerization (3-BrPA) nor dehydrobromination (AA) are observable within 20 h. When the reaction temperature is stepwise raised from 120 to 160 °C, the isomerization grade of 2-BrPA increases and consequently more 3-BrPA is formed. At a reaction temperature of 160°C, the molar

C3-composition is 12 mol% 2-BrPA, 86 mol% 3-BrPA and 2 mol% AA. Interestingly, the molar share of AA does not exceed a value of 2 mol% in the investigated temperature range from 100 – 180 °C when a IL:2-BrPA molar ratio of 1:1 is used. When the 4 Results and Discussion 125

reaction temperature is set to 180 °C, no further increase in 3-BrPA share is observed, hinting at an already reached limit of the isomerization grade at 160 °C. Moreover, arising side reactions, e.g. decarbonylation, cannot be excluded in this experiment.

Hence, 160 °C displays the most suitable reaction temperature for a selective isomerization of 2- to 3-BrPA in [PBu4]Br. The calculated yields of all isomerization experiments are summarized in Table A 5 and Table A 6 (see Appendix).

2-BrPA 100 3-BrPA AA

composition / mol% / composition

3 C

20 Molar Molar

0 100 120 140 160 180 Reaction temperature / °C

Figure 26: Temperature-dependent 3-BrPA synthesis with [PBu4]Br; conditions: 133.5 mmol 2-BrPA,

133.5 mmol [PBu4]Br, 20 h, ambient pressure.

In addition to the qualitative approach of 2-BrPA conversion presented above, a quantitative and time-resolved experiment was monitored at so far optimized isomerization conditions (IL:2-BrPA molar ratio of 1:1, Tr = 160 °C). The corresponding results are presented in Figure 27. The results of a quantitative and time-resolved isomerization experiment at 180 °C can be found in the Appendix (Figure A 21). 126

2-BrPA 3-BrPA 100 AA LA

C3 balance 80

40 Amountmmol /

0 0 5 10 15 20 25 Reaction time / h

Figure 27: Quantitative and time-resolved isomerization of 2-BrPA to 3-BrPA in [PBu4]Br. Conditions:

IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol [PBu4]Br , 24 h, 160 °C, ambient pressure. Error bars illustrate the determined deviation of the C3-balance resulting from quantification errors within the experimental setup.

The time-resolved and quantitative evaluation of the so far optimized isomerization of 2-BrPA to 3-BrPA (Figure 27) shows moderate conversion of 2-BrPA 3-BrPA at

160 °C, reaching a maximum Y3-BrPA of 63 % after 24 h. The remaining amount of

2-BrPA at the end of the reaction is approximately 10 mmol (X2-BrPA = 90 %). LA present from the beginning of the experiment, which is probably formed by 2-BrPA hydrolysis with residual water in the substrate, is slowly degraded over the course of the reaction. Interestingly, the amount of AA is in the range of 3 mmol over the entire reaction time. The course of the molar amounts indicates the presence of two consecutive reactions whereat the intermediate is much more reactive than the reactant

(k1/k2 ≪ 1). Here, dehydrobromination of 2-BrPA triggered in [PBu4]Br leads to the formation of AA and HBr in the first step of the reaction sequence. The subsequent HBr addition, which proceeds in anti-Markovnikov addition fashion (due to the strong inductive effect of the carboxyl group), is comparatively fast. Hence, a quasi-station- ary behavior of AA is observed. However, as can be seen by the decrease of C3-species 4 Results and Discussion 127

over the course of the reaction, reaching a minimum value of approximately

77 mmoles after 24 h, competing pathways are present within the isomerization sequence (decomposition / polymerization). Figure 28 is used to further illustrate the observed isomerization behavior.

100 75 69.3

90 51.8

45 / % /

80 3-BrPA

S -balancemmol /

3 79.1 15 C 77.0

0 70

0 4 8 12 16 20 24 28 Reaction time / h

Figure 28: Time-resolved C3-balance and S3-BrPA of 2-BrPA conversion in [PBu4]Br. conditions: IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol [PBu4]Br, 24 h, 160 °C, ambient pressure. Error bars illustrate the determined deviation of the C3-balance resulting from quantification errors within the experimental setup.

Within the first hour of the reaction an increasing S3-BrPA from 4 to approximately 50 % is observed. As 3-BrPA is the product of the second reaction step within the assumed isomerization sequence, this observation can be explained by the formed intermediate AA limiting S3-BrPA in the beginning of the reaction. With increasing AA concentration, 3-BrPA formation by HBr addition to the vinyl group of AA enhances, which leads to an increased S3-BrPA. Interestingly, S3-BrPA is almost constant for the next three hours (approximately 50 %) while C3-species are lost over the course of the reaction

(see shaded region in Figure 28). This can be seen by a decrease of detectable C3-species from approximately 94 to 79 %. It can be assumed that the present concentration 128 regime within the isomerization sequence between 1 and 4 h enables both 3-BrPA formation and side reaction(s) (decomposition/polymerization) with similar reaction rates. However, C3-loss is inhibited after approximately 4 h and no significant decrease in the C3-balance was observed within the next 20 h of the reaction. In contrast,

S3-BrPA increases again from approximately 52 to 70 % until the end of the experiment.

This indicates that either side reactions are inhibited leading to an increased 3-BrPA formation or undetectable intermediates (e.g. oligomers/polymers) are further converted to 3-BrPA leading to a balanced detection of C3-species.

4.2.2.2 Variation of the IL isomerization matrix – the role of the bromide

In previous isomerization experiments, [PBu4]Br was used as reaction matrix for the conversion of 2-BrPA to 3-BrPA. In order to clarify the role of the bromide anion and ionic liquid, the isomerization matrix was further investigated. In the following screening, different isomerization solvents were investigated in their ability to convert 2-BrPA to 3-BrPA. The corresponding results are summarized in Table 5. The experiments were conducted at a targeted reaction temperature of 160 °C and the molar C3-composition of the reaction mixture was used to compare the isomerization behavior of the different systems.

4 Results and Discussion 129

Table 5: Results of isomerization matrix screening.[a]

Molar C3-composition / mol%

Entry Isomerization matrix Tr / °C 2-BrPA 3-BrPA AA LA olig.

1 [PBu4]Br 160 67 30 3 0 0

2 none 160 100 0 0 0 0

3 H2O 115 59 0 0 41 0

4 H2O/CaBr2 130 62 1 0 37 0

5 CaBr2 160 61 25 0 14 0

6 [PBu4][NTf2] 160 99 0 0 0 1

7 [PBu4][NTf2] / [PBu4]Br 160 32 19 39 3 7

[a]Conditions: 20 h; ambient pressure.

As can be seen in Table 5, the conversion of 2-BrPA in [PBu4]Br (entry 1) represents the benchmark experiment for the screening of alternative isomerization solvents. As already announced above, no conversion of 2-BrPA is observed at 160 °C without the use of additional reactants (entry 2). In contrast, an aqueous solution of 2-BrPA (solvent:reactant molar ratio of 5:1) forms 41 mol% of LA but no reactant isomerization or dehydrobromination is observed (entry 3). The reaction temperature in water was limited to 115°C due to the boiling temperature of the reaction system at ambient pressure. When an equimolar amount of CaBr2 is added (entry 4), traces of 3-BrPA are observable in the 1H NMR spectrum but the main share of the reaction mixture is of 2-BrPA (62 mol%) and LA (37 mol%). Hence, LA formation is strongly favored in aqueous systems. Thus, in yet another experiment, an equimolar and largely water- free mixture of 2-BrPA and CaBr2 (entry 5) was investigated (no further addition of

H2O). Even though the reactant conversion remains unchanged (61 mol% 2-BrPA) the product ratio changes substantial. 25 mol% of 3-BrPA are found whereas the share of

LA decreases to 14 mol%. Unfortunately, the use of a water-free 2-BrPA / CaBr2 leads to solidification of the reaction mixture (at 160 °C) due to a lack of solvent. Obviously, 130 this indicates the role of the IL, which enables high reactant conversions in the liquid phase and a water-free bromide source in one component. To underline the importance of the bromide anion, an experiment with [PBu4][NTf2] (entry 6) was conducted at benchmark reaction conditions. [PBu4][NTf2] was chosen as low melting phosphonium-based salt of high thermal stability. As can be seen in Table 5, neither isomerization nor dehydrobromination is observed. In contrast, further addition of an equimolar amount of [PBu4]Br activates the isomerization matrix (entry 7).

19 mol% 3-BrPA and 39 mol% of AA are observed, indicating the dehydrobromination and isomerization of 2-BrPA. Compared to the benchmark experiment (entry 1), the AA:3-BrPA ratio is shifted to the dehydrobrominated side. This observation is in good accordance with the results of the [PBu4]Br:2-BrPA molar ratio variation (compare Figure 25).

4.2.2.3 Cation variation of the IL used as isomerization matrix

In previous isomerization experiments, the essential role of bromide anion of the IL, used as isomerization matrix, was demonstrated. In this set of experiments, the influence of the cation of the bromide IL was investigated. For these experiments, the optimized IL:2-BrPA molar ratio of 1:1 was selected. To ensure thermal stability of all used cations, the experiments were conducted at 140 °C. In order to achieve high quantities of 3-BrPA, the reaction time of the isomerization experiments was 20 h. The isomerization activity of the investigated ILs is depicted in

Figure 29. An overview of the investigated IL structures and the corresponding nomenclature is given in Figure A 22 (see Appendix). 4 Results and Discussion 131

X2-BrPA 100 Y3-BrPA 80 YAA 90 S3-BrPA

/ % / 60 AA

& Y & / % /

3-BrPA 60

3-BrPA

, Y , S

2-BrPA 50

X 20

0 30

[PBu4]Br [NBu4]Br[1Bu4MePyr]Br[1Et1MePyrro]Br[EMIM]Br [MIMBS]Br Ionic liquid

Figure 29: Cation variation for 2-BrPA conversion in bromide ionic liquids; Conditions: IL:2-BrPA molar ratio of 1:1, 140 °C, 20 h, ambient pressure. [PBu4]Br = tetrabutylphosphonium bromide, [NBu4]Br = tetrabutylammonium bromide, [1Bu4MePyr]Br = 1-butyl-4-methyl-pyridinium bromide, [1Et1MePyrro]Br = 1-ethyl-1-methyl-pyrrolidinium bromide, [EMIM]Br = 1-ethyl-3-methylimidazolium bromide, [MIMBS]Br = 3-methyl-1-(4-butane sulfonic aicd) imidazolium bromide.

2-BrPA conversion in [PBu4]Br at 140 °C is the benchmark for this set of experiments

(X2-BrPA = 54 %, Y3-BrPA = 33 %, YAA = 2 %, S3-BrPA = 62 %). Despite comparable Y3-BrPA

(38 %) and S3-BrPA (58 %) with additional reasonable amounts of AA (YAA = 14 %), tetrabutylammonium bromide ([NBu4]Br) is characterized as non-suitable isomerization matrix, due to thermal instability of the IL. In contrast, pyridinium and pyrrolidinium-based cations, namely 1-butyl-4-methyl-pyridinium bromide

([1Bu4MePyr]Br) and 1-ethyl-1-methyl-pyrrolidinium bromide ([1Et1MePyrro]Br) show significantly enhanced isomerization selectivity of approximately 80 %. In addition, the corresponding yields of 3-BrPA are significantly increased to 59 %

([1Bu4MePyr]Br) and 67 % ([1Et1MePyrro]Br). In both cases, YAA does not exceed 3 %.

The best isomerization performance is achieved when 1-ethyl-3-methylimidazolium bromide ([EMIM]Br) is used as reaction matrix. Y3-BrPA is enhanced to 70 % in high selectivity of 90 %. 3-methyl-1-(4-butane sulfonic acid) imidazolium bromide 132

([MIMBS]Br), which was established as selective reaction matrix and HBr carrier in the bromination of LA (see chapter 4.2.1.3), shows the lowest Y3-BrPA (20 %) and S3-BrPA

(36 %). The significantly decreased isomerization performance is assumed to result from the increased acidity of the reaction matrix, leading to C3-decomposition at reaction temperatures of < 120°C.

While the use of [NBu4]Br and [MIMBS]Br as reaction matrix did not enhance the isomerization performance, the use of [1Bu4MePyr]Br, [1Et1MePyrro]Br and

[EMIM]Br led to an improvement of 2-BrPA isomerization at 140°C. It is assumed that the increased reactivity and selectivity of the reaction can be correlated with the polarity of the isomerization matrix. According to Reichardt[145], the overall solvation capability of the investigated ILs, characterized by ET(30) or ETN, is substantially different. The determined ET(30) (ETN) of [PBu4]Br is 43.5 (0.395), which is in the range of dipolar non-hydrogen-bond donor (“aprotic”) solvents like DMF (ETN = 0.386) or acetone (ET(30) = 42). In contrast, polarity of 1-methyl-3-alkylimidazolium salts, 1-alkyl- and 1,4-dialkylpyridinium salts and 1-methyl-1-(2-methoxyethyl)pyrrolidinium salts corresponds to that of short-chain primary or secondary alcohols and secondary amines (e.g. ethanol or N-methylformamide) with typical ET(30) (ETN) of approximately 52 (0.66). The significantly increased polarity is assumed to influence the chemical equilibrium and reaction rate, for instance, through improved solvation capability for reactants, products and transition states. As conversion of 2-BrPA to

3-BrPA proceeds via consecutive steps including charged transition states, solvation power is assumed to correlate with increased overall selectivity of the reaction.[145]

4.2.2.4 Product isolation from isomerization matrix via distillation

After successful optimization of 2-BrPA isomerization to 3-BrPA in bromide ionic liquids, an efficient isolation is necessary to separate the formed compounds (3-BrPA and AA) from the IL reaction matrix. Particularly with regard to the multi-step

NADA process, isolated 3-BrPA will be converted in a followed process step (see chapter 4.2.3). Therefore, subsequent vacuum distillation was used to remove formed products from the isomerization mixture and to simultaneously purify the ionic liquid reaction matrix for potential reuse. Qualitative distillation experiments showed 4 Results and Discussion 133

that removal of products (AA, 3-BrPA) and unconverted substrate (2-BrPA) is possible at elevated temperatures (100-160 °C) and reduced pressures (2-10 mbar) for the investigated and thermally stable ILs. In Figure 30, an isomerization experiment and the followed isolation / purification is presented ([EMIM]Br:2-BrPA molar ratio of 1:1,

Tr = 160 °C, tr = 20 h), which was followed by 1H NMR analysis.

Figure 30: Partial 1H NMR spectra of an 2-BrPA isomerization experiment with [EMIM]Br and subsequent product isolation via vacuum distillation. 1H NMR of isomerization mixture (1), 1H NMR of purified isomerization mixture (2), and 1H NMRs of received distillate (3). Experiment conducted at 140 °C for 20 h (ambient pressure) with a IL:2-BrPA molar ratio of 1:1. Isolation conducted at 2-10 mbar and a stepwise heated temperature ramp from (100-160 °C) for 30 min, respectively. * DMSO-d6 as solvent. ** [EMIM]Br 1H NMR signals.

As can be seen in Figure 30, C3-compounds are successfully isolated from the

[EMIM]Br isomerization mixture (1) via vacuum distillation (10-2 mbar , 160°C), leading to pure [EMIM]Br (2) with only small traces of 3-BrPA observable, and a C3-rich product distillate (3). Purified IL can be reused for further isomerization experiments.

The 1H NMR of the distillate (3) shows unconverted 2-BrPA, which can theoretically be recycled and the product of the isomerization, namely 3-BrPA. Interestingly, the distillate contains considerable amounts of AA, which is assumed to form during the isolation procedure. This observation will be considered and discussed in chapter 4.3. 134

4.2.2.5 Summary of 2-BrPA isomerization study

The isomerization of 2-BrPA to 3-BrPA triggered in [PBu4]Br was successfully demonstrated. Additionally, the essential role of the bromide anion of the ionic liquid, which initiates the isomerization, was shown. The findings indicate a consecutive reaction sequence via dehydrobromination of 2-BrPA to AA and subsequent addition of HBr to AA forming 3-BrPA. The variation of the IL:2-BrPA molar ratio validate the strong influence of the molar composition of the reaction mixture on the isomerization behavior of 2-BrPA. When equimolar amounts of IL and 2-BrPA are used, an isomerization of 2-to 3-BrPA was achieved as 3-BrPA was the main product in the qualitative isomerization experiments. When an IL excess was used, the share of AA within the molar C3-composition increased significantly, which indicated an inhibited activity for HBr addition to the intermediate of the postulated isomerization sequence. How- ever, quantitative evaluation revealed decomposition and/or oligomerization of C3- species, even under optimized conditions for [PBu4]Br, limiting S3-BrPA to approximately 70 %. Moreover, it was shown that enhanced isomerization grades were accessible when the cation of the bromide IL was changed. The best isomerization performance was achieved with e.g. [EMIM]Br as isomerization matrix. Y3-BrPA was enhanced to 70 % in high selectivity of 90 % (IL:2-BrPA molar ratio of 1:1, 140 °C, 20 h). Furthermore, product isolation as well as purification of the isomerization matrix was successful via vacuum distillation.

4 Results and Discussion 135

4.2.3 Dehydrobromination of 3-BrPA to AA

This chapter provides an overview of the development and investigations on the third reaction of the multi-step NADA process, namely 3-BrPA dehydrobromination to

AA. As described in chapter 2.1.3, the dehydration of 3-HP showed high YAA

(97-98 %). The enhanced selectivity of the reaction compared to LA dehydration is attributed to the elimination mechanisms. Bearing in mind the results of 3-HP dehydration, selective and efficient conversion of 3-BrPA to AA is expected.

For potential use in bio-AA production, atom efficiency of the dehydrobromination of 3-BrPA is crucial. The prerequisites for AA formation from 3-BrPA are efficient and selective HBr elimination, recyclability of the reaction matrix and a safe-handling of

HBr within a closed HBr cycle. A typical protocol for HBr elimination from organic compounds would include refluxing the substrate with a strong base (e.g. ethanolic potassium hydroxide). The hydroxide ion acts as base, removing a hydrogen ion from the carbon atom. The resulting rearrangement of electrons expels the bromine as bromide ion, forming a vinyl group. However, HBr elimination is correlated to the formation of stoichiometric amounts of salt (e.g. KBr). In case of using alkaline earth metal bases or alkali metal bases, the halogen value cannot be recovered in a usable form. Moreover, in most of these processes, recovery of the formed salt is restricted, limiting the atom efficiency of the process.

136

4.2.3.1 3-BrPA conversion in trioctylamine

In this study, elimination of HBr from 3-BrPA in trioctylamine (TOA) was investigated. Liquid state over a broad temperature range and good thermal stability of TOA enable proper handling for liquid-phase dehydrobromination conditions. Stoichio- metric reaction of 3-BrPA with TOA should form AA via base-assisted elimination of

HBr. Simultaneously, the amine “catches” the generated HBr forming an amine salt: trioctylammonium bromide ([TOAH]Br). Amine recovery as well as acid liberation from the amine hydrohalide could for instance be implemented by thermal decomposition of the salt. This procedure is assumed to enable safe and easy HBr storage as well as HBr recovery without (in)organic by-product waste. The concept of base-assisted dehydrobromination of 3-BrPA and the related HBr cycle is illustrated in

Scheme 29.

Scheme 29: Base-assisted 3-BrPA dehydrobromination using TOA and related salt ([TOAH]Br) and HBr recycling as the third and last reaction step of a multi-step NADA sequence.

The first investigation of base-assisted 3-BrPA dehydrobromination with TOA proved a fast and selective conversion of 3-BrPA to AA at mild reaction conditions.

The experiment was conducted at 80 °C with a 3-BrPA:TOA molar ratio of 1:1 and was monitored by 1H NMR analysis. The corresponding results of the time-resolved dehydrobromination of 3-BrPA with TOA are shown in Figure 31. 4 Results and Discussion 137

X3-BrPA

YAA 100 SAA

95 / % /

90 AA

& S & 85

AA , Y ,

3-BrPA X 75

70 0 1 2 3 4 5 Reaction time / h Fig- ure 31: Time-resolved dehydrobromination of 3-BrPA to AA in TOA. Conditions: 50 mmol 3-BrPA, 50 mmol TOA, 80 °C, 5 h, ambient pressure.

Fast substrate conversion (X3-BrPA) of 92 % was observed after the first 15 minutes of the experiment at 80 °C. Within the first hour of the dehydrobromination experiment,

X3-BrPA reaches a maximum value of 96 %, which was determined to be the equilibrium limitation of 3-BrPA dehydrobromination with TOA at the chosen reaction conditions

(80°C, 3-BrPA:TOA molar ratio of 1:1). The yield of acrylic acid (YAA) after 15 minutes is 84 % and constant within the first hour of the reaction. After 2 h, YAA decreases to

78 % and from there on remains constant (78 – 81 %) within the remaining investigated reaction time of 5 h. The selectivity towards acrylic acid (SAA) continuously decreases within the first two hours of the dehydrobromination experiment. High SAA of approximately 91 % was determined after 15 minutes from the start of the experiment. Within the first hour of the reaction, the selectivity exceeds 87 % while a drop of selectivity in AA to approximately 80 % was observed after 2 h. Moreover, dimerization of AA to 3-acryloxypropionic acid (diacrylic acid, DiAA, C6H8O4) was observed with 1H NMR analysis, which is assumed to result from high AA concentrations in the basic reaction environment. Figure 32 illustrates the calculated combined selectivity in AA (SAA-combined), comprising SAA and SDAA. 138

SAA

SDiAA S 100 AA-combined

/ % / 80

AA-combined , S ,

DiAA , S ,

AA 20 S

0 0 1 2 3 4 5 Reaction time / h

Figure 32: Time-resolved dehydrobromination selectivity of 3-BrPA to AA in TOA. Conditions: 50 mmol 3-BrPA, 50 mmol TOA, 80 °C, 5 h, ambient pressure.

As can be seen in Figure 32, dimerization of AA is already observed after 15 minutes

(YDiAA = 9 %). The determined monomer:dimer ratio is approximately 10 within the first hour of the experiments. Due to decreasing quantity of monomeric AA at longer reaction times (YAA drops to 78 % after 2 h), the momomer:dimer ratio also decreases to approximately 8 at reaction times ≥ 2 h. Including the determined amount of DiAA in form of a combined selectivity of 3-BrPA dehydrobromination to AA (SAA-combined) gives approximately 100 % after 15 minutes of the experiment. SAA-combined is ≥ 97 % within the first hour, illustrating the exceptional selectivity of the dehydrobromination of 3-BrPA to AA at short reaction times. SAA-combined slowly decreases over the course of the reaction to approximately 91 %. It is assumed that longer-chain AA oligomers, which are not quantitatively detected with 1H NMR, are the reason for de- clining selectivity in AA. However, the results clarify that exceptional selectivity in

AA is reached when reaction time is shortened to ≤ 1 h.

4 Results and Discussion 139

AA dimerization proceeds via an ionic mechanism where diacrylic acid (DiAA) is formed by intermolecular Michael-type addition. As AA ionization is strongly favored in the alkaline reaction medium (TOA) at 80 °C, fast AA dimerization is even observed after short reaction times. The observed dimerization network is shown in

Scheme 30.

Scheme 30: Postulated reaction network of 3-BrPA dehydrobromination in alkaline environment at elevated temperatures indicating formation of DiAA and AA oligomers from intermolecular Michael-type addition via AA ionization.

To minimize AA oligomerization and simultaneously separate the formed product

(monomeric AA) from the reaction medium, a semi-batchwise dehydrobromination experiment was performed, where AA was isolated via distillation at reduced pressure. The experiment was conducted at 150 °C and 100 mbar to induce AA evaporation and was performed for 30 min as previous experiments showed exceptional SAA at short reaction times. The determined quantities characterizing the semi-batchwise dehydrobromination experiment are summarized in Table 6.

140

Table 6: Determined quantities of semi-batchwise dehydrobromination of 3-BrPA at 150 °C and 100 mbar.

Reaction flask Distillation flask Quantity Overall non-isolated isolated

YAA / % 58 18 76

YDiAA / % 20 - 20

X3-BrPA - - 96

SAA, combined - - ~ 100

The results presented in Table 6 show that a 3-BrPA conversion of 96 % is achieved within 30 minutes at a reaction temperature of 150 °C and a pressure of 100 mbar. The isolated yield of AA, which was determined via q1H NMR of the distillate, is 18 %.

The mass balance can be closed with 1H NMR analysis and the dehydrobromination is characterized as exceptional selective in AA/DiAA (SAA, combined ~100). However, dimerization is not prevented by in-situ AA isolation. Even though AA concentration is reduced by AA removal from the reaction medium, the monomer:dimer ratio in the reaction flask is approximately 3:1, with YDiAA accounting for approximately 20 %. It is assumed that AA dimerization at semi-batchwise reaction conditions is due to the increased reaction temperature compared to the previous batchwise dehydrobromination experiment. Nevertheless, the semi-batchwise dehydrobromination of

3-BrPA shows that AA can be successfully isolated from the TOA/[TOAH]Br reaction medium. Obviously, the reported conditions are far from an optimized procedure but shall be noted as proof of concept for in-situ AA isolation.

4 Results and Discussion 141

4.2.3.2 HBr recovery from [TOAH]Br

After demonstrating a selective base-assisted dehydrobromination of 3-BrPA to AA and DiAA at mild reaction conditions (80 °C, ambient pressure), HBr recovery from the amine hydrohalide is necessary as final step of the overall reaction sequence to close the HBr cycle within the multi-step NADA process. HBr recovery from

[TOAH]Br is briefly touched in this chapter.

Thermally-induced HBr recovery as well as HBr recovery via extraction from

[TOAH]Br were briefly investigated in this study. Details on the experimental procedure can be seen in the Experimental section. Representative results of the HBr recovery study are listed in Table 7.

Table 7: Results of HBr recovery study via thermally-induced HBr release and HBr extraction.

HBr Entry tr / h Tr / °C Ar strip gas / mL/min recovery / % T-1 3 180 300 4.0 T-2 4 200 50 1.5 LLE-1 24 25 - 3.6 LLE-2 24 100 - 7.0

Thermally-induced HBr release from [TOAH]Br showed a maximum HBr recovery of 4 % after 3 hours at 180 °C with an argon flow of 300 mL/min

(entry T-1 in Table 7). This indicated that salt dissociation and related HBr liberation is not sufficiently achieved within the investigated temperature range (up to 200 °C).

The second approach for HBr recovery was HBr extraction from [TOAH]Br into a suitable solvent. Due to high HBr solubility as well as economic and environmental reasons, water was chosen as extracting agent. While the extraction experiment at room temperature did not improve the quantity of liberated HBr from [TOAH]Br (entry LLE-1 in Table 7), an extraction experiment at approximately 100 °C showed 7 % recovered HBr in the aqueous phase (entry LLE-2 in Table 7). However, the results of HBr extraction from [TOAH]Br are assumed to be limited to the protonation equilibrium of [TOAH]Br. Therefore, HBr extraction at elevated temperatures did not sufficiently enhance the performance of the HBr recovery experiment. 142

Even though TOA was well suited for liquid-phase dehydrobromination of 3-BrPA at first glance and also showed excellent performance in AA formation via HBr elimination, first investigations of HBr recovery from the amine hydrohalide showed to be nontrivial. Nevertheless, the mild reaction conditions required for AA formation from 3-BrPA (80 °C, ambient pressure) may also enable the use of alternative and short-chain amines (e.g. triethylamine or tributylamine) or other less-stable bases as reaction medium for base-assisted elimination of HBr from 3-BrPA. These compounds may be able to provide a simplified procedure for amine recycling and HBr recovery. Unfortunately, it was not possible within the time frame of this PhD thesis to test further compounds for the base-assisted elimination of HBr from 3-BrPA and subsequent recovery of HBr.

However, the presented results of HBr recovery should not be seen as a “no-go crite- ria” for a closed HBr cycle within the multi-step NADA process. Even though HBr recovery in above presented experiments was limited to approximately 7 %, literature claims different approaches to overcome the observed limitation. One approach for the recovery of amines and volatile acids from amine salts is described in a patent from ARCO Chemical Technology.[146] In this process, an amine salt reacts with a non- volatile acid to liberate the volatile acid in a first step. The amine can be liberated in a second step by thermal decomposition of formed amine/non-volatile acid salt.[146]

Good results of e.g. HCl recovery from diethylammonium chloride were reported when phosphoric acid (H3PO4, 99 %) or pyrophosphoric acid (H4P2O7, 98 %) are used as non-volatile acids. The reported temperature range for HCl recovery is

90-150 °C.[146] Moreover, the Dow Chemical Company describes the use of reversible bases for the recovering of protic acids.[147] In this context, reversible base refers to an aromatic compound containing at least one sterically-hindered nitrogen atom. This reversible base forms a salt with a strong protic acid, which undergoes dissociation to the aromatic compound and the protic acid at temperatures below the decomposition temperature of the reversible base.[147] A promising candidate with a reported

HBr recovery of 81 % at 125 °C within one hour is 2,6-diphenylpyridine.[147]

4 Results and Discussion 143

4.2.4 Evaluation of the multi-step NADA process

Based on the postulated reaction sequence of the NADA concept and motivated by the limitations along with LA feedstock (PURAC®) in the one-step NADA process, three spatially separated reactions were investigated and the results were described in the previous chapters. In general, all reaction steps along LA conversion to bio-AA via brominated LA species were successfully developed and largely optimized. In the following, an overview of the developed multi-step NADA process is given. Bench- marking with the one-step NADA process using LA as starting material (see chapter

4.1.2) was used to evaluate the multi-step processing of the NADA concept.

The first step of the multi-step NADA process, namely LA bromination to 2-BrPA, was enabled by the use of zwitterionic HBr carriers, serving as bromination agent and reaction medium. The best results were obtained with the protic IL [MIMBS]Br, reaching 60 % yield to 2-BrPA in 100 % selectivity. Optimized bromination performance for [MIMBS]Br was achieved with lactide as starting material, at a reaction temperature of 120 °C, 5 h of reaction time and at ambient pressure. The best performing

[MIMBS]Br:LA molar ratio was 3:1. Product isolation was achieved by LLE with high extraction selectivity in 2-BrPA using 1-heptanol (see chapter 4.2.1.3).

Secondly, we demonstrated the transformation of 2-BrPA to 3-BrPA, which is triggered in aprotic bromide ILs. The best isomerization performance was demonstrated for [EMIM]Br reaching 70 % yield to 3-BrPA in 90 % selectivity. Here, the reaction was performed at a temperature of 140°C and ambient pressure. The

[EMIM]Br:2-BrPA molar ratio was 1:1. The reaction was operated for 20 °h and was not further optimized within this study, allowing a possibility for further optimizing the isomerization productivity. Isolation of the formed product 3-BrPA was successfully demonstrated via distillation at reduced pressure (see chapter 4.2.2).

The third step of the developed reaction sequence, namely dehydrobromination of

3-BrPA to AA, was performed in TOA. On the one hand, the reaction medium TOA served as base, initiating HBr elimination from 2-BrPA. On the other hand, TOA “catched” the released HBr forming the salt [TOAH]Br and therefore enabling HBr storage. High rate dehydrobromination of 3-BrPA was achieved at comparably mild 144 reaction conditions (80 °C, ambient pressure). The used TOA:3-BrPA molar ratio was again 1:1. The reaction was characterized by high selectivity towards AA. After

15 min, YAA was 84 % (SAA = 91 %) and the only by-product found was DiAA, formed by product dimerization (YDiAA = 8 %, SDiAA = 9 %). Bearing in mind that DiAA is formed in a consecutive intermolecular reaction step of two AA monomers, the combined selectivity SAA, combined (SAA + SDiAA) was 100 %. AA isolation from the reaction mixture was enabled by distillation at reduced pressure. However, HBr recovery from the formed salt [TOAH]Br, which is essential to close the HBr cycle within the multi-step NADA process, was demonstrated to be a bottleneck of the current process. Maximum HBr recovery achieved with the base/salt pair TOA/[TOAH]Br was

7 % (see chapter 4.2.3).

A process overview of the developed multi-step NADA process, representing a method for liquid-phase bio-AA production from LA, is shown in Scheme 31:

tr=5h tr=20h tr<0.5h Tr=120°C Tr=140°C Tr=80°C ambientpressure ambientpressure ambientpressure

LA 2-BrPA 3-BrPA [TOAH]Br protic IL aprotic IL (HBr) (Br-) TOA

HBr recovery loop

Step 1 Step 2 Step 3 LA bromination 2-BrPA isomerization to 3-BrPA 3-BrPA dehydrobromination followed by LLE followe d by distillation followed by distillation

Scheme 31: Overview of developed multi-step NADA process.

To evaluate the multi-step process in comparison with the one-step process, characteristic quantities of the latter are compared with the obtained results within this study in Table 8. 4 Results and Discussion 145

Table 8: Comparison of one- and multi-step NADA process in terms of product yield and selectivity.

NADA process operation Quantity 2-BrPA 3-BrPA AA

Yproduct / % 2 6 30 one-step Sproduct / % 2 8 37

Yproduct / % 60 (42)[a] 70 92 multi-step Sproduct / % 100 (95)[a] 90 100

[a] Yield of 2-BrPA and selectivity for PURALACT® as starting material of bromination compared to PURAC® (value in brackets).

By spatial separation of the individual reaction steps and adjustment of the chemical environment and related reaction conditions, respectively, the NADA process was considerably improved. High product selectivity of each individual step combine to a decisively increased NADA process selectivity (SNADA, see equation (11)) for defini- tion). The calculated and theoretical NADA process selectivity for both lactide and

LA as substrate are given in equation (12) and (13). While the overall product yield

(YAA) remained in comparable order (39 % for lactide and 27 % for LA compared to

30 % of one-step NADA process), product selectivity, which was the bottleneck of the one-step NADA, was increased to 90.0 % (85.5 %) for lactide (LA) as process feedstock.

SNADA / % �� × �� × �� (11)

SNADA (lactide) / %: � × �. � × � = ��. � % (12)

SNADA (LA) / %: �. �� × �. � × � = ��. � % (13)

Obviously, multi-step processing of the NADA technology is accompanied with additional process costs for two additional reactors and separation units. However, the possibility of substrate recycling due to excellent process selectivities offers an attractive alternative for an economically viable NADA process and, therefore, for an alternative liquid-phase production of bio-AA from LA.

146

4.3 NADA process shortcuts

After the multi-step NADA process was investigated and largely optimized, ques- tions of possible process shortcuts arised. Based on the improved understanding of each individual reaction step and the related reaction conditions, incompatible reaction step were strictly separated. LA bromination and the required chemical environment, which was namely acidic and HBr-enriched, was determined to be incompatible with the conditions for a selective conversion of bromopropionic acids to AA.

Hence, a shortened NADA reaction sequence would include LA bromination to

2-BrPA and subsequent conversion of 2-BrPA to AA.

On the one hand, direct conversion of 2-BrPA was investigated in basic environment as dehydrobromination of 3-BrPA in TOA showed excellent performance at mild reaction conditions (compare chapter 4.2.3.1). In addition, assuming that 3-BrPA is exclusively formed by anti-Markovnikov type HBr addition to AA, high yields of AA would be accessible if hydrobromination of AA could be suppressed. Especially, isomerization experiments of 2-BrPA to 3-BrPA at high dilutions (IL:reactant molar ratio ≥ 9:1) and elevated temperatures (Tr ≥ 160 °C) previously showed notable amounts of AA. Therefore, in another set of experiments, 2-BrPA dehydrobromination to AA was examined in [PBu4]Br.

4.3.1 Dehydrobromination of 2-BrPA to AA in basic environment

For a potential direct dehydrobromination of 2-BrPA, base-assisted HBr elimination was chosen as a first approach. The behavior of 2-BrPA was characterized in basic environment, as 3-BrPA dehydrobromination was successfully demonstrated under comparable conditions.

The base-assisted dehydrobromination experiment was monitored over 240 min at 80 °C at a 2-BrPA:TOA ratio of 1:1. The time-resolved course of the reaction is shown in Figure 33. For quantification with 1H NMR, C-H molecule fragments were re- garded in the chemical shift region of 3.7-5.5 ppm. The course of 2-BrPA dehydrobromination in TOA monitored by 1H NMR spectroscopy is shown in Figure 34. 4 Results and Discussion 147

To account for the total moles of C3-building blocks, all oligomer-like 1H NMR signals were included for quantification.

2-BrPA 40 linear oligomers lactide

20 mmol

0 50 100 150 200 250 Reaction time / min

Figure 33: Time-resolved profile of base-assisted dehydrobromination of 2-BrPA. Conditions: 50 mmol 2-BrPA, 50 mmol TOA, 240 min, 80 °C, ambient pressure. Lines are only given as a guide to the eye.

First and foremost, no AA is observed over the course of 2-BrPA conversion in TOA at 80 °C. 2-BrPA is converted to approximately 85 % within the first hour of the batchwise performed dehydrobromination experiment. After 2 hours, full conversion of

2-BrPA is observed. The reactivity of 2-BrPA in the basic reaction medium is characterized by intermolecular esterification, resulting in the formation of mainly two products, namely lactide and linear oligomers. Over the first two hours of the experiment, the amounts of the cyclic dimer of LA (lactide) and linear oligomers increase and finally reach a plateau of approximately 8.5 mmoles (lactide) and 30 mmoles (oligomers) after approximately 2 hours. The selective cyclization seems to be limited compared to linear self-esterification, resulting in a lactide:linear oligomers ratio of

0.28. 148

MAK-TOA-1-0_1H-1.jdf 0.11 0.10 substrate 0.09 1

0.08

0.07

0.06

0.05 0 min

Normalized Intensity Normalized 0.04 1 0.03

0.02

0.01

0.045 MAK-TOA-1-05h_1H-1.jdf

5.4 5.3 5.2 5.1 5.0 4.9 4.8 4.7 4.6 4.5 4.4 4.3 4.2 4.1 4.0 3.9 3.8 0.040 Chemical Shift (ppm)

0.035 2

0.030 0.025 main products 0.020 30 min Normalized Intensity Normalized 0.015

0.010

0.005

0.045 MAK-TOA-1-1h_1H-1.jdf

0.040 5.4 5.3 5.2 5.1 5.0 4.9 4.8 4.7 4.6 4.5 4.4 4.3 4.2 4.1 4.0 3.9 3.8 Chemical Shift (ppm) 2 0.035 2 0.030

0.025 0.020 60 min

Normalized Intensity Normalized 0.015

0.010

0.005 0.045 MAK-TOA-1-4h_1H-1.jdf

0.040 5.4 5.3 5.2 5.1 5.0 4.9 4.8 4.7 4.6 4.5 4.4 4.3 4.2 4.1 4.0 3.9 3.8 0.035 Chemical Shift (ppm) 0.030 3 5 0.025 4 0.020 4 240 min

Normalized Intensity Normalized 0.015 0.010 3 5 0.005

5.4 5.3 5.2 5.1 5.0 4.9 4.8 4.7 4.6 4.5 4.4 4.3 4.2 4.1 4.0 3.9 3.8 Chemical Shift (ppm)

Figure 34: Course of 2-BrPA dehydrobromination in TOA monitored by 1H NMR spectroscopy (left): Partial 1H NMR spectra from start (top; 0 min), over the course (mid; 30 min and 60 min) and until the end of the dehydrobromination experiment (bottom; 240 min), indicating ongoing substrate self-esterification. Respective C-H molecule fragment signals are assigned to the related structure (right).

As described in detail in chapter 4.2.3.1, bio-AA was selectively obtained from 3-BrPA by base-assisted HBr elimination in TOA. 2-BrPA, in contrast to its constitutional isomer 3-BrPA, did not allow for a selective dehydrobromination to AA in TOA. The system was characterized by substrate dimerization and oligomerization. The observed reaction pathways for 2- and 3-BrPA are displayed in Scheme 32.

Selective dehydrobromination of 3-BrPA is attributed to the HBr elimination mechanism. It is assumed that the acidic character of �-protons plays a decisive role in this context. In contrast, HBr elimination from 2-BrPA is prevented, which may be due to the aliphatic character of �-protons. Here, exclusively intermolecular substrate esterification was observed, initiated by a nucleophilic attack of the 2-bromopropanoate, which is the preferably available species in alkaline environment. The formed dimer

(2-BrPA)2 can either undergo linear self-esterification, resulting in the formation of linear oligomers of 2-BrPA building blocks of different chain lengths. Moreover, the 4 Results and Discussion 149

cyclic dimer lactide was observed in 1H NMR, emerging from selective cyclization of

(2-BrPA)2 in the nearly anhydrous reaction medium.

Scheme 32: Comparison of observed 2- and 3-BrPA behavior in basic environment (TOA).

A different reactivity of bromopropionic acids in basic and aqueous medium has already been documented in 1905 by Kowski. The results describe an enhanced decomposition and increased formation of AA from the �-modification of the bromopropionic acid under similar conditions.[11d]

4.3.2 Dehydrobromination of 2-BrPA to AA in bromide ILs

Interestingly, base-assisted dehydrobromination of 2-BrPA did not yield AA but exclusively esterification products. To still enable a shortened reaction sequence of the multi-step NADA process, direct conversion of 2-BrPA to AA was investigated in bromide ILs. Bearing in mind the results of 2-BrPA isomerization to 3-BrPA (see chapter 4.2.2), first indications of AA formation in [PBu4]Br were observed when high dilutions of 2-BrPA were used in the isomerization screening. 150

4.3.2.1 Batchwise dehydrobromination of 2-BrPA to AA in bromide ILs

Therefore, the second approach for direct AA production from 2-BrPA was to investigate 2-BrPA conversion in an excess of [PBu4]Br at typical isomerization temperatures of 160-180 °C. The following experiment was conducted at a reaction temperature of 180 °C with an IL:2-BrPA molar ratio of 9:1. The experimental results are compared to the conversion of 2-BrPA at an IL:2-BrPA molar ratio of 1:1 in Figure 35.

Comparative experiments at a reaction temperature of 160 °C and an illustration of the temperature-dependent conversion of 2-BrPA at a [PBu4]Br:2-BrPA molar ratio of

1:1 (3 h) are given in the Appendix (Figure A 23 and Figure A 24).

2-BrPA 2-BrPA 3-BrPA 3-BrPA 100 AA 12 AA LA LA

C3 balance C3-balance 10 80

8 60

Amountmmol / Amountmmol /

20 2

0 0 0 5 10 15 20 25 0 5 10 15 20 25 Reaction time / h Reaction time / h

Figure 35: Time-resolved conversion of 2-BrPA in [PBu4]Br. Conditions: IL:2-BrPA molar ratio of 1:1 (left) or 9:1 (right), 100 mmol 2-BrPA (left) or 11 mmol (right), 100 mmol [PBu4]Br, 24 h, 180 °C, ambient pressure.

When an IL:2-BrPA molar ratio of 9:1 is used at 180 °C (see Figure 35, right), a different product ratio is observed over the course of 2-BrPA conversion. Instead of exclusively forming 3-BrPA (IL:2-BrPA molar ratio of 9:1, left), substrate dilution leads to preferred formation of AA in [PBu4]Br. Fast 2-BrPA conversion is observed exceeding

90 % substrate conversion within one hour. Simultaneously, AA and 3-BrPA are formed in high rates. A maximum YAA of approximately 48 % is reached after 45 min and the highest Y3-BrPA of 25 % was determined after one hour. Additionally, regarding the molar balance of the experiment, severe C3-decomposition is observable, resulting from degradation of AA and 3-BrPA over the course of the experiment. After 8 hours, over 50 % of the used substrate quantity is lost and after one day, almost 90 % C3-compounds are decomposed. Interestingly, decomposition is already observed in the first 4 Results and Discussion 151

hour of the reaction, which means that even during AA and 3-BrPA formation, decomposition is present. The decomposition of AA and 3-BrPA in [PBu4]Br at elevated temperatures is attributed to decarbonylation to CO and AcH as well as decarboxylation to CO2 and ethylene.

Again, the IL ([PBu4]Br) activates the C-Br bond and therefore enables AA and 3-BrPA formation in the above shown experiment. Assuming that 3-BrPA is exclusively formed from AA, a diluted reaction system decreases the relative share of 3-BrPA formation by suppression of HBr addition to the vinyl functionality of AA. AA decomposition is predominant under respective reaction conditions, which can be seen by a characteristic sharp bend in the course of the C3-balance at 1.5 hours (see Figure

35, left). As long as noticeable amounts of 3-BrPA are formed, the available AA tends to fast decomposition via the above addressed mechanisms. The formed 3-BrPA is more stable in the reaction medium, leading to a significantly decelerated decomposition.

However, the results indicate hampered 3-BrPA formation in diluted reaction medium, resulting in enhanced AA formation. The formed AA was characterized to be instable under reaction conditions. Moreover, it should be noted that the diluted reaction system is accompanied by a drastically reduced productivity / STY and therefore less economical.

4.3.2.2 Semi-batchwise dehydrobromination of 2-BrPA to AA in bromide ILs

To overcome the above described barriers, the final approach for direct dehydrobromination of 2-BrPA to AA was a semi-batchwise process operation at equimolar

IL:2-BrPA molar ratio. In contrast to previous and batchwise operated dehydrobromination experiments, AA and HBr were isolated from the hot reaction mixture. On one hand, reduced C3-decomposition was assumed due to shorten product residence time. On the other hand, fast separation of AA and HBr was estimated to influence the AA:3-BrPA product ratio, assuming that 3-BrPA is exclusively produced by HBr addition to the vinyl functionality of AA. 152

For the semi-batchwise process operation, an adapted experimental setup was used to enable in-situ product removal by either reduced pressure or a flow of argon strip gas and downstream processing of liquid (AA and 3-BrPA) and gaseous (HBr) products. Moreover, separated product trapping was assumed to be essential to achieve high isolated AA amounts. HBr absorption in a downstream washing flask was proven by continuous pH monitoring. A detailed description of the experimental procedure and setup is given in the Experimental section.

Different parameters and setup configurations were screened to investigate the influence of in-situ HBr and product removal on the reactivity and product selectivity of

2-BrPA conversion in [PBu4]Br. An overview of the screened configurations and the corresponding results are summarized in the Appendix. Two exemplary semi-batchwise conducted experiments were compared with a standard batch experiment at a reaction temperature of 180 °C. As clearly visible from

Figure 36, the product share of 2-BrPA dehydrobromination in [PBu4]Br at 180 °C can strongly be influenced by different experimental operation modes. It should be further noted that the reported yields of the semi-batchwise experiments refer to an isolated product amount, compared to the standard experiment (non-isolated product yields).

The results shown in Figure 36 report the semi-batchwise operation by either use of reduced pressure or strip gas are compared to a batch experiment of 2-BrPA conversion in [PBu4]Br at 180 °C. Characterizing quantities of the benchmark experiment are

85 % X2-BrPA, 6 % YAA and 51 % Y3-BrPA after 3 h. The share of hydrobromination of AA is strikingly reduced by in-situ product removal. HBr addition to the vinyl functionality of AA is decreased, leading to increased isolated amounts of AA. In the shown semi-batchwise example with product removal at a reduced pressure of 25 mbar, YAA was 41 % compared to Y3-BrPA of 25 %. 47 mmol of HBr are recovered by HBr trapping and subsequent absorption in a pH monitored scrubber unit. Including the detected amount of 3-BrPA, this corresponds to 73 % of the used HBr. When a flow of argon strip gas of 500 mLn/min was used, a YAA of 25 % could be achieved, compared to a

Y3-BrPA of 15 %. Here, 68 mmol of HBr are absorbed in the downstream washing flask, 4 Results and Discussion 153

resulting in 83 % (15 mmol 3-BrPA) balanced HBr of the initial HBr amount. As already mentioned, SAA and S3-BrPA are strongly influenced by the operation mode of the dehydrobromination of 2-BrPA. Quite the contrary is true for the combined product selectivity (SAA+S3-BrPA), which is approximately 68 % for both dehydrobrominations at ambient pressure and reduced pressure, respectively. This indicates that substrate decomposition via decarbonylation and decarboxylation are still highly competing reaction pathways. The determined combined product selectivity for the dehydrobromination of 2-BrPA with a flow of argon strip gas of 500 mLn/min was approximately

45 %. The observed drop in selectivity is most likely due to insufficient trapping of liquid products and therefore based on the experimental setup and not due to reduced reaction selectivity. It should be mentioned that dehydrobromination of

2-BrPA at temperatures ≥ 180 °C and reduced pressures showed visible deposition of white material on the glass surface of the experimental setup. Under these conditions, AA polymerization occurs in the concentrated product stream, which has been verified by NMR spectroscopy.

X2-BrPA

YAA

50 Y3-BrPA 100

40 80 / % /

30 60

3-BrPA

2-BrPA X

& Y & 20 40

AA Y

10 20

0 0 batch semi-batch / vacuum semi-batch / stripgas

operation mode

Figure 36: Conversion of 2-BrPA in [PBu4]Br at different experimental operation modes. Conditions: IL:2-

BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol [PBu4]Br, 3 h, 180 °C, ambient pressure (batch), 25 mbar (vacuum) and 500 mLn/min flow of Ar (strip gas). 154

To clarify the vast influence of the process operation mode on the dehydrobromination of 2-BrPA in [PBu4]Br, the dehydrobromination product distribution within the screened set of parameters is summarized in Table 9.

Table 9: Product distribution of 2-BrPA conversion in [PBu4]Br at different process operations modes .

Temperature / °C Product distribution[a] batch & ambient semi-batch & vac- semi-batch & strip

pressure uum[b] gas[c]

160 0.14 1.64 0.82

180 0.12 3.67 1.67

200 0.13 n.a. 1.93

[a] Conditions: IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol [PBu4]Br, 3 h. Product distribution is defined as n(AA) [mmol] / n(3-BrPA] [mmol]. [b] Pressure of experiment was 5 . [c] (160 °C) and 25 mbar (180 °C). Flow of argon strip gas was 500 mLn/min.

Batchwise dehydrobromination of 2-BrPA in a temperature range of 160-180 °C results in product distributions of 0.12-0.14, reflecting the high share of AA hydrobromination in [PBu4]Br at elevated temperatures. In contrast, the use of strip gas increased the dehydrobromination product distribution to 0.82 (160 °C), 1.67 (180 °C) or even 1.93 (200 °C). This means that the formed AA is not only isolated from the reaction mixture but also, for instance, increased by a factor of approximately 4 at a reaction temperature of 200 °C (27 mmoles vs. 7 mmoles). A similar trend was observed at reduced pressure of 5 mbar (160 °C). The highest product distribution of 3.67 was obtained when the dehydrobromination was performed at reduced pressure of 25 mbar and 180 °C. The isolated amount of AA (41 mmoles) is increased by a factor of approximately 6, compared to non-isolated amount (7 mmoles) of the batchwise dehydrobromination experiment. 4 Results and Discussion 155

4.3.3 Summary of process shortcuts

To sum up, neither base-assisted dehydrobromination of 2-BrPA in TOA nor batchwise 2-BrPA dehydrobromination in [PBu4]Br were suitable shortcuts for a selective bio-AA production. In the first concept, esterification species were determined as main products of the reaction. Batchwise 2-BrPA conversion in [PBu4]Br was characterized by either 3-BrPA formation or massive C3-decomposition. In contrast, the presented results clearly demonstrated the opportunity of direct dehydrobromination of

2-BrPA by shifting the dehydrobromination product distribution towards AA at reduced pressure or a flow of argon strip gas. Simultaneously, 3-BrPA formation was significantly reduced under HBr-removing reaction conditions. Moreover, this observation led to the conclusion that 3-BrPA is exclusively formed by hydrobromination of the vinyl functionality of AA and therefore proved the postulated NADA mechanism. The use of strip gas simplified the process operation by enabling continuous

HBr absorption in a downstream scrubber unit, which was continuously monitored with a pH meter. Furthermore, an argon-diluted product gas stream minimizes AA polymerization in downstream processing.

A further advantage of the developed two-step NADA concept compared to the

3-step process is the opportunity of HBr trapping in aqueous solution, which can theoretically be reused in the first step of the reaction sequence, namely LA bromination.

No additional HBr recovery process step is needed to close the HBr cycle within the process (see 3-step NADA process, chapter 4.2.3 and 4.2.4).

This chapter can be seen as a proof of concept for 2-BrPA dehydrobromination to AA and HBr. However, this study also revealed several challenges for selective bio-AA production from 2-BrPA: First and foremost, an optimization of the experimental setup and reaction parameters is necessary to improve product isolation. Key issue is an optimized separated trapping of AA and HBr to reduce 3-BrPA production within downstream processing. Inhibition of AA polymerization in the high concentrated fluids and safe and efficient HBr recovery plays also an important role for the feasibility of the process. Finally, it should be noted that the experiments in this series of dehydrobromination experiments were performed in [PBu4]Br, due to excellent ther- 156 mal stability, for instance. Though, isomerization selectivity was proven to be significantly improved in e.g. [EMIM]Br (up to 90 %). Reduced AA and 3-BrPA decomposition may therefore be enabled in imidazolium-based ILs, especially at lower reaction temperatures. [EMIM]Br and ILs of similar structure hold great promise for a future selective 2-BrPA dehydrobromination to AA.

5 Summary 157

5 Summary

The aim of this thesis was the development of a multi-step liquid-phase process for the dehydration of LA to bio-AA. The work is based on a novel technology developed at the Institute of Chemical Reaction Engineering (CRT) of the Friedrich-Alexander-Uni- versity Erlangen-Nürnberg (FAU), namely “Nucleophile Assisted Dehydration to

Acrylates” (NADA). This innovative technology was developed in close collaboration with the industrial partner P&G from 2014 to 2017 and enables a liquid-phase dehydration of LA and its derivatives towards bio-AA. The NADA technology can be summarized as a HBr-triggered dehydration of LA (or derivatives thereof like lactide, lactates, lactic oligomers, 2-APA, etc.) at reaction temperatures of 120-220 °C in bromide

ILs (chapter 4.1.1).

In the beginning, a benchmarking of LA dehydration in a one-step NADA process was performed (chapter 4.1.2). The dehydration was conducted at typical process conditions of the one-step NADA process (150 °C, LA:[PBu4]Br molar ratio of 1:2,

5 mmol of 2-BrPA as acidic additive, 168 h). The obtained YAA was 30 % and selectivity towards the desired product (SAA) was limited to < 40 %. However, 30 % AA at 150 °C and ambient pressure was the highest YAA ever reported in a liquid-phase process based on commercial and aqueous LA as feedstock (PURAC®) at this time.

Based on the observed limitations of direct LA dehydration using the NADA technology and the postulated reaction mechanism, a multi-step NADA process was conceptualized: The designed process can be described as a spatially separated process operation of three individual reaction steps:

1) Bromination of LA to 2-BrPA

2) Isomerization of 2- to 3-BrPA

3) Dehydrobromination of 3-BrPA to bio-AA

158

The decoupling of acid- and base-catalyzed reaction steps was assumed to allow for increased degrees of freedom for optimized conditions of each individual reaction and the corresponding environment. Thereupon, a multi-step NADA process operation, including the above mentioned reaction steps, should be developed and optimized (chapter 4.1.3).

The first and most complex reaction step, namely bromination of the sensitive feedstock LA to 2-BrPA (chapter 4.2.1), was investigated in an aqueous reaction medium using HBr(aq) as bromination agent. In the aqueous reaction system, acid-catalyzed substrate decarbonylation and/or selfesterification / polymerization predominated.

Under optimized reaction conditions (5 h, 120 °C, LA:HBr molar ratio of 1:3), bromination selectivity was limited to approx. 50 % and the highest obtained Y2-BrPA was 25 %.

Subsequently, two concepts were investigated to enhance the performance of the aqueous bromination medium. In-situ LLE of 2-BrPA did not result in an enhancement of LA bromination with HBr(aq). Nevertheless, ex-situ LLE of 2-BrPA from the aqueous reaction solution was successfully demonstrated, which could be used for product isolation at a later stage. The use of [PBu4]Br as ionic liquid additive, which was assumed to enhance bromination yield by shifting the reaction equilibrium towards the product site (increased bromide concentration), resulted in a slightly increased Y2-BrPA (32 %) and S2-BrPA (63 %), compared to the aqueous benchmark.

To overcome the limitations of LA bromination with HBr(aq), zwitterions were investigated as HBr carrier. In this concept, the zwitterionic carrier molecule was loaded with HBr by protonation of the sulfonyl group, forming an acidic bromide IL. When the zwitterionic-based molten salt is loaded with HBr, the formed ionic liquid has been shown to act as a selective and efficient brominating agent and reaction matrix for the bromination of LA to 2-BrPA. The IL-based bromination system provided several benefits compared to HBr(aq): Reduced H2O-content of the bromination agent, low vapor pressure of HBr and increased nucleophilic character of the bromide anion.

Moreover, the zwitterionic HBr carrier combined the role of solvent, proton source, buffer agent, and brominating agent within one substance. Especially, imidazole- based ILs showed exceptional selectivity in the synthesis of 2-BrPA from lactide and

LA. The best results were obtained with [MIMBS]Br and [EIMBS]Br reaching 60% 5 Summary 159

yield to 2-BrPA and > 99% selectivity when lactide was used as substrate. This finding was in sharp contrast to the same reaction performed in HBr(aq) where decarbonylation and self-oligomerization dominated and a maximum yield of only 20% could be achieved under otherwise comparable reaction conditions (120°C, 5 h, and ambient pressure). Moreover, product and substrate isolation, as well as brominating agent recycling, have successfully been demonstrated. This novel method for the synthesis of 2-BrPA using SO3H-functionalized bromide ILs was also investigated with PU-

RAC® (88 wt% LA solution in water) as substrate for the reaction. The reduced Y2-BrPA

(42 %) was attributed to the strongly increased H2O-content of the reaction system. However, comparably high selectivity towards 2-BrPA (95 %) and reduced feedstock costs make PURAC® a viable alternative as feedstock for renewable production of

2-BrPA en route to bio-AA. Even the aim was to develop a renewable and environmental-friendly synthesis of 2-BrPA starting from biogenic LA, we additionally gained fundamental and detailed insights into technical potential of zwitterionic-type molten salts as HBr carrier for brominating high-functional and sensitive substrates.

This method is expected to be suitable for a selective bromination of various OH- containing, complex substrates.

An emerging isomerization of 2- to 3-BrPA, which is the second step of the designed multi-step NADA process (chapter 4.2.2), was observed in [PBu4]Br at elevated temperatures. It should be mentioned here that the IL [PBu4]Br is the state-of-the-art reaction matrix for the one-step NADA process and was further used to benchmark direct LA dehydration. The essential role of the bromide anion of the used IL was shown, initiating the isomerization reaction. The observations indicated a consecutive reaction sequence via dehydrobromination of 2-BrPA to AA and subsequent addition of HBr to AA, forming 3-BrPA. Highest isomerization grades were achieved with an equimolar ratio of IL:2-BrPA. Under optimized conditions, maximum S3-BrPA was approximately 70 % in [PBu4]Br, limited by competing substrate decomposition and/or oligomerization. A cation variation of the IL further enhanced the isomerization performance. The best isomerization matrix was [EMIM]Br giving Y3-BrPA of 70 % in 90 % selectivity (IL:2-BrPA molar ratio of 1:1, 140 °C, 20 h). Product isolation and purification of the IL was accessible via vacuum distillation. 160

The third and last reaction of the multi-step NADA process, namely 3-BrPA dehydrobromination to AA, was studied in TOA (chapter 4.2.3). The amine served as base, enabling HBr elimination from 3-BrPA and simultaneously “catched” the formed

HBr, giving a salt ([TOAH]Br). This procedure allowed for safe and efficient HBr storage. Even at mild conditions (80 °C, ambient pressure), high rate and selective dehydrobromination of 3-BrPA was achieved. Within the first 15 minutes of the experiment, YAA reached 84 % with selectivity towards AA of 91 %. The only by-product observed was the dimerization product DiAA, resulting from a consecutive intermolecular reaction of two AA monomers. Combined selectivity (SAA + SDIAA) was > 99 %.

However, HBr recovery from the formed salt remained the bottleneck of the current multi-step NADA process. The maximum HBr recovery achieved within this study was below 10 %.

All in all, spatial separation of the individual reaction steps and adjustment of the related chemical environment and reaction conditions, respectively, significantly improved the dehydration of LA with the NADA concept. The overall product selectivity (AA), which is a decisive factor for a novel and profitable technology, was increased from < 40 % to 90.0 % (85.5 %) for lactide (LA) as feedstock. By means of multi-step processing, it was possible to adapt the NADA environment to the actually required conditions of each reaction step and therefore enable selective dehydration starting from LA. Thus, the basic aim of this thesis to develop a multi-step and liquid-phase dehydration process for the production of bio-AA from LA was achieved

(chapter 4.2.4).

Based on the improved knowledge of each individual reaction step of the NADA mechanism, possible process shortcuts were finally examined. In contrast to 3-BrPA, its constitutional isomer 2-BrPA did not allow for direct dehydrobromination to AA in TOA via base-assisted HBr elimination. The reaction system was characterized by substrate dimerization and oligomerization (chapter 4.3.1).

In a further approach (chapter 4.3.2), the behavior of 2-BrPA was examined in excess of [PBu4]Br at typical isomerization temperatures of 160-180 °C (IL:2-BrPA molar ratio of 9:1). By using an excess of bromide IL, the share of AA formation was increased. 5 Summary 161

YAA of approximately 48 % was reached within 45 minutes. Again, severe C3-decomposition was observed, resulting in a loss of AA and 3-BrPA over the course of the reaction. However, the results indicated a hampered 3-BrPA formation in a diluted reaction system, resulting in an increased AA formation (minimized HBr addition to

AA vinyl group).

On the basis of these findings, a final approach for direct dehydrobromination of

2-BrPA to AA was a semi-batchwise operated process at equimolar IL:2-BrPA molar ratio (chapter 4.3.2). Here, AA and HBr were isolated from the hot reaction zone to reduce C3-decomposition and to suppress AA hydrobromination. The semi-batchwise dehydrobromiation (either realized by reduced pressure or a flow of argon strip gas) demonstrated the opportunity of direct AA formation from 2-BrPA in [PBu4]Br at elevated temperatures (160-200 °C). 3-BrPA formation was significantly reduced by HBr removal and the share of formed AA was simultaneously increased. The highest share of AA was achieved at 25 mbar and 180 °C with a product distribution

(nAA/n3-BrPA) of 3.67, which corresponded to a sixfold increase compared to the batchwise process operation at similar conditions. In addition to a shortened reaction sequence, such a two-step process offers a further advantage: The eliminated HBr could directly be removed and trapped as aqueous solution, which can be reused in the first step of the reaction sequence (bromination of LA). In such a case, no additional HBr recovery step is needed to close the HBr cycle within the NADA process. Further optimization of the experimental setup, procedure and reaction conditions is needed to enable an efficient bio-AA production from 2-BrPA. However, within the scope of this thesis and the presented work, a proof of concept for 2-BrPA dehydrobromination to bio-AA and HBr was achieved and thus, the foundation of a two-step NADA process was laid.

The knowledge gained during this thesis allows a deeper chemical understanding of bio-AA production from LA with the NADA concept and in bromide ILs, proceeding via brominated species. Even though the young NADA technology needs to be further optimized, it can already be considered as a real breakthrough technology towards a future economic production of bio-AA in liquid phase processes.

162

6 Bibliography

[1] Procter & Gamble, https://us.pg.com/sustainability/at-a-glance/our-goals, accessed 01-05-2018. [2] a) R. Beerthuis, G. Rothenberg, N. R. Shiju, Green Chemistry 2015, 17, 1341- 1361; b) R. A. Sheldon, Green Chemistry 2014, 16, 950-963. [3] a) R. K. Grasselli, F. Trifirò, Rendiconti Lincei 2017, 28, 59-67; b) P. A. Holman, D. R. Shonnard, J. H. Holles, Industrial & Engineering Chemistry Research 2009, 48, 6668-6674; c) A. Talebian-Kiakalaieh, N. A. S. Amin, H. Hezaveh, Renewable and Sustainable Energy Reviews 2014, 40, 28-59; d) X. Xu, J. Lin, P. Cen, Chinese Journal of Chemical Engineering 2006, 14, 419-427. [4] a) Acrylic Acid - A Summary of Safety and Handling, 4th ed., Basic Acrylic Monomer Manufacturers, Inc. (BAMM), 2013; b) A. Behr, D. W. Agar, J. Jörissen, A. J. Vorholt, Einführung in die Technische Chemie, Springer Berlin Heidelberg, 2017; c) M. M. Lin, Applied Catalysis A: General 2001, 207, 1-16; d) K. Weissermel, H.-J. Arpe, Industrial Organic Chemistry, 4th ed., Wiley-VCH, Weinheim, 2003; e) H. A. Wittcoff, B. G. Reuben, J. S. Plotkin, Industrial Organic Chemicals, 2nd ed., Wiley-VCH, Weinheim, 2004; f) B. Yan, L.-Z. Tao, Y. Liang, B.-Q. Xu, ACS Catalysis 2014, 4, 1931-1943. [5] a) The Roadmap for Biomass Technologies in the U.S., US Department of Energy, 2002; b) R. Datta, M. Henry, Journal of Chemical Technology & Biotechnology 2006, 81, 1119-1129; c) M. Dusselier, P. Van Wouwe, A. Dewaele, E. Makshina, B. F. Sels, Energy & Environmental Science 2013, 6, 1415-1442; d) P. Mäki-Arvela, I. L. Simakova, T. Salmi, D. Y. Murzin, Chemical Reviews 2014, 114, 1909-1971. [6] R. E. Holmen, US2859240, 1958. [7] a) J. E. Godlewski, J. Villalobos, D. I. Collias, J. E. Velasquez, US2013/0274513, 2013; b) J. E. Velasquez, D. I. Collias, J. E. Godlewski, J. Villalobos Lingoes, US2015/0105584, 2015; c) J. E. Velasquez, D. I. Collias, J. E. Godlewski, F. C. Wireko, US2017/0057901, 2017. [8] C. T. Lira, P. J. McCrackin, Industrial & Engineering Chemistry Research 1993, 32, 2608-2613. [9] a) T. M. Aida, A. Ikarashi, Y. Saito, M. Watanabe, R. L. Smith Jr, K. Arai, The Journal of Supercritical Fluids 2009, 50, 257-264; b) W. S. L. Mok, M. J. Antal, M. Jones, The Journal of Organic Chemistry 1989, 54, 4596-4602. [10] a) F.-F. Kuppinger, F. Klasovsky, A. May, M.-Z. Oh, WO2014/111363, 2014; b) F. G. Terrade, J. van Krieken, B. J. V. Verkuijl, E. Bouwman, ChemSusChem 2017, 10, 1904-1908; c) F. G. Terrade, B. Verkuijl, E. Bouwman, EP15201469, 2015. [11] a) C. Hell, Berichte der deutschen chemischen Gesellschaft 1881, 14, 891-893; b) J. Volhard, Justus Liebigs Annalen der Chemie 1887, 242, 141-163; c) N. Zelinsky, Berichte der deutschen chemischen Gesellschaft 1887, 20, 2026-2026; d) E. Kowski, Justus Liebigs Annalen der Chemie 1905, 342, 124-138; e) E. C. Kendall, B. McKenzie, Org. Synth. 1923, 3, 25; f) E. C. Kendall, B. McKenzie, Org. Synth. 1923, 3, 57; g) F. K. Beilstein, B. Prager, P. H. Jacobson, P. Schmidt, D. Stern, 6 Bibliography 163

F. Richter, Handbuch der organischen Chemie: Die Literatur bis 1. Januar 1910 umfassend, Vol. 4, Springer, 1944. [12] a) J. P. Hallett, T. Welton, Chemical reviews 2011, 111, 3508-3576; b) T. Welton, Chemical Reviews 1999, 99, 2071-2084; c) P. Wasserscheid, W. Keim, Angewandte Chemie International Edition 2000, 39, 3772-3789. [13] a) S. Bell, D. Dave, PEP Report 6F Acrylic Acid, IHS Markit Chemical, 2015; b) Global Market Insights, https://www.gminsights.com/industry-analysis/bio- acrylic-acid-market, , accessed 11-28-2017; c) IHS Chemicals, https://www.ihs.com/products/chemical-technology-pep-acrylic-acid- 2015.html, , accessed 29-09-2017; d) Markets and Markets, https://www.marketsandmarkets.com/Market-Reports/acrylic-acid-market- 683.html?gclid=CjwKCAjwos7NBRAWEiwAypNCey3UZkzT_J1HMoWB1Q BVPZksdSaK9d64BxmbSZ_peGI-RvLGWlGYKBoCGVwQAvD_BwE, accessed 11-28-2017. [14] a) BASF Spec Sheet Acrylic Acid glacial, BASF, Ludwigshafen, 2011; b) F. Ho, O. S. Fruchey, R. L. Roundy, W.G. Etzkorn, C. T. Reeves, M. Jawaid, E. J. Yang, D. W. Jewell, US8530700, 2013; c) T. Blum, P. Zurowski, T. J. Korn, S. Haremza, T.n Friese, F.-G. Martin, U. Jäger, S. Rissel, V. Schliephake, US9096513 B2, 2015. [15] a) N. Shimizu, I. Yanagisawa, M. Takata, T. Sato, US4147885, 1979; b) M. Tanimoto, K. Uekawa, T. Kawajiri, US6069271, 2000. [16] a) O. Machhammer, T. Dockner, G.-J. Engert, T. Proll, US5817865, 1998; b) J. Schröder, G. Nestler, K. J. Müller-Engel, US6498272, 2002. [17] a) S. Umemura, K. Ohdan, H. Asada, US4267385, 1981; b) S. Umemura, K. Ohdan, K. Suzuki, F. Adachi, US4045478, 1977. [18] a) A. Diefenbacher, U. Hammon, V. Schliephake, G. Sieder, US7566804, 2009; b) A. Hengstermann, A new approach to industrial melt crystallization of acrylic acid, Shaker, Aachen 2010 (Delft University of Technology, Diss., 2010); c) S. Maezawa, H. Yoshikawa, M. Hashimoto, K. Sakamoto, J. Fujii, US3725208, 1973; d) H. Panda, Handbook On Chemical Industries (Alcohol Based), Asia Pacific Business Press Incorporated, 2002; e) K. Wada, Y. Seo, A. Iwamoto, A. Sudo, F. Sakai, K. Shiraishi, H. Nowatari, US6028220A, 2000; f) D. R. Wagner, S. J. Miko, US6737546, 2004. [19] Light Olefin Solutions for Propylene and Ethylene Production, UOP LLC., Des Plaines, 2011. [20] S. Andre, T. Hundertmark, R. Rao, Using microeconomics to guide investments in petrochemicals, McKinsey & Company, Houston, 2012. [21] Direct biological production of Propylene, Global Bioenergies, Evry, 2014. [22] a) Global Bioenergies, http://www.global-bioenergies.com/erster-erfolg-bei- der-masstabsgetreuen-umsetzung-des-isobutenverfahrens-in-der- demonstrationsanlage-in-leuna/?lang=de, accessed 12-05-2017; b) Global Bioenergies, http://www.global-bioenergies.com/erste-flaschenabfullung-in- der-demonstrationsanlage-leuna/?lang=de, accessed 12-05-2017. [23] M. Iwamoto, S. Mizuno, M. Tanaka, Chemistry – A European Journal 2013, 19, 7214-7220. [24] B. Katryniok, S. Paul, M. Capron, F. Dumeignil, ChemSusChem 2009, 2, 719- 730. 164

[25] A. S. de Oliveira, S. J. S. Vasconcelos, J. R. de Sousa, F. F. de Sousa, J. M. Filho, A. C. Oliveira, Chemical Engineering Journal 2011, 168, 765-774. [26] L. Shen, H. Yin, A. Wang, Y. Feng, Y. Shen, Z. Wu, T. Jiang, Chemical Engineering Journal 2012, 180, 277-283. [27] N. R. Shiju, D. R. Brown, K. Wilson, G. Rothenberg, Topics in Catalysis 2010, 53, 1217-1223. [28] T. Toraya, T. Tobimatsu, M. Yamanishi, K. Hideki, USP2008/0131945, 2008. [29] S. Vollenweider, C. Lacroix, Applied Microbiology and Biotechnology 2004, 64, 16-27. [30] P. Slininger, R. Bothast, Applied and environmental microbiology 1985, 50, 1444- 1450. [31] A. Chieregato, F. Basile, P. Concepción, S. Guidetti, G. Liosi, M. D. Soriano, C. Trevisanut, F. Cavani, J. M. L. Nieto, Catalysis Today 2012, 197, 58-65. [32] A. Witsuthammakul, T. Sooknoi, Applied Catalysis A: General 2012, 413–414, 109-116. [33] J.-L. Dubois, US8143454, 2012. [34] R. M. Contractor, M. W. Andersen, D. Campos, G. Hecquet, C. Pham, M. Simon, M. Stojanovic, R. Kotwica, J.-P. Schirmann, US6310240, 2001. [35] X. Li, Y. Zhang, ACS Catalysis 2016, 6, 143-150. [36] a) P. Korovchenko, N. Shiju, A. Dozier, U. Graham, M. Guerrero-Pérez, V. Guliants, Topics in Catalysis 2008, 50, 43-51; b) N. R. Shiju, V. V. Guliants, S. H. Overbury, A. J. Rondinone, ChemSusChem 2008, 1, 519-523; c) N. R. Shiju, X. Liang, A. W. Weimer, C. Liang, S. Dai, V. V. Guliants, Journal of the American Chemical Society 2008, 130, 5850-5851. [37] T. Abe, Y. Kambara, US5476883, 1995. [38] M. Shen, Y.-G. Zheng, Z.-Q. Liu, Y.-C. Shen, Journal of microbiology and biotechnology 2009, 19, 582-587. [39] A. Kamal, M. S. Kumar, C. G. Kumar, T. B. Shaik, J Microbiol Biotechnol 2011, 21, 37-42. [40] M. O. Guerrero-Pérez, M. A. Bañares, ChemSusChem 2008, 1, 511-513. [41] V. Calvino-Casilda, M. O. Guerrero-Pérez, M. A. Bañares, Green Chemistry 2009, 11, 939-941. [42] J. Le Notre, E. L. Scott, M. C. R. Franssen, J. P. M. Sanders, Green Chemistry 2011, 13, 807-809. [43] H. Belitz, W. Grosch, P. Schieberle, Food Chemistry, Springer, Berlin, 2009. [44] L. Craciun, G. P. Benn, J. Dewing, G. W. Schriver, W. J. Peer, B. Siebenhaar, U. Siegrist, US7538247, 2009. [45] J. Chwae, N. S. Park, M. H. Lee, J. W. Kim, US0171681, 2014. [46] R. O. M. A. de Souza, L. S. M. Miranda, R. Luque, Green Chemistry 2014, 16, 2386-2405. [47] A. J. A. v. Maris, W. N. Konings, J. P. v. Dijken, J. T. Pronk, Metabolic Engineering 2004, 6, 245-255. [48] C. H. Kim, J. W. Seo, L. LUO, US20130095541, 2013. [49] The Dow Chemical Company, https://www.dow.com/en-us/news/press- releases/dow-and-opxbio-collaborating-on-renewable-route-to-acrylic-acid, accessed 12-07-2017. [50] a) BASF SE, https://www.basf.com/en/company/news-and-media/news- releases/2013/07/p-13-356.html, accessed 12-07-2017; b) BASF SE, 6 Bibliography 165

https://www.basf.com/en/company/news-and-media/news- releases/2014/09/p-14-336.html, accessed 07-12-2017. [51] Novozymes, https://www.novozymes.com/en/news/news- archive/2015/01/novozymes-cargill-continue-bio-acrylic-acid-partnership- basf-exits, accessed 07-12-2017. [52] M. L. Lejkowski, R. Lindner, T. Kageyama, G. É. Bódizs, P. N. Plessow, I. B. Müller, A. Schäfer, F. Rominger, P. Hofmann, C. Futter, Chemistry-A European Journal 2012, 18, 14017-14025. [53] N. Huguet, I. Jevtovikj, A. Gordillo, M. L. Lejkowski, R. Lindner, M. Bru, A. Y. Khalimon, F. Rominger, S. A. Schunk, P. Hofmann, M. Limbach, Chemistry – A European Journal 2014, 20, 16858-16862. [54] a) M. Akedo, C. L. Cooney, A. J. Sinskey, Nature Biotechnology 1983, 1, 791- 794; b) G. Schweiger, W. Buckel, FEBS letters 1985, 185, 253-256; c) A. J. Straathof, Chemical reviews 2013, 114, 1871-1908. [55] A. J. J. Straathof, S. Sie, T. T. Franco, L. A. M. van der Wielen, Applied Microbiology and Biotechnology 2005, 67, 727-734. [56] a) Y. Fan, C. Zhou, X. Zhu, Catalysis Reviews 2009, 51, 293-324; b) T. Werpy, G. Petersen, A. Aden, J. Bozell, J. Holladay, J. White, A. Manheim, D. Eliot, L. Lasure, S. Jones, Department of Energy Washington DC, 2004. [57] J. P. Greene, Sustainable Plastics: Environmental Assessments of Biobased, Biodegradable, and Recycled Plastics, Wiley, 2014. [58] a) Allied Market Research, https://www.alliedmarketresearch.com/PLA- polylactic-acid-market, accessed 12-08-2017; b) L. Nattrass, A. Higson, Renewable Chemicals Factsheet: Lactic acid, NNFCC, York, 2011. [59] a) F. A. Castillo Martinez, E. M. Balciunas, J. M. Salgado, J. M. Domínguez González, A. Converti, R. P. d. S. Oliveira, Trends in Food Science & Technology 2013, 30, 70-83; b) Y. Wee, J. Kim, H. Ryu, Food Technology and Biotechnology 2006, 44, 163-172. [60] Y.-J. Wee, J.-N. Kim, H.-W. Ryu, Food Technology and Biotechnology 2006, 44, 163-172. [61] a) R. Auras, Poly (lactic acid), Wiley Online Library, 2010; b) C. H. Fisher, E. M. Filachione, Properties and reactions of lactic acid - a review, United States Department of Agriclture, Philadelphia, 1950. [62] K. Masutani, Y. Kimura, in Poly(lactic acid) Science and Technology: Processing, Properties, Additives and Applications, The Royal Society of Chemistry, 2015, pp. 1-36. [63] P. Van Wouwe, M. Dusselier, E. Vanleeuw, B. Sels, ChemSusChem 2016, 9, 907-921. [64] a) M. Lutz, J. v. Krieken, Acta Crystallographica Section C 2010, 66, 401-405; b) D. T. Vu, A. K. Kolah, N. S. Asthana, L. Peereboom, C. T. Lira, D. J. Miller, Fluid Phase Equilibria 2005, 236, 125-135. [65] a) S. Bhattacharyya, S. Palit, A. Das, Industrial & Engineering Chemistry Product Research and Development 1970, 9, 92-95; b) S. P. Chahal, J. N. Starr, Ullmann's Encyclopedia of Industrial Chemistry 2006. [66] a) R.V. Chaudhari, D. Roy, S. Bala, US0253067, 2012; b) D. Roy, B. Subramaniam, R. V. Chaudhari, Acs Catalysis 2011, 1, 548-551. 166

[67] a) M. A. Abdel-Rahman, Y. Tashiro, K. Sonomoto, Journal of Biotechnology 2011, 156, 286-301; b) M. A. Abdel-Rahman, Y. Tashiro, K. Sonomoto, Biotechnology advances 2013, 31, 877-902. [68] M. Singhvi, D. Gokhale, RSC Advances 2013, 3, 13558-13568. [69] R. P. John, K. M. Nampoothiri, A. Pandey, Applied Microbiology and Biotechnology 2007, 74, 524-534. [70] a) P. R. Gruber, E. S. Hall, J. J. Kolstad, M. L. Iwen, R. D. Benson, R. L. Borchard (Cargill Corp.), WO1995/009879, 1995; b) P. R. Gruber, E. S. Hall, J. J. Kolstad, M. L. Iwen, R. D. Benson, R. L. Borchardt (Cargil Inc.), USP5142023, 1992; c) P. R. Gruber, E. S. Hall, J. J. Kolstad, M. L. Iwen, R. D. Benson, R. L. Borchardt (Cargil Inc.), USP5247059, 1993; d) P. R. Gruber, E. S. Hall, J. J. Kolstad, M. L. Iwen, R. D. Benson, R. L. Borchardt (Cargil Inc.), USP5357035, 1994; e) P. R. Gruber, E. S. Hall, J. J. Kolstad, M. L. Iwen, R. D. Benson, R. L. Borchardt (Cargil Inc.), USP5484881, 1996; f) M. S. Lopes, A. L. Jardini, R. M. Filho, Procedia Engineering 2012, 42, 1402-1413; g) H. Nishida, T. Mori, S. Hoshihara, Y. Fan, Y. Shirai, T. Endo, Polymer degradation and stability 2003, 81, 515-523. [71] a) A. Lucke, A. Göpferich, European journal of pharmaceutics and biopharmaceutics 2003, 55, 27-33; b) W. S. L. Mok, M. J. Antal Jr, M. Jones Jr, The Journal of Organic Chemistry 1989, 54, 4596-4602; c) D. T. Vu, A. K. Kolah, N. S. Asthana, L. Peereboom, C. T. Lira, D. J. Miller, Fluid Phase Equilibria 2005, 236, 125-135. [72] a) E. Blanco, C. Lorentz, P. Delichere, L. Burel, M. Vrinat, J. Millet, S. Loridant, Applied Catalysis B: Environmental 2016, 180, 596-606; b) J.-M. Lee, D.-W. Hwang, Y. K. Hwang, S. B. Halligudi, J.-S. Chang, Y.-H. Han, Catalysis Communications 2010, 11, 1176-1180; c) B. M. Murphy, M. P. Letterio, B. Xu, Journal of Catalysis 2016, 339, 21-30; d) M. S. Tam, J. E. Jackson, D. J. Miller, Industrial & Engineering Chemistry Research 1999, 38, 3873-3877; e) B. Wang, C. li, Q. Zhu, T. Tan, RSC Advances 2014, 4, 45679-45686; f) Z. Zhang, Y. Qu, S. Wang, J. Wang, Journal of Molecular Catalysis A: Chemical 2010, 323, 91-100. [73] R. Beerthuis, M. Granollers, D. R. Brown, H. J. Salavagione, G. Rothenberg, N. R. Shiju, RSC Advances 2015, 5, 4103-4108. [74] G. C. Gunter, R. H. Langford, J. E. Jackson, D. J. Miller, Industrial & Engineering Chemistry Research 1995, 34, 974-980. [75] J. Zhang, J. Lin, P. Cen, The Canadian Journal of Chemical Engineering 2008, 86, 1047-1053. [76] a) J.-G. Na, B. E. Yi, J. N. Kim, K. B. Yi, S.-Y. Park, J.-H. Park, J.-N. Kim, C. H. Ko, Catalysis Today 2010, 156, 44-48; b) H.-Y. Oh, J.-H. Park, Y.-W. Rhee, J.-N. Kim, Journal of Industrial and Engineering Chemistry 2011, 17, 788-793. [77] H. Wang, D. Yu, P. Sun, J. Yan, Y. Wang, H. Huang, Catalysis Communications 2008, 9, 1799-1803. [78] P. Sun, D. Yu, Z. Tang, H. Li, H. Huang, Industrial & Engineering Chemistry Research 2010, 49, 9082-9087. [79] J. Zhang, Y. Zhao, M. Pan, X. Feng, W. Ji, C.-T. Au, ACS Catalysis 2011, 1, 32- 41. [80] B. Yan, L.-Z. Tao, A. Mahmood, Y. Liang, B.-Q. Xu, ACS Catalysis 2017, 7, 538-550. 6 Bibliography 167

[81] V. C. Ghantani, M. K. Dongare, S. B. Umbarkar, RSC Advances 2014, 4, 33319- 33326. [82] R. Burns, D. T. Jones, P. D. Ritchie, Journal of the Chemical Society (Resumed) 1935, 400-406. [83] a) O. S. Fruchey, T. A. Malisezewski, J. E. Sawyer, WO2013/036389, 2013; b) O. S. Fruchey, T. A. Malisezewski, J. E. Sawyer, WO2014/172540, 2014. [84] a) M. Bicker, J. Hirth, H. Vogel, Green Chemistry 2003, 5, 280-284; b) M. Bicker, D. Kaiser, L. Ott, H. Vogel, The journal of supercritical fluids 2005, 36, 118-126; c) V. Lehr, M. Sarlea, L. Ott, H. Vogel, Catalysis Today 2007, 121, 121- 129; d) L. Ott, M. Bicker, A. Kruse, H. Vogel, Chemie Ingenieur Technik 2005, 77, 124-128; e) L. Ott, S. Kohl, M. Bicker, H. Vogel, Chemical engineering & technology 2005, 28, 1561-1568; f) L. Ott, V. Lehr, S. Urfels, M. Bicker, H. Vogel, The Journal of supercritical fluids 2006, 38, 80-93; g) T. Richter, H. Vogel, Chemical engineering & technology 2001, 24, 340-343. [85] P. E. Savage, Chemical reviews 1999, 99, 603-622. [86] a) J. H. Hong, J. M. Lee, H. Kim, Y. K. Hwang, J. S. Chang, S. B. Halligudi, Y. H. Han, Applied Catalysis A: General 2011, 396, 194-200; b) P. Sun, D. Yu, Z. Tang, H. Li, H. Huang, Industrial and Engineering Chemistry Research 2010, 49, 9082-9087; c) J. Zhang, Y. Zhao, M. Pan, X. Feng, W. Ji, C. T. Au, ACS Catalysis 2011, 1, 32-41. [87] a) M. Avramoff, J. Weiss, O. Schaechter, J. Org. Chem. 1963, 28, 3256-3258; b) D. V. C. Awang, S. Wolfe, Can. J. Chem. 1969, 47, 706-709; c) R. Bloch, Synthesis 1978, 1978, 140-142; d) U. Bora, G. Bose, M. K. Chaudhuri, S. S. Dhar, R. Gopinath, A. T. Khan, B. K. Patel, Org. Lett. 2000, 2, 247-249; e) S. Chandrappa, M. Umashankara, K. Vinaya, C. S. A. Kumar, K. S. Rangappa, Tetrahedron Lett. 2012, 53, 2632-2635; f) H. J. Dauben, L. McCoy, J. Am. Chem. Soc. 1959, 81, 4863-4873; g) R. R. Dim, C. R. Melgarejo, M. T. P. López- Espinosa, I. I. Cubero, J. Org. Chem. 1994, 59, 7928-7929; h) C. Djerassi, Chem. Rev. 1948, 43, 271-317; i) F. L. Greenwood, M. D. Kellert, J. Sedlak, Org. Synth. Collect. Vol. 1958, 38, 8; j) I. V. Koval’, Russ. J. Org. Chem. 2002, 38, 301-337; k) A. Wohl, Ber. Dtsch. Chem. Ges. B 1919, 52B, 51-63. [88] a) S. Dunstan, H. B. Henbest, J. Chem. Soc. 1957, 4905; b) R. Heilmann, P. Baret, C. R. Acad. Sci. C: Sci. Chim. 1968, 267, 579; c) J. D. Prugh, W. C. McCarthy, Tetrahedron Lett. 1966, 7, 1351; d) P. S. Radhakrishnamurthy, N. C. Sahu, Indian J. Chem. 1981, 20A, 269; e) B. Singh, R. Srivastava, Curr. Sci. 1985, 54, 415; f) N. Venkatasubramanian, V. Thiagarajan, Can. J. Chem. 1969, 47, 694. [89] a) D. D. Agarwal, V. Agarwal, J. Indian Chem. Soc. 1991, 68, 360; b) A. F. Patrocínio, P. J. S. Moran, J. Organomet. Chem. 2000, 603, 220. [90] a) N. A. Petasis, I. A. Zavialov, Tetrahedron Lett. 1996, 37, 567; b) C. Thiebes, G. K. Surya Prakash, N. A. Petasis, G. A. Olah, Synlett 1998, 1998, 141. [91] a) B. Akhlaghinia, A. Pourali, Turk. J. Chem. 2010, 34, 753; b) D. Azarifar, H. Ghasemnejad Bosra, M. Tajbaksh, J. Heterocycl. Chem. 2007, 44, 467; c) S. Habibzadeh, H. Ghasemnejad, M. Faraji, Helv. Chim. Acta 2011, 94, 429; d) A. T. Khan, S. Ali, A. A. Dar, M. Lal, Tetrahedron Lett. 2011, 52, 5157; e) A. T. Khan, M. M. Khan, A. Adhikary, Carbohydr. Res. 2011, 346, 673; f) M. A. Pasha, V. P. Jayashankara, Indian J. Chem. 2007, 46B, 1328; g) S. K. Talluri, A. Sudalai, Org. Lett. 2005, 7, 855. 168

[92] T. Bastock, M. Holding, Bromination and production techniques for fine chemicals, Contract Chem Ltd., Prescot, 2004. [93] Y.-S. Hwang, H. A. Navvab-Gojrati, M. S. Mulla, Journal of Agricultural and Food Chemistry 1978, 26, 1293-1296. [94] a) A. Bathe, S. Emmert, B. Helfert, H. Boettcher, Preparation of 5-(1- Piperazinyl)-benzofuran-2-carboxamide 2001; b) P. B. Deshpande, B. B. Boda, S. S. Surti, P. P. Shah, Process for preparation of Tadalafil and its intermediate 2007; c) M. Hans, Phenylaminoacetamides for regulating plantgrowth 1977; d) S. A. Harris, K. Folkers, Journal of the American Chemical Society 1939, 61, 1245-1247; e) B. S. Lee, I. K. Park, S. F. Shin, Preparation of 6-hydroxy-2-oxo-1,2,3,4- tetrahydroquinoline 2000; f) X. T. Li, J. Sui, X. M. Yang, Environment friendly synthesis of 1-chloroethyl-4-isobutylphenyl ketone 2005; g) C. Loncaric, W. D. Wulff, Organic Letters 2001, 3, 3675-3678; h) Z. Zhang, Q. Zhang, J. Wang, X. Shi, J. Zhang, H. Song, Carbohydrate Polymers 2010, 79, 628-632. [95] M. A. Kaldrikyan, N. S. Minasyan, R. G. Melik-Ogandzanyan, Russian Journal of General Chemistry 2016, 86, 305-311. [96] a) J. R. Dimmock, R. N. Puthucode, J. M. Smith, M. Hetherington, J. W. Quail, U. Pugazhenthi, T. Lechler, J. Stables, J Med Chem 1996, 39, 3984-3997; b) Z. Polivka, J. Holubek, E. Svatek, J. Metys, M. Protiva, Collect Czech Chem Commun 1984, 49, 621-623; c) J. Ragavendran, D. Sriram, S. Patil, I. V. Reddy, N. Bharathwajan, J. Stables, P. Yogeeswari, Eur J Med Chem 2007, 42, 146-151; d) R. P. Shank, D. R. Doose, A. J. Streeter, M. Bialer, Epilepsy Res 2005, 63, 103-112; e) P. Yogeeswari, R. Thirumurugan, R. Kavya, J. S. Samuel, J. Stables, D. Siram, Eur J Med Chem 2004, 39, 729-734. [97] a) M. Kehrer, J. Mehler, N. Taccardi, J. Nagengast, J. Kadar, D. Collias, P. Dziezok, P. Wasserscheid, J. Albert, ChemSusChem 2018, 11, 1063-1072; b) M. Kehrer, J. Nagengast, J. Kadar, D. Collias, P. Dziezok, J. Albert, P. Wasserscheid, "Novel liquid-phase technology for the production of bio- acrylic acid", 51. Jahrestreffen dt. Katalytiker, Weimar, oral presentation, 15.- 17.03.2017; c) J. Nagengast, S. Hahn, N. Taccardi, M. Kehrer, J. Kadar, D. Collias, P. Dziezok, P. Wasserscheid, J. Albert, ChemSusChem, Accepted Author Manuscript. doi:10.1002/cssc.201800914; d) P. Wasserscheid, J. Albert, "Novel , selective catalytic routes to organic acids from biomass", 3rd International Symposium on Green & Sustainable Chemistry (ISGC), La Rochelle, plenary presentation, 16.-19.05.2017. [98] a) K. A. Cooper, M. L. Dhar, E. D. Hughes, C. K. Ingold, B. J. MacNulty, L. I. Woolf, Journal of the Chemical Society (Resumed) 1948, 2043-2049; b) E. D. Hughes, C. K. Ingold, Journal of the Chemical Society (Resumed) 1935, 244-255. [99] a) J. K. Laerdahl, E. Uggerud, International Journal of Mass Spectrometry 2002, 214, 277-314; b) D. J. Raber, J. M. Harris, Journal of Chemical Education 1972, 49, 60-64. [100] A. R. Katritzky, B. E. Brycki, Journal of Physical Organic Chemistry 1988, 1, 1- 20. [101] D. G. Hall, C. Gupta, T. H. Morton, Journal of the American Chemical Society 1981, 103, 2416-2417. [102] L. R. Domingo, J. Andrés, V. Moliner, V. S. Safont, Journal of the American Chemical Society 1997, 119, 6415-6422. 6 Bibliography 169

[103] a) Product Stewardship Summary: Alkyl bromides, Great Lake Solutions, West Lafayette, 2015; b) J. Houben, T. Weyl, Methoden der organischen Chemie, 4th ed., Thieme, Stuttgart, 1960; c) B. M. Trost, I. Fleming, E. Winterfeldt, Comprehensive Organic Synthesis, Vol. 6, Elsevier, Oxford, 1991. [104] A. Kekulé, Justus Liebigs Annalen der Chemie 1864, 130, 11-31. [105] D. R. Kirklin, E. S. Domalski, The Journal of Chemical Thermodynamics 1988, 20, 743-754. [106] J. P. Schaefer, D. S. Weinberg, The Journal of Organic Chemistry 1965, 30, 2635- 2639. [107] M. M. Ponpipom, S. Hanessian, Carbohydrate Research 1971, 18, 342-344. [108] B. Classon, Z. Liu, B. Samuelsson, The Journal of Organic Chemistry 1988, 53, 6126-6130. [109] R. Appel, Angewandte Chemie International Edition in English 1975, 14, 801-811. [110] V. S. C. de Andrade, M. C. S. de Mattos, Current Organic Synthesis 2015, 12, 309-327. [111] a) R. O. Recknagel, Pharmacological Reviews 1967, 19, 145-208; b) J. B. Sullivan, G. R. Kneges, Clinical Environmental Health and Toxic Exposures, 2 ed., Lippincott Willians & Willians, Philadelphia, 2001. [112] a) M. König, A. Linhardt, O. Brüggemann, I. Teasdale, Monatshefte für Chemie - Chemical Monthly 2016, 147, 1575-1582; b) K. A. Roper, H. Lange, A. Polyzos, M. B. Berry, I. R. Baxendale, S. V. Ley, Beilstein Journal of Organic Chemistry 2011, 7, 1648-1655. [113] H. A. van Kalkeren, A. L. Blom, F. P. J. T. Rutjes, M. A. J. Huijbregts, Green Chemistry 2013, 15, 1255-1263. [114] a) R. M. Denton, J. An, B. Adeniran, A. J. Blake, W. Lewis, A. M. Poulton, The Journal of Organic Chemistry 2011, 76, 6749-6767; b) H. A. van Kalkeren, S. H. A. M. Leenders, C. R. A. Hommersom, F. P. J. T. Rutjes, F. L. van Delft, Chemistry – A European Journal 2011, 17, 11290-11295. [115] M.-F. Pouliot, O. Mahé, J.-D. Hamel, J. Desroches, J.-F. Paquin, Organic Letters 2012, 14, 5428-5431. [116] J. D. Pelletier, D. Poirier, Tetrahedron Letters 1994, 35, 1051-1054. [117] a) K. P. C. Vollhardt, N. E. Schore, Organische Chemie, John Wiley & Sons, 2011; b) J. March, Advanced organic chemistry: reactions, mechanisms, and structure, John Wiley & Sons, 1992. [118] R. K. Bansal, A Textbook of Organic Chemistry, New Age International, 2007. [119] M. Rossberg, W. Lendle, G. Pfleiderer, A. Tögel, E.-L. Dreher, E. Langer, H. Rassaerts, P. Kleinschmidt, H. Strack, R. Cook, U. Beck, K.-A. Lipper, T. R. Torkelson, E. Löser, K. K. Beutel, T. Mann, in Ullmann's Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, 2000. [120] P. Wasserscheid, T. Welton, Ionic liquids in synthesis, John Wiley & Sons, 2008. [121] S. Gabriel, J. Weiner, European Journal of Inorganic Chemistry 1888, 21, 2669- 2679. [122] P. Walden, Bull. Acad. Imper. Sci.(St. Petersburg) 1914, 1800. [123] a) T. Welton, Coordination Chemistry Reviews 2004, 248, 2459-2477; b) Y. Beste, M. Eggersmann, H. Schoenmakers, Chemie Ingenieur Technik 2005, 77, 1800- 1808. [124] a) Sigma-Aldrich, Tetrabutylphosphonium bromide, Safety Data Sheet, https://www.sigmaaldrich.com/MSDS/MSDS/DisplayMSDSPage.do?country 170

=DE&language=de&productNumber=189138&brand=ALDRICH&PageToGo ToURL=https%3A%2F%2Fwww.sigmaaldrich.com%2Fcatalog%2Fproduct% 2Faldrich%2F189138%3Flang%3Dde, accessed 01-02-2018; b) D. Xu, Q. Yang, B. Su, Z. Bao, Q. Ren, H. Xing, The Journal of Physical Chemistry B 2014, 118, 1071-1079; c) M. Stalpaert, F. G. Cirujano, D. E. De Vos, ACS Catalysis 2017, 7, 5802-5809. [125] C. J. Bradaric, A. Downard, C. Kennedy, A. J. Robertson, Y. Zhou, Green Chemistry 2003, 5, 143-152. [126] M. A. Kareem, F. S. Mjalli, M. A. Hashim, M. K. O. Hadj-Kali, F. S. G. Bagh, I. M. Alnashef, Fluid Phase Equilibria 2012, 333, 47-54. [127] a) N. Mayoufi, D. Dalmazzone, A. Delahaye, P. Clain, L. Fournaison, W. Furst,̈ Journal of Chemical & Engineering Data 2011, 56, 2987-2993; b) T. Suginaka, H. Sakamoto, K. Iino, Y. Sakakibara, R. Ohmura, Fluid Phase Equilibria 2013, 344, 108-111; c) T. Suginaka, H. Sakamoto, K. Iino, S. Takeya, M. Nakajima, R. Ohmura, Fluid Phase Equilibria 2012, 317, 25-28. [128] J. F. Knifton, Journal of the American Chemical Society 1981, 103, 3959-3961. [129] a) N. Gupta, G. L. Kad, J. Singh, Journal of Molecular Catalysis A: Chemical 2009, 302, 11-14; b) N. E. Leadbeater, H. M. Torenius, H. Tye, Tetrahedron 2003, 59, 2253-2258; c) H.-P. Nguyen, P. Kirilov, H. Matondo, M. Baboulène, Journal of Molecular Catalysis A: Chemical 2004, 218, 41-45; d) B. C. Ranu, R. Jana, European Journal of Organic Chemistry 2005, 2005, 755-758; e) R. X. Ren, J. X. Wu, Organic Letters 2001, 3, 3727-3728. [130] J.-M. Mas, P. Metivier, Synthetic Communications 1992, 22, 2187-2191. [131] V. Stegmann, K. Massonne, WO2005/026089A2, 2005. [132] M. Strebelle, WO2012/084624A1, 2012. [133] H. Li, J. Liu, J. Zhu, H. Wang, Journal of the Korean Chemical Society 2011, 55, 685-690. [134] M. Yoshizawa, M. Hirao, K. Ito-Akita, H. Ohno, Journal of Materials Chemistry 2001, 11, 1057-1062. [135] A. C. Cole, J. L. Jensen, I. Ntai, K. L. T. Tran, K. J. Weaver, D. C. Forbes, J. H. Davis, Journal of the American Chemical Society 2002, 124, 5962-5963. [136] a) J. H. Hong, J.-M. Lee, H. Kim, Y. K. Hwang, J.-S. Chang, S. B. Halligudi, Y.-H. Han, Applied Catalysis A: General 2011, 396, 194-200; b) B. Katryniok, S. Paul, F. Dumeignil, Green Chemistry 2010, 12, 1910-1913. [137] Z. Zhai, X. Li, C. Tang, J. Peng, N. Jiang, W. Bai, H. Gao, Y. Liao, Industrial & Engineering Chemistry Research 2014, 53, 10318-10327. [138] D. C. Wadley, M. S. Tam, P. B. Kokitkar, J. E. Jackson, D. J. Miller, Journal of Catalysis 1997, 165, 162-171. [139] a) V. Piemonte, S. Sabatini, F. Gironi, Journal of Polymers and the Environment 2013, 21, 640-647; b) D.-K. Yoo, D. Kim, Polymer Bulletin 2009, 63, 637-651. [140] R. Montgomery, Journal of the American Chemical Society 1952, 74, 1466-1468. [141] a) P. Lan, W. Wang, Journal of Fiber Bioengineering & Informatics 2009, 2, 137- 140; b) E. Lillie, R. C. Schulz, Die Makromolekulare Chemie 1975, 176, 1901- 1906; c) K. A. M. Thakur, R. T. Kean, E. S. Hall, J. J. Kolstad, E. J. Munson, International Journal of Polymer Analysis and Characterization 1998, 4, 379-391. [142] S. Brownstein, A. E. Stillman, The Journal of Physical Chemistry 1959, 63, 2061- 2062. 171

[143] a) A. Berthod, S. Carda-Broch, Journal of Chromatography A 2004, 1037, 3-14; b) E. Müller, R. Berger, E. Blass, D. Sluyts, A. Pfennig, in Ullmann's Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, 2000. [144] a) Y. Wang, X. Gong, Z. Wang, L. Dai, Journal of Molecular Catalysis A: Chemical 2010, 322, 7-16; b) D.-Q. Xu, J. Wu, S.-P. Luo, J.-X. Zhang, J.-Y. Wu, X.-H. Du, Z.-Y. Xu, Green Chemistry 2009, 11, 1239-1246. [145] C. Reichardt, Green Chemistry 2005, 7, 339-351. [146] E. T. Shawl, H. S. Kesling, US5227483, 1993. [147] R. D. Devries, US4640831, 1987.

172

7 Appendix

7.1 Further data and information concerning LA bromination

7.1.1 Summary of LA bromination data

Table A 1: LA bromination with HBr(aq) using different LA:HBr molar ratios: Screening of reaction parameters[a].

LA:HBr molar ra- Temperature / Time / Y2-BrPA / S2-BrPA / Loss / tio °C h % % mol%

1:1 100 24 15.0 36.9 38

1:1 110 5 8.6 67.0 4

1:1 120 5 16.7 64 9.3

1:3 90 24 5.0 71.0 2

1:3 100 72 19.6 37.7 31

1:3 100 24 12.7 53.0 13

1:3 110 24 25.0 49.3 23

1:3 110 5 10.0 62.5 6

1:3 120 24 26.6 39.9 39

1:3 120 5 20.0 50.0 20

1:6 100 72 24.1 50.7 25

1:6 110 5 10 56.5 8

1:6 120 5 21 48 23

1:9 100 72 25.4 54.0 22

1:12 100 72 27.0 55.0 23

[a] 0.05 mol lactide, X mol HBr(aq), ambient pressure, quantified via 1H NMR using AcOH as internal standard.[C]

7 Appendix 173

Table A 2: Concentration-dependent Y2-BrPA, S2-BrPA and XLA. [a]

Substrate concentration / wt % XLA / % Y2-BrPA / % S2-BrPA / %

20 0 0 -

40 8 5 62

60 26 13 50

80 38 20 53

88 (PURAC®) 45 23 51

100 (PURALACT®) 66 26 39

[a] 0.20 mol lactide, 0.60 mol HBr(aq), X mol H2O,105 °C, 72 h, ambient pressure, quantified via 1H NMR using AcOH as internal standard.[C]

Table A 3: Variation of the reaction temperature of 2-BrPA synthesis with HBr(aq).[a]

Reaction temperature / °C XLA / % Y2-BrPA / % S2-BrPA / % Loss / mol% 110 16.0 10.0 62.5 6

120 40.0 20.0 50.0 20

[a] 0.5 mol lactide, 0.30 mol HBr(aq), 5 h, ambient pressure, quantified via 1H NMR using AcOH as internal standard.[C]

174

Table A 4: Summary of LA bromination with HBr(aq) and [PBu4]Br at various LA:HBr:IL molar ratios. [a]

LA:HBr:IL T / t / Y S Y Y / Loss / r r 2-BrPA 2-BrPA 3-BrPA AA Substrate molar ratio °C h / % / % / % % mol% 1:1:1 100 72 21.4 46.2 - - 25.0 PURALACT®

1:1:5 100 72 10.0 33.8 - - 19.6 PURALACT®

1:1:5 120 5 3.5 32.0 - - 7.4 PURAC®

1:3:0.33 120 5 10.0 57.0 - - 7.5 PURAC®

1:1:1 120 5 7.4 66.0 - - 3.8 PURAC®

1:3:1 120 5 9.0 54.5 - - 7.5 PURAC®

1:3:5 120 5 12.5 47.0 - - 13.9 PURALACT®

1:3:5 140 5 31.4 63.0 - - 19.6 PURALACT®

1:3:5 140 24 32.7 63.5 16.2 - 27.0 PURALACT®

1:3:5 140 5 23.0 52.0 - - 21.0 PURAC®

1:1:5 140 5 4.0 10.0 - - 37.0 PURAC®

1:1:5 160 5 <1.0 18.2 5.6 10.8 73.6 PURAC®

1:3:10 140 5 15.0 25.0 41.0 PURAC®

[a] 0.05 mol lactide/0.10 mol PURAC, X mol HBr(aq) and [PBu4]Br, ambient pressure, quantified via 1H NMR using AcOH as internal standard.[C]

7 Appendix 175

7.1.2 Bromide salt and precursor synthesis

4-(3-Methylimidazolium)-butane-1-sulfonate (MIMBS)

1-Methylimidazole (60.26 g, 0.734 mol, 1 eq.) and 1,4-butanesultone (100 g, 0.734 mol,

1 eq.) were dissolved in 500 mL of acetonitrile and degassed with argon for 15 min.

The reaction mixture was heated to reflux (82°C) and stirred for 48 h, where the product precipitated over course of reaction as a white solid. After cooling to room temperature, 500 mL of cold ethyl acetate were added and the reaction mixture was stirred for additional 15 min. After filtration, the residue was first washed with toluene (3 x 100 mL), then with ether (3 x 100 mL) and finally dried in vacuum to give the product as a white powder (154.33 g, 0.707 mol, 96.3%).

1H NMR (400 MHz, MeOH-d4, ppm): δ= 9.00 (s, 1H, 1), 7.67 (s, 1H, 2), 7.58 (s, 1H, 3), 4.27 (t, 2H, 5), 3.93 (s, 3H, 4), 2.83 (t, 2H, 8), 2.05 (m, 2H, 6), 1.75 (m, 2H, 7).

13C NMR (100 MHz, MeOH-d4, ppm): δ= 136.81 (1C, 1), 123.69 (1C, 2), 122.40 (1C, 3), 50.26 (1C, 8), 48.96 (1C, 5), 35.22 (1C, 4), 28.64 (1C, 6), 21.53 (1C, 7).

176

Figure A 1: 1H NMR spectrum of MIMBS in MeOH-d4.

Figure A 2: 13C NMR spectrum of MIMBS in MeOH-d4. 7 Appendix 177

4-(3-Ethylimidazolium)-butane-1-sulfonate (EIMBS):

1-Ethylimidazol (48.07 g, 0.5 mol, 1 eq.) and 1,4-butanesultone (68.09 g, 0.5 mol, 1 eq.) were dissolved in 300 mL of acetonitrile and degassed with argon for 15 min. The reaction mixture was heated to reflux (82°C) and stirred for 48 h, where the product precipitated over course of reaction as a white solid. After cooling to room temperature, 300 mL of cold ethyl acetate were added and the reaction mixture was stirred for an additional 15 min. After filtration, the residue was first washed with toluene

(3 x 100 mL), then with ether (3 x 100 mL) and finally dried in vacuum to give the product as a white powder (107.73 g, 0.464 mol, 92.7%).

1H NMR (400 MHz, MeOH-d4, ppm): δ= 9.10 (s, 1H, 1), 7.70 (m, 2H, 2/3), 4.31 (m, 4H, 4/6), 2.85 (t, 2H, 9), 2.05 (m, 2H, 7), 1.75 (m, 2H, 8), 1.52 (t, 3H, 5).

13C NMR (100 MHz, MeOH-d4, ppm): δ= 135.85 (1C, 1), 125.56 (1C, 2), 122.21 (1C, 3), 50.36 (1C, 9), 49.03 (1C, 6), 44.78 (1C, 4), 28.70 (1C, 7), 21.63 (1C, 8), 14.39 (1C, 5).

178

Figure A 3: 1H NMR spectrum of EIMBS in MeOH-d4.

Figure A 4: 13C NMR spectrum of EIMBS in MeOH-d4. 7 Appendix 179

4-(3-Butylimidazolium)-butane-1-sulfonate (BIMBS):

1-Butylimidazol (62.09 g, 0.5 mol, 1 eq.) and 1,4-butanesultone (68.09 g, 0.5 mol, 1 eq.) were dissolved in 200 mL of acetonitrile and degassed with argon for 15 min. The reaction mixture was heated to reflux (82°C) and stirred for 48 h. After cooling to room temperature and addition of 300 mL of cold ethyl acetate, a crude product precipitated as a white solid. To ensure complete precipitation, the reaction mixture was stirred for additional 15 min. After filtration, the residue was first washed with toluene (3 x 200 mL), then with ether (3 x 200 mL) and finally dried in vacuum to give the product as a white powder (125.22 g, 0.481 mol, 96.2%).

1H NMR (400 MHz, MeOH-d4, ppm): δ= 9.09 (s, 1H, 1), 7.70 (s, 1H, 2), 7.67 (s, 1H, 3) 4.24 (m, 4H, 4/8), 2.84 (t, 2H, 11), 2.05 (m, 2H, 9), 1.78 (m, 4H, 10/5), 1.35 (m, 2H, 6), 0.96 (t, 3H, 7).

13C NMR (100 MHz, MeOH-d4, ppm): δ= 136.10 (1C, 1), 122.52 (2C, 2/3), 50.27 (1C, 11), 49.31 (1C, 8), 49.02 (1C, 4), 31.78 (1C, 5), 28.66 (1C, 9), 21.57 (1C, 10), 19.17 (1C, 6), 12.51 (1C, 7).

180

Figure A 5: 1H NMR spectrum of BIMBS in MeOH-d4.

Figure A 6: 13C NMR spectrum of BIMBS in MeOH-d4. 7 Appendix 181

1-Octylimidazole (OIM):

An aqueous sodium hydroxide (NaOH) solution (100 mL, 5 M, 0.5 mol, 1.25 eq.) was heated to 50 °C in an oil-bath and 1-H-imidazole (27.23 g, 0.4 mol, 1 eq.) was added in small portions under stirring. Due to formation of sodium imidazolate, the solution turned slowly yellow. The reaction mixture was cooled to room temperature and 1- octylbromide (75.32 g, 0.39 mol, 0.975 eq.) was added slowly. The temperature was increased to 70°C and the reaction mixture was stirred for 72 h, which led to the formation of a second, organic phase. The phases were separated and the organic layer was distilled in vacuum at 105°C and 1 mbar to yield the product as viscous, colorless oil (51.06 g, 0.283 mol, 71.8%).

1H NMR (400 MHz, DMSO-d6, ppm): δ= 7.55 (s, 1H, 1), 7.09 (s, 1H, 2), 6.83 (s, 1H, 3), 3.88 (t, 2H, 4), 1.63 (m, 2H, 5), 1.18 (m, 10H, 6-10), 0.78 (t, 3H, 11).

Figure A 7: 1H NMR spectrum of OIM in DMSO-d6.

182

4-(3-Octylimidazolium)-butane-1-sulfonate (OIMBS):

1-Octylimidazol (51 06 g, 283 mmol, 1 eq.) and 1,4-butanesultone (38 57 g, 283 mmol,

1 eq.) were dissolved in 200 mL of acetonitrile and degassed with argon for 15 min.

The reaction mixture was heated to reflux (82°C) and stirred for 48 h. After cooling to room temperature and addition of 500 mL of cold ethyl acetate, a crude product formed as a white, oily, bottom layer. The solvent layer was decanted off, 700 mL of ether were added and the mixture was stirred for 30 min, whereby the crude product solidified. After filtration, the residue was washed with ether (5 x 200 mL each) and finally dried in vacuum to give the product as a white powder (80.58 g, 254 mmol, 89.9%).

1H NMR (400 MHz, MeOH-d4, ppm): δ= 9.07 (s, 1H, 1), 7.68 (s, 1H, 2), 7.65 (s, 1H, 3), 4.23 (m, 4H, 4/12), 2.84 (t, 2H, 15), 2.04 (m, 2H, 13), 1.88 (m, 2H, 5), 1.77 (m, 2H, 14), 1.28 (m, 10H, 6-10), 0.88 (t, 3H, 11).

13C NMR (100 MHz, MeOH-d4, ppm): δ= 136.04 (1C, 1), 122.53 (2C, 2/3), 50.18 (1C, 15), 49.57 (1C, 12), 49.02 (1C, 4), 31.59–26.00 (m, 6C, 5-9,13), 22.37 (1C, 10), 21.51 (1C, 14), 13.14 (1C, 11).

7 Appendix 183

Figure A 8: 1H NMR spectrum of OIMBS in MeOH-d4.

Figure A 9: 13C NMR spectrum of OIMBS in MeOH-d4. 184

3-(3-Methylimidazolium)-propane-1-sulfonate (MIMPS):

1-Methylimidazol (16.80 g, 205 mmol, 1 eq.) and 1,3-propanesultone (25 g, 205 mmol,

1 eq.) were dissolved in 125 mL of acetonitrile and degassed with argon for 15 min.

The reaction mixture was heated to reflux (82°C) and stirred for 3 h, where the product precipitated over course of reaction as a white solid. After cooling to room temperature, 100 mL of cold toluene were added and the reaction mixture was stirred for additional 15 min. After filtration, the residue was first washed with toluene (3 x 100 mL), then with ether (3 x 50 mL) and finally dried in vacuum to give the product as a white powder (40.00 g, 196 mmol, 95.7%).

1H NMR (400 MHz, D2O, ppm): δ= 8.59 (s, 1H, 1), 7.36 (s, 1H, 2), 7.27 (s, 1H, 3), 4.19 (t, 2H, 5), 3.72 (s, 3H, 4), 2.75 (t, 2H, 7), 2.14 (m, 2H, 6).

13C NMR (100 MHz, D2O, ppm): δ= 136.22 (1C, 1), 123.78 (1C, 2), 122.20 (1C, 3), 47.74 (1C, 7), 47.22 (1C, 5), 35.73 (1C, 4), 25.13 (1C, 6).

7 Appendix 185

Figure A 10: 1H NMR spectrum of MIMPS in D2O.

Figure A 11: 13C NMR spectrum of MIMPS in D2O. 186

4-(N-pyridinium)-butane-1-sulfonate (PyrBS)

Pyridine (97.06 g, 1.23 mol, 1 eq.) and 1,4-butanesultone (175 44 g, 1.29 mol, 1.05 eq.) were dissolved in 600 mL of acetonitrile. The reaction mixture was heated to reflux

(82°C) and stirred for 48 h, where the product precipitated over course of reaction as a white solid. After cooling to room temperature, 400 mL of cold toluene were added and the reaction mixture was stirred for additional 15 min. After filtration, the residue was first washed with toluene (3 x 200 mL), then ether (3 x 200 mL) and finally dried in vacuum to give the product as a white powder (252.70 g, 1.17 mol, 95.4%).

1H NMR (400 MHz, D2O, ppm): δ= 8.72 (m, 2H, 1), 8.39 (m, 1H, 3), 7.92 (m, 2H, 2), 4.49 (t, 2H, 4), 2.50 (t, 2H, 7), 2.02 (m, 2H, 5), 1.64 (m, 2H, 6).

13C NMR (100 MHz, D2O, ppm): δ= 145.72-144.31 (m, 3C, 1/3), 128.35 (2C, 2), 61.23 (1C, 4), 50.05 (1C, 7), 29.39 (1C, 5), 20.95 (1C, 6).

7 Appendix 187

Figure A 12: 1H NMR spectrum of PyrBS in D2O.

Figure A 13: 13C NMR spectrum of PyrBS in D2O. 188

4-(Triphenylphosphonium)-butane-1-sulfonate (PPh3BS):

Triphenylphosphine (39.44 g, 150 mmol, 1 eq.) was dissolved in mesitylene (150 mL) and 1,4-butanesultone (20.43 g, 150 mmol, 1 eq.) was added in one portion under stirring. The reaction mixture was degassed for 15 min, heated to reflux (165°C) and stirred for 72 h. Over course of reaction the product precipitated as a white solid. Af- ter cooling to room temperature, 100 mL of cold toluene were added and the reaction mixture was stirred for additional 15 min. After filtration, the residue was first washed with toluene (3 x 50 mL), then with ether (3 x 50 mL) and finally dried in vacuum to give the product as a white powder (49.45 g, 124 mmol, 82.7%).

1H NMR (400 MHz, DMSO-d6, ppm): δ= 7.87-7.69 (m, 15H, 1-3), 3.54 (m, 2H, 4), 2.42 (m, 2H, 7), 1.73 (m, 2H, 6), 1.60 (m, 2H, 5).

13C NMR (100 MHz, DMSO-d6, ppm): δ= 137.84 (3C, 3), 133.91-133.71 (m, 12C,

1/2), 125.79 (3C, CPh-P), 74.77 (1C, 4), 48.39 (1C, 7), 23.50 (1C, 5), 21.52 (1C, 6). 7 Appendix 189

Figure A 14: 1H NMR spectrum of PPh3BS in DMSO-d6.

Figure A 15: 13C NMR spectrum of PPh3BS in DMSO-d6. 190

4-(Tri-n-butylphosphonium)-butane-1-sulfonate (PBu3BS):

Under inert gas (Ar), a 500 mL three-necked flask was charged with 250 mL of acetonitrile and further degassed via three vacuum-argon cycles. Tri-n-butylphosphine

(22.78 g, 113 mmol, 1 eq.) and 1,4-butanesultone (16.1 g, 118 mmol, 1.05 eq.) were inserted into the reactor using standard Schlenk techniques. The reaction mixture was heated to reflux (82°C) and stirred for 48 h. After cooling to room temperature, acetonitrile was removed under reduced pressure, until the overall volume of the solution was reduced to approximately 100 mL. The reaction mixture was poured on

400 mL of cold ethyl acetate, where a white solid precipitated immediately. After filtration, the residue was first washed with ethyl acetate (3 x 50 mL), then with ether

(3 x 50 mL) and finally dried in vacuum to give the product as a white powder

(31.50 g, 93 mmol, 82.4%).

1H NMR (400 MHz, MeOH-d4, ppm): δ= 2.85 (m, 2H, 8), 2.17 (m, 8H, 1/5), 1.75 (m, 2H, 6), 1.75 (m, 2H, 7), 1.51 (m, 12H, 2/3), 0.98 (m, 9H, 4).

13C NMR (100 MHz, MeOH-d4, ppm): δ= 48.97 (1C, 8), 24.98 (1C, 7), 22.93-22.18 (m, 6C, 2/3), 18.99 (1C, 6), 17.11-16.63 (m, 4C, 1/5), 11.53 (12C, 4). 7 Appendix 191

Figure A 16: 1H NMR spectrum of PBu3BS in MeOH-d4.

Figure A 17: 13C NMR spectrum of PBu3BS in MeOH-d4. 192

[MIMBS]+ [MIMPS]+

[EIMBS]+ [BIMBS]+

[OIMBS]+ [PyrBS]+

[PBu3BS]+

[PPh3BS]+

Figure A 18: Overview of the applied cations and their respective abbreviations.

7 Appendix 193

7.1.3 Error estimation of standard LA bromination with [MIMBS]Br

80 25 calculated mean value`Y and`n

experimentally determined Y2-BrPA experimentally determined n 20 70 2-BrPA

15 / % /

60 mmol /

2-BrPA 10

2-BrPA

Y n

50 5

40 0

I II III IV `Y /`n ± sz

Figure A 19: Error estimation of standard bromination of lactic acid in [MIMBS]Br. conditions: 12.5 mmol lactide, 75 mmol [MIMBS]Br, 120°C, 5 h. Error bar of calculated mean �2-BrPA illustrates the determined root mean square deviation σz from four individual experiments I to IV.

194

7.1.4 Synthesis of a [MIMBS]Br bromination agent with low water content and high HBr content

In the optimized HBr loading procedure, 75 mmoles MIMBS were loaded with HBr(aq) and the resulting HBr loading grade was determined via acid-base titration. The IL was dried at 100°C for 16 h (residual water content <1.5 wt%) and the determined loss of HBr was 18 mmoles (24%). Compensation of the HBr loss was achieved by the addition of 3.03 g of HBr(aq) (48 wt%, 18 mmol HBr) prior to the bromination. The resulting IL contained 0.32 g residual water from the loading procedure and additional

1.58 g of H2O from the addition of HBr(aq) resulting in an overall water content of

1.90 g (approximately 7.5 wt%) and 100 mol% HBr loading grade. In a further experiment, instead of only compensating for the HBr-loss, a 33 mol% excess of HBr (4.21 g,

48 wt%, 25 mmol) was added after full IL drying, resulting in 133 mol% HBr loading grade and an overall water content of 14 wt%. This “over-saturation” of [MIMBS]Br led to free and unbuffered HBr in the reaction matrix.

7 Appendix 195

7.2 Further data and information concerning 2-BrPA isomerization to 3-BrPA

Figure A 20: 1H NMR spectrum of pure 2-BrPA in CDCl3 after thermal treatment up to 180 °C.

2-BrPA 3-BrPA 100 AA LA

C3 balance 80

40 Amountmmol /

0 0 5 10 15 20 25 Reaction time / h

Figure A 21: Quantitative and time-resolved isomerization of 2-BrPA to 3-BrPA in [PBu4]Br. conditions: IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol 24 h, 180 °C, ambient pressure. Error bars 196

illustrate the determined deviation of the C3-balance resulting from quantification errors within the experimental setup.

7.2.1 Summary of isomerization data

Table A 5: Results of 2-BrPA isomerization experiments with [PBu4]Br.

Entry IL:2-BrPA molar ratio / - tr / h Tr / °C Y3-BrPA / % YAA / %

1 1:1 20 120 36 2

2 1:1 20 140 30 2

3 1:1 20 160 79 3

4 1:1 20 180 62 1

5 1:1 3 160 25 2

6 9:1 1 160 19 32

7 9:1 2 160 26 41

8 9:1 3 160 29 47

9 9:1 4 160 30 47

10 9:1 6 160 33 47

11 9:1 3 100 0 0

12 9:1 3 120 4 17

13 9:1 3 140 15 20

14 9:1 3 180 52 45

15 1:3 3 160 4 0

16 3:1 3 160 38 23

17 6:1 3 160 39 31

18 12:1 3 160 28 41

19 15:1 3 160 23 50

20 18:1 3 160 22 54

7 Appendix 197

Table A 6: Results from 2-BrPA isomerization experiments with different bromide ILs at a IL:2-BrPA molar ratio of 1:1.

Entry IL / - Tr / h tr / °C Y3-BrPA / % YAA / % S3-BrPA / %

1 [PBu4]Br 20 140 33 3 62

2 [NBu4]Br 20 140 38 14 58

3 [1B4MPyr]Br 20 140 65 1 82

4 [1E1MPyrro]Br 20 140 69 2 81

5 [EMIM]Br 20 140 74 0 90

6 [MIMBS]Br 20 140 20 0 36

Figure A 22: Overview of used ILs in isomerization of 2-BrPA and their respective abbreviations.

198

2-BrPA 2-BrPA 3-BrPA 3-BrPA 100 AA AA LA LA 10 C3 balance C3-balance 80 8

60 6

Amountmmol / Amountmmol /

20 2

0 0 0 5 10 15 20 25 0 5 10 15 20 25 Reaction time / h Reaction time / h

Figure A 23: Time-resolved conversion of 2-BrPA in [PBu4]Br. Conditions: IL:2-BrPA molar ratio of 1:1

(left) or 9:1 (right), 100 mmol 2-BrPA (left) or 11 mmol (right), 100 mmol [PBu4]Br 24 h, 160 °C, ambient pressure.

100 X2-BrPA Y 90 AA Y3-BrPA

80 / % /

70 3-BrPA 60

& Y & 50 AA

40 , Y , 30

2-BrPA 20 X 10

0 160 180 200 Reaction temperature / °C

Figure A 24: Temperature-dependent conversion of 2-BrPA in [PBu4]Br. Conditions: IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol [PBu4]Br, 3 h, ambient pressure.

7 Appendix 199

7.3 Further data and information concerning 2-BrPA dehydrobromination

Table A 7: Quantities of 2-BrPA conversion in [PBu4]Br at varying process operations.

Temperature / °C quantity batch & ambient semi-batch & vac- semi-batch & strip

pressure uum[a] gas[b]

X2-BrPA = 50 % X2-BrPA = 68 % X2-BrPA = 57 % 160 YAA = 3 % YAA = 11 % YAA = 13 % Y3-BrPA = 21 % Y3-BrPA = 3 % Y3-BrPA = 16 %

X2-BrPA = 85 % X2-BrPA = 96 % X2-BrPA = 92 % 180 YAA = 6 % YAA = 41 % YAA = 25 % Y3-BrPA = 51 % Y3-BrPA = 25 % Y3-BrPA = 15 %

X2-BrPA = 93 % X2-BrPA = 93 % 200 YAA = 7 % n.a. YAA = 27 % Y3-BrPA = 52 % Y3-BrPA = 14 %

[a] Conditions: IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol [PBu4]Br, 3 h. Pressure of . [b] experiment was 5 (160 °C) and 25 mbar (180 °C). Flow of argon strip gas was 500 mLn/min.

200

7.4 List of Figures

Figure 1: Exemplary and simplified flowsheet of commercial AA production...... 9 Figure 2: Structural formula of L-(+)- and D-(-)-lactic acid enantiomers...... 20

Figure 3: Selection of common cations (i) and anions (ii) in terms of IL chemistry (Rn represent alkyl residues, which can be of same or different chain length)...... 49 Figure 4: Flowsheet of batch reactor for bromination of LA and lactide...... 62 Figure 5: Flowsheet of batch reactor for isomerization of 2-BrPA and subsequent product isolation via distillation at reduced pressure...... 64 Figure 6: Flowsheet of the experimental rig for semi-batchwise dehydrobromination of 2-BrPA...... 68 Figure 7: Time-resolved profile of the NADA process: Lactic acid conversion to

acrylic acid in [PBu4]Br. Conditions: 50 mmol LA (PURAC®, 88 wt% LA in

water), 100 mmol [PBu4]Br , 5 mmol 2-BrPA, 168 h, 150 °C, ambient pressure. Lines are only given as a guide to the eye...... 76

Figure 8: Time-dependent LA conversion (XLA), AA yield (YAA) and selectivity (SAA)

of AA production from LA with [PBu4]Br and 2-BrPA. Conditions: 50 mmol

LA (PURAC®, 88 wt% LA in water), 100 mmol [PBu4]Br , 5 mmol 2-BrPA, 168 h, 150 °C, ambient pressure...... 78 Figure 9: Concentration-dependent 2-BrPA synthesis with HBr(aq). Conditions: 0.05 mol lactide, 0.05-0.60 mol HBr(aq), 72 h, 100 °C, ambient pressure.[C] ...... 87

Figure 10: Temperature-dependent 2-BrPA synthesis with HBr(aq). Conditions:

0.05 mol lactide, 0.30 mol HBr(aq), 24 h, ambient pressure.[C] ...... 90

Figure 11: 3D-plot of temperature- and time-dependent Y2-BrPA (top) and S2-BrPA (bottom) in aqueous reaction medium. Conditions: 0.05 mol lactide, 0.30 mol

HBr(aq), ambient pressure, quantified via 1H NMR using AcOH as internal standard.[C] ...... 93 Figure 12: Time-dependent 2-BrPA synthesis in aqueous reaction medium using toluene for in-situ product removal compared to a standard experiment (120

°C) without LLE.0.10 mol PURAC®, 0.30 mol HBr(aq), 34 mL extracting agent, 100 °C, ambient pressure.[C] ...... 97 Figure 13: Extractor-screening with promising organic solvents for in-situ product removal compared to a standard experiment without LLE. 100 mmol

PURAC®, 300 mmol HBr(aq), 24 h, ambient pressure.[C] ...... 99 Figure 14: Time-dependent 2-BrPA synthesis in aqueous reaction medium with a bromide ionic liquid additive. 12.5 mmol lactide, 75 mmol HBr(aq), 125 mmol [PBu4]Br, 120 °C, ambient pressure.[C] ...... 101

Figure 15: Influence of [PBu4Br] concentration and reaction temperature Tr on the synthesis of 2-BrPA with ionic liquid additive. 12.5-125 mmol PURAC®,

37.5-75 mmol HBr(aq), 75-125 mmol [PBu4]Br, ambient pressure.[C] ...... 102 7 Appendix 201

Figure 16: Influence of cation lipophilicity (top) and acidity (bottom) of the zwitterion-based bromide IL on 2-BrPA synthesis; Conditions: 12.5 mmol lactide, 75 mmol IL, 90 mol% HBr loading grade of IL, water content of ~ 3 wt%, 5 h, 120°C; [EIMBS]Br= 1-(4-butane sulfonic acid)-3-ethyl imidazolium bromide, [BIMBS]Br= 1-(4-butane sulfonic acid)-3-butyl imidazolium bromide, [OIMBS]Br= 1-(4-butane sulfonic acid)-3-octyl imidazolium bromide, [MIMPS]Br= 1-(3-propanesulfonic acid)-3-methyl imidazolium bromide.[C] ...... 107 Figure 17: Temperature variation in the 2-BrPA synthesis from lactide using [MIMBS]Br as bromination agent; Conditions: 12.5 mmol lactide, 75 mmol IL, 75 mol% HBr loading grade of IL, water content of < 1.5 wt%, 5 h, ambient pressure.[C] ...... 109 Figure 18: Concentration-dependent 2-BrPA synthesis with [MIMBS]Br; Conditions: X mmol lactide, 75 mmol IL, ~ 90 mol% HBr loading grade of IL, water content of ~ 3 wt%, 5 h, 120°C, ambient pressure.[C] ...... 111 Figure 19: Influence of water content on 2-BrPA synthesis with [MIMBS]Br. Varying

amounts of H2O added prior to bromination after complete drying procedure

(top) and varied residual H2O content by incomplete drying procedure (bottom). Conditions: 12.5 mmol lactide, 75 mmol IL, 5 h, 120°C, ambient pressure.[C] ...... 113 Figure 20: Influence of HBr loading grade and water content on 2-BrPA synthesis with [MIMBS]Br. Conditions: 12.5 mmol lactide, 75 mmol IL, 5 h, 120°C, ambient pressure. IL loading grade varied by preparation method (1, 2) and

addition of HBr(aq) after drying procedure (3, 4).[C] ...... 114 Figure 21: Time-dependent 2-BrPA synthesis with [MIMBS]Br. Five independent bromination experiments with different reaction times. Conditions: 12.5 mmol lactide, 75 mmol IL, full HBr loading grade of IL by compensating HBr lost during drying procedure, IL water content of 7.5 wt%, 120°C.[C] ...... 115 Figure 22: The effect of recycled [MIMBS]Br with 1-heptanol and DBE on the

isolated n2-BrPA and Y2-BrPA of bromination of LA to 2-BrPA. Lines are only given as a guide to the eye...... 118 Figure 23: Substrate-dependent 2-BrPA synthesis with [MIMBS]Br. Conditions: 25 mmol lactic acid (PURAC®) or 12.5 mmol lactide (PURALACT®), 75 mmol IL, full HBr loading grade of IL, water content of < 1.5 wt%, 5 h 120 C, ambient pressure.[C] ...... 119 Figure 24: Partial 1H NMR spectra of an 2-BrPA isomerization experiment with

[PBu4]Br. 1H NMR of pure 2-BrPA (1), 1H NMR of pure [PBu4]Br (2, **traces of water) and 1H NMRs of samples taken at a reaction time of 3 (3) and 20 h (4). Experiment conducted at 140 °C and ambient pressure with a IL:2-BrPA molar

ratio of 1:1. *DMSO-d6 as solvent...... 122 202

Figure 25: Concentration-dependent 3-BrPA synthesis with [PBu4]Br; conditions:

X mmol 2-BrPA, 133.5 mmol [PBu4]Br, 3 h, 160 °C, ambient pressure...... 123

Figure 26: Temperature-dependent 3-BrPA synthesis with [PBu4]Br; conditions:

133.5 mmol 2-BrPA, 133.5 mmol [PBu4]Br, 20 h, ambient pressure...... 125 Figure 27: Quantitative and time-resolved isomerization of 2-BrPA to 3-BrPA in

[PBu4]Br. Conditions: IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA,

100 mmol [PBu4]Br , 24 h, 160 °C, ambient pressure. Error bars illustrate the

determined deviation of the C3-balance resulting from quantification errors within the experimental setup...... 126

Figure 28: Time-resolved C3-balance and S3-BrPA of 2-BrPA conversion in [PBu4]Br. conditions: IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol

[PBu4]Br, 24 h, 160 °C, ambient pressure. Error bars illustrate the determined

deviation of the C3-balance resulting from quantification errors within the experimental setup...... 127 Figure 29: Cation variation for 2-BrPA conversion in bromide ionic liquids; Conditions: IL:2-BrPA molar ratio of 1:1, 140 °C, 20 h, ambient pressure. [PBu4]Br = tetrabutylphosphonium bromide, [NBu4]Br = tetrabutylammonium bromide, [1Bu4MePyr]Br = 1-butyl-4- methyl-pyridinium bromide, [1Et1MePyrro]Br = 1-ethyl-1-methyl- pyrrolidinium bromide, [EMIM]Br = 1-ethyl-3-methylimidazolium bromide, [MIMBS]Br = 3-methyl-1-(4-butane sulfonic aicd) imidazolium bromide...... 131 Figure 30: Partial 1H NMR spectra of an 2-BrPA isomerization experiment with [EMIM]Br and subsequent product isolation via vacuum distillation. 1H NMR of isomerization mixture (1), 1H NMR of purified isomerization mixture (2), and 1H NMRs of received distillate (3). Experiment conducted at 140 °C for 20 h (ambient pressure) with a IL:2-BrPA molar ratio of 1:1. Isolation conducted at 2-10 mbar and a stepwise heated temperature ramp from (100-

160 °C) for 30 min, respectively. * DMSO-d6 as solvent. ** [EMIM]Br 1H NMR signals...... 133 Figure 31: Time-resolved dehydrobromination of 3-BrPA to AA in TOA. Conditions: 50 mmol 3-BrPA, 50 mmol TOA, 80 °C, 5 h, ambient pressure...... 137 Figure 32: Time-resolved dehydrobromination selectivity of 3-BrPA to AA in TOA. Conditions: 50 mmol 3-BrPA, 50 mmol TOA, 80 °C, 5 h, ambient pressure. . 138 Figure 33: Time-resolved profile of base-assisted dehydrobromination of 2-BrPA. Conditions: 50 mmol 2-BrPA, 50 mmol TOA, 240 min, 80 °C, ambient pressure. Lines are only given as a guide to the eye...... 147 Figure 34: Course of 2-BrPA dehydrobromination in TOA monitored by 1H NMR spectroscopy (left): Partial 1H NMR spectra from start (top; 0 min), over the course (mid; 30 min and 60 min) and until the end of the dehydrobromination experiment (bottom; 240 min), indicating ongoing substrate self-esterification. 203

Respective C-H molecule fragment signals are assigned to the related structure (right)...... 148

Figure 35: Time-resolved conversion of 2-BrPA in [PBu4]Br. Conditions: IL:2-BrPA molar ratio of 1:1 (left) or 9:1 (right), 100 mmol 2-BrPA (left) or 11 mmol

(right), 100 mmol [PBu4]Br, 24 h, 180 °C, ambient pressure...... 150

Figure 36: Conversion of 2-BrPA in [PBu4]Br at different experimental operation modes. Conditions: IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol

[PBu4]Br, 3 h, 180 °C, ambient pressure (batch), 25 mbar (vacuum) and 500 mLn/min flow of Ar (strip gas)...... 153

Figure A 1: 1H NMR spectrum of MIMBS in MeOH-d4...... 176

Figure A 2: 13C NMR spectrum of MIMBS in MeOH-d4...... 176

Figure A 3: 1H NMR spectrum of EIMBS in MeOH-d4...... 178

Figure A 4: 13C NMR spectrum of EIMBS in MeOH-d4...... 178

Figure A 5: 1H NMR spectrum of BIMBS in MeOH-d4...... 180

Figure A 6: 13C NMR spectrum of BIMBS in MeOH-d4...... 180

Figure A 7: 1H NMR spectrum of OIM in DMSO-d6...... 181

Figure A 8: 1H NMR spectrum of OIMBS in MeOH-d4...... 183

Figure A 9: 13C NMR spectrum of OIMBS in MeOH-d4...... 183

Figure A 10: 1H NMR spectrum of MIMPS in D2O...... 185

Figure A 11: 13C NMR spectrum of MIMPS in D2O...... 185

Figure A 12: 1H NMR spectrum of PyrBS in D2O...... 187

Figure A 13: 13C NMR spectrum of PyrBS in D2O...... 187

Figure A 14: 1H NMR spectrum of PPh3BS in DMSO-d6...... 189

Figure A 15: 13C NMR spectrum of PPh3BS in DMSO-d6...... 189

Figure A 16: 1H NMR spectrum of PBu3BS in MeOH-d4...... 191

Figure A 17: 13C NMR spectrum of PBu3BS in MeOH-d4...... 191 Figure A 18: Overview of the applied cations and their respective abbreviations. . 192 Figure A 19: Error estimation of standard bromination of lactic acid in [MIMBS]Br. Conditions: 12.5 mmol lactide, 75 mmol [MIMBS]Br, 120°C, 5 h. Error bar of calculated mean �2-BrPA illustrates the determined root mean square

deviation σz from four individual experiments I to IV...... 193

Figure A 20: 1H NMR spectrum of pure 2-BrPA in CDCl3 after thermal treatment up to 180 °C...... 195 Figure A 21: Quantitative and time-resolved isomerization of 2-BrPA to 3-BrPA in

[PBu4]Br. conditions: IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol 24 h, 180 °C, ambient pressure. Error bars illustrate the determined 204

deviation of the C3-balance resulting from quantification errors within the experimental setup...... 195 Figure A 22: Overview of used ILs in isomerization of 2-BrPA and their respective abbreviations...... 197

Figure A 23: Time-resolved conversion of 2-BrPA in [PBu4]Br. Conditions: IL:2-BrPA molar ratio of 1:1 (left) or 9:1 (right), 100 mmol 2-BrPA (left) or 11 mmol

(right), 100 mmol [PBu4]Br 24 h, 160 °C, ambient pressure...... 198

Figure A 24: Temperature-dependent conversion of 2-BrPA in [PBu4]Br. Conditions:

IL:2-BrPA molar ratio of 1:1, 100 mmol 2-BrPA, 100 mmol [PBu4]Br, 3 h, ambient pressure...... 198

205

7.5 List of Schemes

Scheme 1: Polymerization and esterification as main product pathways starting from the AA monomer, according to Beerthuis et al.[2a] ...... 5 Scheme 2: Two-step propene oxidation as reaction pathway for commercial production of AA...... 7 Scheme 3: Overview of production routes to AA, starting from biomass feedstock (green) or crude oil (black) and proceeding via bio-based platform chemicals (blue) or petro-based intermediates (black); LA = lactic acid; 3-HPA = 3- hydroxypropionaldehyde; 3-HP = 3-hydroxypropionic acid; 2-APA = 2-acetoxypropionic acid; investigated route of this thesis is highlighted by bold arrows; adapted from Beerthuis et al.[2a]...... 11

Scheme 4: Reversible dimer (L2A) formation from two LA molecules via intermolecular esterification (i) and formation and hydrolysis of LA oligomers (ii); adapted from Dusselier et al.[5c] ...... 19 Scheme 5: Overview of different LA production methods: Commercial and biotechnological process (i) and chemical synthesis from bio-based (ii) and petro-based (iii) feedstocks; adapted from Beerthuis et al.[2a] ...... 21 Scheme 6: Simplified metabolic pathway from glucose to lactate; according to Auras et al.[61a] ...... 22 Scheme 7: The central role of LA (cyan) as platform chemical for the synthesis of LA derivatives and PLA (orange) or LA-derived platform chemicals (blue)...... 23 Scheme 8: LA oligomers and lactide as self-esterified LA intermediates en route to PLA...... 25 Scheme 9: LA derivatives formed in the presence of acids and alcohols or bases. .... 26 Scheme 10: Mechanisms of LA decarbonylation (i) and decarboxylation (ii) triggered by activation of carboxylic group via protonation (i) or nucleophilic attack (ii)...... 27

Scheme 11: Dehydration of LA to AA over NaSO4/CaSO4 catalyst at 400 °C according to Holmen (1958).[6]...... 29 Scheme 12: Proposed mechanism for the catalytic cracking of lactide to AA by MSA

in PPh4Br; according to Terrade et al.[10b] ...... 36 Scheme 13: Structural formula of 2- (left) and 3-bromopropionic acid (right) isomers...... 38 Scheme 14: Mechanism of the Hell-Volhardt-Zelinsky reaction for the production of 2-BrPA from petroleum-based PA...... 38 Scheme 15: Bromination of LA with HBr including both mechanistic opportunities

SN1 (top) and SN2 (bottom)...... 41 Scheme 16: Hydrobromination of AA via addition of HBr...... 45 Scheme 17: Base-promoted dehydrobromination of 3-BrPA to AA...... 46 206

Scheme 18: Quaternization reaction of tributylphosphine and 1-bromobutane to

[PBu4]Br; adapted from Stalpaert et al. and Bradaric et al. [124c, 125] ...... 50 Scheme 19: Alcohol bromination using HBr-carrying zwitterions, adapted from Li et al.[133] R = H, alkyl, aryl, etc; n = 3,4...... 54 Scheme 20: The NADA concept: An HBr-triggered dehydration of LA to bio-AA. .. 73 Scheme 21: In-situ HBr formation within the [PBu4]Br reaction matrix...... 74 Scheme 22: Postulated mechanism for LA dehydration to AA triggered by HBr in the NADA molten salt reaction matrix...... 75 Scheme 23: Overview of the multi-step NADA process...... 80

Scheme 24: Observed reaction mechanism of lactide bromination in HBr(aq) at elevated temperatures...... 83 Scheme 25: Partial 1H NMR spectra of an aqueous bromination experiment with lactide (0.05 mol) and HBr(aq) (0.3 mol). 1H NMR of pure lactide (top, **traces of water) and 1H NMRs of samples taken at a reaction time of 0 (mid) and 5 h (bottom). Experiment conducted at 120 °C and ambient pressure. *DMSO-d6 as solvent...... 83

Scheme 28: Observed isomerization reaction of 2-BrPA to 3-BrPA in [PBu4]Br at elevated temperatures...... 121 Scheme 29: Base-assisted 3-BrPA dehydrobromination using TOA and related salt ([TOAH]Br) and HBr recycling as the third and last reaction step of a multi- step NADA sequence...... 136 Scheme 30: Postulated reaction network of 3-BrPA dehydrobromination in alkaline environment at elevated temperatures indicating formation of DiAA and AA oligomers from intermolecular Michael-type addition via AA ionization...... 139 Scheme 31: Overview of developed multi-step NADA process...... 144 Scheme 32: Comparison of observed 2- and 3-BrPA behavior in basic environment (TOA)...... 149

207

7.6 List of Tables

Table 1: Variation of the reaction temperature of 2-BrPA synthesis with HBr(aq).[a]... 89 Table 2: Results of ex-situ room temperature LLE screening.[a] ...... 95 Table 3: LA bromination using different brominating agents: comparison of aqueous

HBr (48 wt%) and different SO3H-functionalized bromide ionic liquids...... 106 Table 4: Results of ex-situ LLE solvent screening. [a] ...... 116 Table 5: Results of isomerization matrix screening.[a] ...... 129 Table 6: Determined quantities of semi-batchwise dehydrobromination of 3-BrPA at 150 °C and 100 mbar...... 140 Table 7: Results of HBr recovery study via thermally-induced HBr release and HBr extraction...... 141 Table 8: Comparison of one- and multi-step NADA process in terms of product yield and selectivity...... 145

Table 9: Product distribution of 2-BrPA conversion in [PBu4]Br at different process operations modes ...... 154

Table A 1: LA bromination with HBr(aq) using different LA:HBr molar ratios: Screening of reaction parameters[a]...... 172

Table A 2: Concentration-dependent Y2-BrPA, S2-BrPA and XLA. [a] ...... 173

Table A 3: Variation of the reaction temperature of 2-BrPA synthesis with HBr(aq).[a] ...... 173

Table A 4: Summary of LA bromination with HBr(aq) and [PBu4]Br at various LA:HBr:IL molar ratios. [a] ...... 174

Table A 5: Results of 2-BrPA isomerization experiments with [PBu4]Br...... 196 Table A 6: Results from 2-BrPA isomerization experiments with different bromide ILs at a IL:2-BrPA molar ratio of 1:1...... 197

Table A 7: Quantities of 2-BrPA conversion in [PBu4]Br at varying process operations...... 199