Compact, Cryogen-Free High-Resolution 60 MHz Permanent Magnet NMR Systems for Reaction Monitoring and Online/At-Line Process Control Observing 1H, 19F, and 31P

John C. Edwards1, David A. Foley2, Mark T. Zell2, Brian L. Marquez2, Tal Cohen3, Paul J. Giammatteo1

1. Process NMR Associates, LLC, 87A Sand Pit Rd, Danbury, CT 06810 USA, 2. Worldwide Global Research & Development, Center for Discovery & Development Sciences, Pfizer Inc, Groton, CT, USA

3. Aspect AI, Shoham, Israel

A compact high resolution NMR system will be described that can be situated on the bench-top or in the fume hood to be used as a continuous or stop-flow detector and/or an “in-situ” reaction monitoring system. The same system can be fully integrated into on-line shelters for on-line process control or utilized by engineers and technicians in an “at-line” environment. The system uses a unique 1.5 Tesla permanent magnet that can accommodate sample tube diameters of 3-10 mm with half-height spectral resolution (water resonance) approaching 1-3 Hz depending on the sample volume size and with excellent single pulse sensitivity. These systems can be utilized in a traditional NMR methodology approach or combined with chemometric approaches that allow NMR data to predict chemical and physical properties of materials via regression analyses that establish correlations between observed spectral variability and sample-to-sample property variance [1]. The systems utilized since the early 1990's are capable of single channel operation on higher sensitivity nuclei (1H, 19F, 31P, 23Na, 7Li, 11B). A new generation of NMR systems are now being manufactured featuring multi-channel operation giving the possibility to monitor two nuclei at once or to perform 1H-13C-DEPT and higher sensitivity approaches to 13C observation. In pharmaceutical applications the Aspect-AI 60 MHz system was utilized in a reaction monitoring scenario where a reaction was monitored simultaneously on a split sample loop by the 60 MHz NMR and a 400 MHz Bruker Avance III superconducting NMR spectrometer [2]. The results obtained on the two systems were virtually identical indicating that the 60 MHz NMR system can be used to transfer PAT knowledge generated on pharmaceutical reactions in the research lab to the manufacturing areas for production monitoring. 1) “Process NMR Spectroscopy: Technology and On-line Applications”, John C. Edwards, and Paul J. Giammatteo, in Process Analytical Technology: Spectroscopic Tools and Implementation Strategies for the Chemical and Pharmaceutical Industries, 2nd Ed., Editor Katherine Bakeev, Blackwell-Wiley, 2010. 2) "Application of a 60 MHz Permanent Magnet NMR System to Online NMR Reaction Development in the Pharmaceutical Industry", David A. Foley, Mark T. Zell, Brian L. Marquez, John C. Edwards, and Paul J. Giammatteo, Presented at PittCon 2013, Philadelphia, PA, March 21, 2013. Poster PDF available at www.process-nmr.com.

Abstract

Example Application: Steam Cracking Optimization Installed at BASF, Ludwigshaven 2000Cracker Facility Capacity: 600,000 Tonnes per YearControl Strategy: Feed Forward Detailed Hydrocarbon Analysis to SPYRO OptimizationNMR Analysis: 3-4 Minute Cycle (Single Stream)NMR PLS Outputs: Naphtha – Detailed Hydrocarbon PONA Analysis, DensityC4-C10 normal-paraffin, iso-paraffin, aromatics, naphthenes

1st Generation NMR AnalyzerInvensys –Foxboro - 1998-2003

2nd Generation NMR Analyzer Qualion Ltd. 2003-2011

Typical NMR Analyzer EnvironmentShelter House at Base of Vaccuum Tower

Sample system providing 2 conditioned streams to NMRAnalyzer

3rd Generation NMR AnalyzerModcon-Xentaur-Aspect AI

Actual Toluene (Wt%)

Pre

dic

ted

To

luen

e (W

t%)

( F9

C1

)

PLS Model Toluene Wt% by PIONA GC

1

23

45

6

7

8

9

10

11

12

1314

15

16

17

18

1920

21

22

2325

26

2728

2930

31

32

33

34

35

36

37

383940

4142

43

44

45

46

4950

51

53

5456

58

59

6263

66

68

69

70

71

73

75

76

77

78

79

80

82

83

84

85

86

87

88

90

91

929394

95

9697

98

99

101

102

103

104

105

106

108

109

110

111

112

113

114

115

116

117

118119120

121

122123

124

125126

127128

129130131132133134135136137

138139140141142143

144145146

147148149

150151152

156157158159160161

162163164

165166167168169170

171172173174175176

177178179

180181182

183184185186187188

189190191

192193194195196197

198199200201202203

204205206

207208209210211212

213214215

216217218219220221 222223224

225226227228229230

231232233

234235236

237238239240241242243244245246247248249250251

252253254

255256257258259260261262263

264265266

267268269270271272

273274275

279280281282

283284285

286287288

289290291

292293294

295296297298299300

301302303

304305306

307308309

310311312

313314315

319320321

322323324325326327

328329330

331332333334335336

337338339

340341342343344345

346347348349350351

352353354

355356357358359360

361362363364365366367368369

370371372

373374375

376377378379380381382383384

385386387

388389390391392393

394395396397398399400401402

403404405

406407408

409410411

412413414

427428429 430431432433434435

436437438

439440441

445446447

448449450

451452453454455456

457458459

460461462

463464

465466467

471472473

477478

481482

483484485 489490493494495496

-.5

1

2.5

4

5.5

0 1.5 3 4.5

Spectral Units ( )

Bet

a C

oef

fici

ent

( F9

C1

)

0

10 40 70 100 130

-1.5

1.5

10 40 70 100 130

Cyclopentane

Date

Wt% GC

NMR

0

2

4

6

8

10

12

14

16

1 147 293 439 585 731 877 1023 1169 1315 1461 1607 1753

iso-C5

iso-C6

iso-C7

iso-C8

iso-C9

Online Validation Process for PAT: 4 Month Comparison of Online NMR Prediction( (4 out of 33) and Laboratory GC Analysis

Predictive Vector for Toluene PLS Model

96 Hours of Online Variability Observed by NMR Analyzer for iso-paraffin components

0

0.5

1

1.5

2

2.5

3

3.5

4

1 167 333 499 665 831 997 1163 1329 1495 1661

Benzene

Toluene

Ethyl-Benzene

Xylenes

96 Hours of Online Variability Observed by NMR Analyzer for aromatic components

Examples of the expected resolution obtained on various essential oils at 60 MHz demonstrating the resolution that is obtained on a static 5mm NMR sample. Essential Oil analysis is currently being developed to allow identification and authentification on a compact 60 MHz NMR system that requires no specialized personnel or custom laboratory space.

60 MHz NMR Reaction 5mm NMR Tube

T-Butyl AlcoholReacting with Acetic AnhydrideIn the presence of dilute acid.

Sucrose Hydrolysis Kinetics

Sucrose a-glucose

b-glucose

2.5 2.0 1.5 ppm 1.0 ppm

A

B

CD

E

E

B

D

A

C

Esterification of t-BuOH

Superimposed Spectrum Plot

2.5 2.0 1.5 1.0 ppm

Esterification of t-BuOH

Integral GraphAnd Integration Plot

Acetyl Anhydride

Acetic Acid

T-BuOH

T-Bu-Ester

Ac-Ester

1-Propanol Esterification with Acetic AnhydridePure Solution – No Solvent –No Acid Catalyst - 8 Hour Reaction Profile

1,2

1

2

3

3

4

4

5

56

6

7

7

8

8,9

9

AcAn (7)

Esterification of 1-Propanol by Acetic AnhydrideIntegral Plots for Reaction Profile

Ac-EsterAcetic Acid

1-Pr-Ester

1-PrOH

1H NMR of a range of naphtha samples obtained every 3-4 minutes in a stop-flow situation – spectra variability of the integration binned and normalized spectra are regressed against primary property results obtained by GC-PIONA analysis

EQUIPMENT SET-UP AT PFIZER

PORTABLE ONLINE NMR MOBILE ANALYSIS – The size, mobility and standard utility requirements facilitates

analysis at the site of chemistry, at the fume-hood, kilo lab or manufacturing site.

REACTION MONITORING – Capable of monitoring multiple reaction components in

solution in real-time.

Reaction Processes were monitored simultaneously using an AI-60-RMS NMR Reaction

Monitoring System (60 MHz) from Cosa Xentaur and Bruker 400 MHz Avance III. NMR

spectra were recorded at regular intervals over the course of the reaction.

Experimental conditions:

Typical reaction concentration: 0.2 mol/L.

Flow rate: 4 mL/min

Transfer time from reaction vessel to detection: <1 min.

TRANSESTERIFICATION RESULTS

ACKNOWLEDGMENTS

The authors would like to thank Cosa Xentaur, and Aspect AI, for development and

provision of the 60 MHz online and laboratory NMR systems installed at PNA and Pfizer as

well as MestreLab Research for development of NMR reaction monitoring software.

More details of the compact, cryogen free 60 MHz NMR systems utilized in these studies

can be obtained at www.process-nmr.com

Online systems are marketed by Cosa-Xentaur and/or Modcon Syst

Laboratory Systems are marketed by Cosa-Xentaur and Aspect AI

Application support and development is performed by Process NMR Associates.

For further discussion contact :

John Edwards

Email: john@process-nmr.com

Tel: +1 (203)-744-5905

Skype: jcepna Twitter: jcepna

Web: www.process-nmr.com

Blog: www.nmblog.comREFERENCES

1.. D. A. Foley, M. T. Zell, B. L. Marquez, and A. Kaerner,

Pharm. Tech. S19-S21 (2011).

IMINE FORMATION

0

20

40

60

80

100

0 20 40 60 80 100 120

%

Time (min)

Isopropanol 400 MHz Ester 400 MHz

Isopropanol 60 MHz Ester 60 MHz

Aldehyde

Imine

CDI COUPLING

0

20

40

60

80

100

0 10 20 30 40 50 60

%

Time (min)

Benzaldehyde 400 MHz Imine 400 MHz

Benzaldeyde 60 MHz Imine 60 MHz

60 MHz NMR was used to monitor three reaction processes; imine

formation, CDI mediated amide coupling and transesterification.

Each reaction was monitored at regular intervals by both 400 and 60

MHz NMR and the data was overlaid to compare the profiles

obtained at the two different field strengths.

NMR data generated from these three reactions demonstrates the

application of low field NMR as a PAT tool for reaction monitoring.

It is planned to utilize chemometric analysis in the future to enhance

reaction profiling and investigate further complex reaction processes.

0

20

40

60

80

100

0 20 40 60 80 100 120

%

Time (min)

Acid 400 MHz CDI Intermediate 400 MHz Amide 400 MHz

Acid 60 MHz CDI Intermediate 60 MHz Amide 60 MHz

INTRODUCTION TO ONLINE NMROnline NMR is routinely employed as a reaction monitoring tool in the process development

area at Pfizer. Investigation of organic reaction processes by online NMR at 400 MHz

provides detailed process understanding for development chemists.

Here we outline the details of expansion of this reaction monitoring platform to include a 60

MHz portable NMR. The utilization of a compact and portable 60 MHz instrument provides

increased flexibility and cost benefits over traditional cryogenically-cooled super-conducting

magnets. These advantages allow the analysis to be performed at the location where the

chemistry is being conducted, rather than bringing the chemistry to the lab space specifically

designed for online NMR.

.WHAT INFORMATION CAN ONLINE NMR PROVIDE?

INFORMATION RICH DATA FROM A SINGLE EXPERIMENT – Online NMR is a powerful

analytical tool that enables a plethora of information to be gathered from a single

experiment. It provides a real-time, detailed picture of what is occurring in the process.

CONTINUOUS ONLINE SAMPLING1 – Stream of reaction mixture removed from vessel

and returned following analysis. No sample preparation or isolation allows detection of

labile species in solution.

PROTONATED SOLVENTS & REAGENTS – No necessity for expensive deuterated

solvents or isotopic enhancement.

REACTION KINETICS – Quantitative nature of the technique furnishes reliable kinetic data.

REACTION CHARACTERIZATION – Structural information of individual components in

the mixture is obtained at 400 MHz which aids assignment at 60 MHz.

REACTION OPTIMIZATION – Real-time analysis permits “on the fly” adjustments to

reaction conditions.

MECHANISTIC INSIGHT – Combination of reaction profile and intermediate identification

sheds new light on reaction mechanisms.

.

CH3

O O

O

CH3

+CH3

CH3

CH3

OH

CH3

CH3

CH3

O

OCH3

+ CH3

O

OH

Comparison of 60 and 300 MHz NMR data obtained onIbuprofen capsule content.

Invertase induced conversion of sucrose to fructose and glucose followed by 60 MHz 1H NMR.

CH3

O O

O

CH3

+ O

OCH3

CH3

+ CH3

O

OHCH3

OH

Table shows the chemometric modeling Results PLS regression for naphtha PIONA.

Beta

Coefficients

Spectral Units ( )

Be

ta C

oe

ffic

ien

t (

F9

C1

)

-

2

-

.5

1

2.

5

10 40 70 100 130

-

2

-

.5

1

2.5

10 40 70 100 130 Pre

dic

ted

Cyclo

he

xa

ne

( F

9 C

1 )

1

4

7

10

1 4 7 10 1 4 7 10

1

2345

6

78

910

11

12

1314

15

16

17

18

19

2021

22

23242526

2728

29

30

3132

33

34

35

3637

3839

4041

42

4346

47

48

49

5051

52

53

54

55

56

58

596067

68

69

7071

7273

74

75

76

77

7879

80

81

82

8485

8687

88

89

91

92

93

9495

9697

98

99

100

101

102103

104

105

106

108109

110111

112

113

115116

117118119120121

122123124125126127

128129130131132133

134135136

137138139

140141142

143144145146147148149150151152153154161162163164165166

167168169170171172173174175176177178

179180181182183184

185186187191192193194195196

197198199

200201202

203204205

206207208

209210211

212213214

215216217

218219220

221222223

224225226

227228229230231232

233234235236237238239240241

243244

245246247248249250251252253

254255256

257258259

260261262

263264265

266267268

269270271272

273274275

276277278

279280281

282283284

285286287288289290

291292293294295296

297298299

300301302

303304305

309310311

312313314315316317

318319320

321322323

324325326

327328329

330331332333334335

336337338339340341

342343

345346347348349350

351352353354355356357358359

360361362

363364365

366367368

369370371372373374375376377

378379380381382383

384385386

387388389390391392393394395396397398399400401

402403404

405406407408409410

414415416417418419420421422423424425429430431432433434438439440441442443

444445446447448449

453454455456457

458459460464465466

482483484485486487488489

1

4

7

10

1 4 7 10

Actual Cyclohexane (Wt%)

60 MHz

300 MHz

60 MHz

300 MHz

60 MHz

300 MHz

60 MHz

300 MHz

Polymer Application: Adhesive Prepolymers – Developing chemometric models relating 300 MHz calculated monomer concentrations to 60 MHz 1H NMR data.

-1.5-1.0-0.50.00.51.01.52.02.53.03.54.04.55.05.56.06.57.07.58.08.59.09.510.010.511.011.512.0f1 (p pm)

K ior-0 0 1-HB io- Oil V1 20 91 3- 04 KHDT Liq.Fr ac._DO_ Low Conv1H NMR in DM SOJCE -PNA-Merc3 0 0

60 MHz

300 MHz

Labile OH

Groupsand

Aldehydes

Water

andResidual

Alcohol/Ether

Aromatics

Olefin

alpha

Protons

Aliphatic

CH2/CH3

TMS

0.51.01.52.02.53.03.54.04.55.05.5f1 (ppm)

Omega-015-H

Omega-015-H

Genceutic Wild & Pure Omega-3 Fish Oil 1400mg

EPA=530mg DHA=280mg 1H NMR in CDCl3JCE-PNA-MVX300

5.32

NMR is ideally suited to petroleum and alternative energyapplications where water and black samples make NIR and Ramanprocess analytical approaches very difficult . The linear response of the NMR experiment, the spectral orthogonality of the chemistry Types, and the lack of response to sample color variability makeIt an excellent candidate for the development of robust and lowmaintenance chemometric calibrations.

Biomass Pyrolysis yields bio oils that are readily analyzed by 1HNMR. Here is a comparison of the quality of data obtained on60 and 300 MHz NMR system.

60 MHz 1H NMR of fish oils is being investigated to predict automatic omega-3 and omega-6 fatty acid concentrations as well as fatty acid distributions.

60 MHz NMR Reaction 5mm NMR Tube

Normal propyl alcoholreacting with Acetic Anhydridein the presence of dilute acid. Over course of 8 hours

1H 60 MHz NMR shows great promise to be utilized in USP standard methods developedfor ID and purity of small moleculeexcipients and API

19F NMR is straightforward on the1.5 Tesla NMR system and is usefulfor authentification tagging analysisand pharmaceutical drug analysis.

Fish and Edible Oils Analysis

Biomass to Fuels - Process ControlLeft: examples of NMR chemometric

prediction vectors obtained frompartial least-squares regression models of toluene and cyclohexane. Each spectrum was a 140 point binned and normalized spectrum with each bin representing 0.1 ppm. The prediction vector represents the 140 coefficients by which the spectrum is multiplied to obtain the wt% of the molecular component. Each molecular component has an individual PLS regression calibration developed based upon primary test data (GC) provided by the customer.

Right: Actual vs Predicted plots that demonstrate the correlation between1H NMR and GC-PIONA analysisfor cyclohexane and toluene content.

Predictive Vector for Cyclohexane PLS Model PLS Regression ModelCyclohexane Wt% by GC-PIONA

Naphtha is a light fraction of crude oil that is not utilized in fuel production due to pooroctane values. Many refiners with petrochemical complexes utilize this material as a feed to produce ethylene/propylene/butylene based on market economical factors. The cracking condition for naphtha is controlled by a SPYRO kinetic model that can predict the cracking product distribution based on a detailed GC PIONA analysis.