Solid-Phase Combinatorial Synthesis: Wittig Reaction for C=C Formation in Tamoxifen Derivatives David Yu-Chung Lee

Rochester Institute of Technology RIT Scholar Works

Theses Thesis/Dissertation Collections

2003 Solid-phase combinatorial synthesis: Wittig reaction for C=C formation in Tamoxifen derivatives David Yu-Chung Lee

Follow this and additional works at: http://scholarworks.rit.edu/theses

Recommended Citation Lee, David Yu-Chung, "Solid-phase combinatorial synthesis: Wittig reaction for C=C formation in Tamoxifen derivatives" (2003). Thesis. Rochester Institute of Technology. Accessed from

This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusion in Theses by an authorized administrator of RIT Scholar Works. For more information, please contact [email protected]. Solid-Phase Combinatorial Synthesis: Wittig Reaction for

C=C Formation in Tamoxifen Derivatives

David Yu-Chung Lee

June 2003

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in chemistry

Approved: ___K_a----'y~T_u_r_n_e_r __ Thesis Advisor

T. C. Morrill Department Head

Department of Chemistry Rochester Institute of Technology Rochester, NY 14623-5603 Copyright Release Fonn

Solid-Phase Combinatorial Synthesis: Wittig Reaction for

C=C Formation in Tamoxifen Derivatives

I, David Yu-Chung Lee, hereby grant permission to the Wallace Memorial Library, of RIT, to reproduce my thesis in whole or in part. Any use will not be for commercial use or profit.

Signature: ______Daniel Y. Lee _

Date: 'l/ { / ~.3 ; , Table of Contents

Abstract i

List of Tables ii

List of Figures hi

List of Schemes iv Chapter 1. Introduction 1 1.1 Tamoxifen 1 1.2 The Concept of Combinatorial Chemistry 6 1.3 Methods in Generating Combinatorial Libraries 7 1.4 Solid-Phase Synthesis 9 1.5 Resin: The Solid Support 10

Chapter 2. Purpose of Research 15

Chapter 3. Materials and Methods 21

3.1 Solid-Phase Synthesis of Stilbene Derivatives 21

3.1.1. Procedure No. 1 Solid-Phase Synthesis of Stilbene Derivatives Based on Procedures by Bernard and Ford (Scheme VII) and Argonaut (Table I, column 1) 24 3.1.2. Procedure No.2 Optimized Version of Procedure No. 1 24 3.1.3. Procedure No.3 Based on the Procedure Provided by NovaBiochem (Table I, column 2) 25 3.2 Synthesis of Tamoxifen Derivatives 25 3.2. 1 Procedure No. 4 Solution-Phase Triphenylethylene Synthesis 25 3.2.2 Procedure No. 5 Solid-Phase Synthesis of Tamoxifen Derivatives (Using Triphenylethylene Synthesis as an Example) 26 3.3 Product Analysis 27

Chapter 4. Results and Discussion 30 4.1 The Initial Stilbene Synthesis 30 4.2 Optimization: From Procedure No. 1 to No.2 32 4.3 The First Stilbene Library 35 4.3.1 Reactivities of the Benzyl Halides 35 4.3.1.1 Special Study for 4-Methoxy Benzyl Chloride 35 4.3.1.2 Electron Donating/Withdrawing Effects 36 4.3.2 The Effects of Functional Groups on Aldehydes 38 4.4 The Second Stilbene Library 40 4.4.1. Closer Examination ofArgonaut's Data 40 4.4.2. The NovaBiochem Resin 42 4.4.3. Data Interpretation 45 4.5 Liquid-Phase Synthesis of Triphenylethylene 45 4.6 Solid-Phase Synthesis ofTamoxifen Derivatives 49 Chapter 5. Conclusion and Suggestions for Future Study 53 Appendix Gas Chromatograms of the Second Stilbene Library 55 Abstract

Tamoxifen has a backbone structure of triphenylethylene and is a drug that has been used for over 20 years to treat patients with breast cancer. It has been discussed in the literature that the C=C formation would be the desired synthetic approach because of its flexibility in the introduction of suitable functional groups using readily available reagents. Solid-phase combinatorial approach is adopted for the purpose of product diversity and simplicity in product analysis. This research focused on optimization of each reaction step in the Wittig mechanism to overcome low yield caused by employing solid-state chemistry and the steric hindrance of C=C formation. It has been discovered that commercially available polystyrene-triphenylphosphine resins require extreme reaction conditions and may not be applicable for the pharmaceutical industry. List of Tables

Table I. Commercially Available Crosslinked Polystyrene-Triphenylphosphine Resins Used in This Research 23

Table II. GCMS Conditions 29

Table III. Adjustments from Procedure No. 1 to Procedure No.2 33

Table IV. Comparison of Triphenylphosphonium Salt Formations from Benzyl Halides

with Different Functional Groups 38 Table V. Results Published by Argonaut 42 List of Figures

Figure 1. A Simple Illustration of Parallel Synthesis 7

Figure 2. An example of generating 1,000 compounds from ten reagents and three

reaction steps using (a) parallel synthesis and (b) split-pool synthesis 8

Figure 3. General Procedure of Solid-Phase Synthesis. Reprinted from Hinks and Swali(2002) 10

Figure 4. Functionalization of a Crosslinked Polystyrene Resin 13

Figure 5. Yields of a Five-by-Five Library of Tamoxifen Derivatives Synthesized by Brown and Armstrong 18

Figure 6 . Structures of Reagents and Products in the Initial Stilbene Synthesis 31

Figure 7. Structures and Percent Yields of the First Stilbene Library 34

Figure 8. Formation of a Triphenylphosphonium Salt via Nucleophilic Substitution

Reaction between a Triphenylphosphine and a Benzyl Halide with a Functional Group 37 Figure 9. Mechanism of the Last Step in the Wittig Reaction, (a) Formation of an Oxaphosphetane. (b) Formation of a Betaine 40 Figure 10. Structures, Percent Yields, and cis : trans ratios of the Second Stilbene Library 44

Figure 11. Gas Chromatogram of Liquid-Phase Triphenylethylene Synthesis 47

Figure 12. Mass Spectra of (a) the Peak with Retention Time = 15.32 min from Figure 11 (b) Triphenylethylene (c) Triphenylphosphine 48 Figure 13. Summary of Four Solid-Phase Reactions for Tamoxifen Derivative Synthesis 49

Figure 14. Gas Chromatogram of Solid-Phase Synthesized Triphenylethylene 51

Figure 15. Mass Spectra of (a) Solid-Phase Synthesized Triphenylethylene and (b) Standard Triphenylethylene (from GCMS Internal Library) 52 List of Schemes

Scheme I. An Example of Non-Stereospecific Synthesis of Tamoxifen 4

Scheme II. An Example of Stereospecific Synthesis oftrans-Tamoxifen 5

Scheme III. Synthesis of Tamoxifen Derivatives via the McMurry Reaction 16

Scheme IV. Solid-Phase Synthesis of Tamoxifen Derivatives from Five Alkynes and Five

Aryl Halides via Resin Capture 16 Scheme V Proposed Solid-Phase Wittig Reaction for Tamoxifen Synthesis 18 Scheme VI. Schlosser Modification of the Wittig Reaction 20 Scheme VII. Bernard and Ford's Procedure for Solid-Phase Wittig Synthesis of Stilbene. 22 Chapter 1. Introduction

1.1 Tamoxifen

Tamoxifen (Nolvadex) is a drug that has been used for over 20 years to

treat patients with advanced breast cancer. It is used as adjuvant, or additional

cancer.1 therapy following primary treatment for early stage breast Tamoxifen

is classified as an antiestrogen. An antiestrogen or estrogen blocker works by

blocking estrogen in breast tissue. While estrogen may not actually cause

breast cancer, it may stimulate its growth, feeding the cancer. With estrogen

blocked, the cancer cells that need it may not grow at all. In other words,

antiestrogens may keep cancer from developing in the breast2.

Research has shown that when Tamoxifen is used as adjuvant therapy for

early stage breast cancer, it reduces the risk of recurrence of the original

cancer and also reduces the risk of developing new cancers in the other breast.

Based on these findings, the National Cancer Institute (NCI) funded a large

research study to determine the usefulness of Tamoxifen in preventing breast

disease.1 cancer in women who have an increased risk of developing the This

study, known as the Breast Cancer Prevention Trial (BCPT), involved 13,000

premenopausal and postmenopausal women at high risk of breast cancer.

Results of this trial were announced on April 6, 1998, and published in the

1 Tamoxifen: Questions and Answers. National Cancer Institute. http://cis.nci.nih.gOv/fact/7 16.htm (accessed April 2003) 2 NOLVADEX (tamoxifen citrate) Home Page, http://www.nolvadex.com/consumer/home.asp (accessed April 2003) Journal of the National Cancer Institute five months later on Sept. 16. In the

BCPT, half the women took Tamoxifen and half took a placebo (an inactive

pill that looked like Tamoxifen). Participants taking Tamoxifen also had fewer

fractures of the hip, wrist, and spine than women taking the placebo. However,

the drug increased the women's chances of developing four potentially

life-threatening health problems: endometrial cancer (cancer of the lining of

the uterus), deep vein thrombosis (blood clots in large veins), pulmonary

embolism (blood clot in the lung), and possibly stroke. The U.S. Food and

Drug Administration (FDA) approved the use of Tamoxifen to reduce the

incidence of breast cancer in women at increased risk of the disease in

October 1998.

/ /raws-Tarnoxifen

*This isomer is also

"Z" assigned in the \ / literature, which does sometimes cause confusion. Only cisltrans designations are used in this thesis.

Tamoxifen has a backbone structure of triphenylethylene (see above).

Only the trans isomer is an antiestrogen and active in the treatment of breast

Raloxifene Under Across North America. The Cancer Clinic Study of Tamoxifen and (STAR) Way Hereditary (accessed April at Duke Comprehensive Cancer Center, http://cancer.duke.edu/hcc/siar.nsp 2003) uses.4 cancer. Its cis isomer has no clinical Several papers have been

published regarding the synthesis of Tamoxifen5'6,7,89, from which two

exemplary synthesis, stereospecific and non-stereospecific, are shown in

I5 II6 Scheme and Scheme respectively.

Magdalen College Oxford. Wilson C Tamoxifen (Tamoxifen Citrate).

ar..Mk/mom/Tamoxifin/Tamoxifin.htm (accessed April httr//..'m, nhPm n 2003) Synthesis of the E and Z Isomers of the Antiestrogen Tamoxifen and Its Robertson D. W.; Katzenellenbogen, J. A. Tritium-Labeled Form. J. Org. Chem. 1982. 47. 2387-2393 Metabolite Hydroxytamoxifen, in Stereospecific Synthesis of (Z) Tamoxifen via Carbometalation ofAlkynylsilanes. Miller, R. B.; Al-Hassan, M. I. J. Ore. Chem. 1985, 50, 2121-2123 of Functionalized Ketones Low Valent Titanium (The McMurray Coe P L Scriven, C. E. Crossed Coupling by Tamoxifen. Chem. Soc. Perkin. Trans. I. 1986, 475-477 A New Stereoselective Synthesis of - 1,2,- Formation from the Dehydration of l-Cp-(Alkox\ phenyl) McCague R Stereoselective Olefin Tamoxifen. J. Chem. Soc. Perkin Trans. I. 1011-1015 diphenylb'utan-1-ols: Application to the Synthesis of 1987, Tamoxifen and 4-Hydroxytamoxifen Super-Base-Metalated Olier-Reuchet, C. et. al. Synthesis of Using 36(45), 8221-8224 Propylbenzene, Tetrahedron Lett. 1995, 5 3 K W M rt s

X E

s *- s >*> in u s '3 OI Q. 6 O 01 u OI -w

I C O Z

a E \ / CS w s

X u E a> z S CO u

U K U \ / ai oi od c a 'H o E

S 5!

GO u

01 _ OI -*- CO <

01 "S E OS X w B

^^ 1.2 The Concept of Combinatorial Chemistry

Combinatorial chemistry is a sub-field of chemistry that has undergone

rapid growth over that past decade. The goal of combinatorial synthesis is to

generate a large number of closely related molecules from a small number of

starting materials in combinations defined by a given reaction scheme. The

philosophy behind this new field of chemistry is that as the number and

diversity of the products increase, so does the probability of success. Nature

has already used a combinatorial approach to generate diverse functional

macromolecules, such as the large number of antibodies that recognize

molecules.10 non-self Combinatorial chemists applied this approach

evolutionarily to a wide array of topics, including the parallel synthesis of

individual organic compounds, the synthesis of complex mixtures of organic

compounds, the synthesis of the vast phage display and nucleic acid libraries,

and all the associated technologies for handling these libraries and obtaining

useful molecules from them.

10 Bunin, B. A. The Combinatorial Index; Academic Press: San Diego, 1998; Chapter 1. 11 Szostak, J. W. Introduction: Combinatorial Chemistry. Chem. Rev. 1997, 97 (2), 347-348 1.3 Methods in Generating Combinatorial Libraries

There are two common ways of generating combinatorial libraries. The

most straightforward approach is called Parallel Synthesis, which can be

simply explained by the following three-by-four matrix (Figure 1). In this

method every reaction takes place simultaneously in an individual reaction

vessel, therefore no further separation of the products is needed.

Ri R2 R3 R4

Ra Pal Pal Pa3 Pa4

Rb Pbl Pb2 PbS Pb4

Re Pel Pel PcS Pc4

Figure 1. A Simple Illustration of Parallel Synthesis. (R = Reactants; P = Products)

The other method is called Split-Pool Synthesis, which is commonly

used to synthesize a very large number of molecules. However, towards the

end of the synthesis there will be a significant amount (depending on the scale

of the library) of products in each reaction vessel, which makes the final

al.12 separation and purification much more complicated. Schreiber et have

demonstrated the power of this method by synthesizing a collection of 2.18

12 of over Two Million Compounds Structural Schreiber, S. L. et al. Stereoselective Synthesis Having Features Both Reminiscent ofNatural Products and Compatible with Miniaturized Cell-Based Assays. J. Am. Chem. Soc., 1998, 120, 8565-8566 million distinct compounds with only six reaction steps. A procedural

explanation of split-pool synthesis can be found in the article written by

al.n 14 Hruby et Figure 2 is an exemplary comparison between these two

methods.

100 100 100 100 100 100 100 100 100 100 Number of (a) ^ " " "' ^~ ^ wmm " mm *1 reactions

+ ABCDEFGH I J 1,000

10 each

1,000

F G H I J 1,000

Total number of reactions = 3,000

Number of (b) ABODE F G H I J_ reactions

10 each 10

Number of reactions = 30

compounds ten Figure 2. An example of generating 1,000 from parallel synthesis and reagents and three reaction steps using (a) (b) Copyright 1998 split-pool synthesis. Reprinted from Czarnik (1998). American Chemical Society.

13 of Synthetic Peptide for Identifying Ligand-Binding Activity. Nature 1991, Hruby V. J. et al. A New Type Library S54, 82-84 14 Chemistry. Anal. 70 (1 1), 379A-386A Czarnik, A. W. Combinatorial Chem., 1998, 1.4 Solid-Phase Synthesis

The great advantage of combinatorial chemistry is that it has the

potential to synthesize compounds faster than classical organic synthesis15.

Solid-phase synthesis is particularly suited in combinatorial chemistry

because there are only three routine procedures involved in all the reaction

steps: addition of reagents, filtration, and washing, which allows many simple

automated procedures to be developed. Also, only the final product of a

multi-step synthesis needs to be purified, and high concentrations of reagents

are often used to drive reactions to completion. All of these advantages can be

demonstrated by Figure 3: Solid-phase synthesis begins with the design of an

appropriate solid support, often a substituted polystyrene. This is covalently

attached to a linker molecule. The linker allows for attachment of the first

monomer to the solid support and subsequently facilitates the cleavage of the

final compound from it. At all other points during the synthesis the linker is

inert. Monomers are efficiently attached to the solid support by using large

excesses of reagents, which drive reactions to completion. Excess reagents are

"purification" removed by filtration and washing. This simple way of means

that automation is possible.

15 Hermkens, P. H. H.; Ottenjeijm, H. C. J; Rees, D. Solid-Phase Organic Reactions: A Review of the Recent Literature. 'Tetrahedron, 1996, 52 (13), 4527-4554

16 - Hinks, J.; Swali, V. Organic Practicals A Collaborative Affair. Educ. Chem. 2002, 39 (2), 50 -53 A B Linker Linker A + A A B + 4

0 Cleavage o k-t'r-C + C ^fr + ABC l 11

Key Molecule covalently linked to polymer support Polymer Support 9 which will allow for attachment of this monomer

. Monomers that are sequentially Collected in the filtrate filtration and g reacted to produce target during washing structures

1.5 Resin: The Solid Support

The most commonly used resin in solid-phase synthesis is hydrophobic

polystyrene, crosslinked with 1-2% divinylbenzene. The later polymer is the

preferred one for reactions at higher temperatures or for reactions with

reagents.15 organometallic Suspension polymerization is the conventional

polymers.17 method of preparing these crosslinked

The central element of utilizing these crosslinked polymers in organic

synthesis is the means of covalently attaching the initial building blocks

(reagents) to the polymer support. This is accomplished through a linker. The

linker consists of three parts: a resin-anchoring linkage, a building block

17 Balakrishnan T.; Ford, W. T Particle Size Control in Suspension Copolymerization of Styrene, Chloromethylstyrene, and Divinylbenzene. J. Appl. Polym. Sci. 1982, 27, 133-138

10 anchoring/cleavable functional group, and a structural template that

group.18 influences the stability of the functional The functionalization step

may occur either before or after the polymerization step (Figure 4). Many

linkers have been designed to accommodate different types of solid supports

and a variety of reaction conditions19. Loading, defined as the number of

reaction sites or functional groups per gram of resin, is typically in the range

mmole/g.20 of- 1-2 These polymer resins withstand a wide range of reaction

conditions, but some limitations can be observed15:

1) They are compatible with a wide range of polar and apolar solvents:

e. g. DMF, NMP, alcohols, THF, acetonitrile, DCM.

2) Prolonged use of mechanical stirring can cause mechanical damage of

the resin.

3) The temperature range for reactions described to date (1996) is

although most common used room about -78C to 155C, the is

temperature.

4) A broad range or reactants is compatible with the conventional resins

g. bases (e. g. employed, including acids (e. TFA, POCl3), DBU, RLi),

Lewis acids (e. g. BF3OEt2), reducing agents (e. g. LiAltL,, DIBALH,

18 for Solid-Phase Synthesis. In A Practical Guide to Combinatorial Kiely J. S. et al. Synthesis Tools Chemistry; Chemical Society: 99-122 Czarnik, A. W., DeWitt, S. H., Eds.; American Washington, DC, 1997; pp 19 Strategies in Solid-Phase Organic Synthesis and Guillier F. et al. Linkers and Cleavage Combinatorial Chemistry. Chem. Rev. 2000, 100, 2091-2157 20 Seneci P. Solid-Phase Synthesis and Combinatorial Techniques; John Wiley & Sons, Inc.: New York, 2000;

Chapter 1 .

11 NaCNBH3), oxidizing agents, homogeneous transition metal complexes

(e. g. Pd(OAc)2, Pd2dba3.CHCl3), and soluble salts (e. g. KOtBu).

Solid-support based reagents are not usually used due to the solid-solid

interactions. Sometimes a severely limiting factor in the choice of

reagents is the nature of the linker attaching the molecule to the

polymer.

5) The large interior surface area of the resins is easily accessible for

solvents and reagents. Extensive washing procedures are required in order

to remove excess reagents and high boiling solvents from these interior

spaces.

12 Radical Initiator

Styrene 1,4 divinylbenzene Functionalized Monomer Monomer

^l Reaction Site

Polystyrene

Figure 4. Functionalization of a Crosslinked Polystyrene Resin

As shown in Figure 4, in order to allow reactants to react with the reaction sites inside the resin, the resin has to provide enough room that is bigger than the sizes of reactants and solvents. This space provided is a crucial factor in solid-phase reaction kinetics, and it is heavily dependent upon the nature of solvents and percentage of crosslinking of the polymer support. The crosslinking of commonly used resins is between 1 and 2 %, which gives a reasonable compromise between a good ability to swell

(resulting from a low level of crosslinking) and the stability of the beads (low levels of crosslinking result in beads that are very fragile). Generally speaking, hydrophobic polymer (e. g. polystyrene) resins swell properly in aprotic

solvents (from 3 to 8 times of their original size) and poorly in polar protic

solvents such as water and alcohol20. One resin would swell differently in

different solvents, and the swelling abilities of many solvents have been

reported.21 examined and Thus, it is not surprising that the reaction rate in

solid phase reactions is generally slower than the traditional solution-phase

reactions. As mentioned above, one of the advantages of solid-phase synthesis

is that high concentrations of reagents can be used to drive reactions to

completion; but in each step of a multi-step reaction, excess reagents have to

be removed before the next step takes place. Although this removal is

accomplished by several repetitions of filtration and solvent washing, resins

that have swelled too much can make this procedure extremely

time-consuming. Therefore in selecting the solvent for each reaction step, it is

important to find the balance between good resin swelling ability and the time

needed for filtration.

21 Santini, R.; Griffith, M. C; Qi, M. A Measure of Solvents Effects on Swelling of Resins for Solid Phase Organic Synthesis. Tetrahedron Letters, 1998, 39, 8951-8954

14 Chapter 2. Purpose of Research

"In seeking a simpler synthesis to these interesting compounds (Tamoxifen derivatives) we looked for a method which would be flexible to allow the introduction of suitable functional groups using readily available reagents. Considerations of possible disconnections in the target molecule suggested the possibility of C=C formation as being a desirable step; three possibilities are, the Wittig reaction and its various modifications, the use of silicon-stabilized anion, and the McMurry reaction." Paul L. Coe and Clare E. Scriven

Based upon the availability of the starting materials, Coe and Scriven

chose the last method, with low valent titanium catalyst, to synthesize

Tamoxifen derivatives (Scheme III).

15 HO R=CH3 o = C1CH2CH2 = BrCH2CH2 = 0*30*4 RX = (CH3)2NCH2CH2 CX=F,C13r,D = H

RO OrO ROXX TiCl3/Li ^^"^^C

or TiCtyZn

Scheme III. Synthesis of Tamoxifen Derivatives via the McMurry Reaction.

Hi 6 x?

1-5" DMF 9-* P-H 80 C *2

a/A : R, = H, R2 = Me A-E b/B:R, =H, R2 = Et 1 : X = I. R3 = OCH2CH2N(Me)2 c/C:R1 = KH2-Pf 2 : X =Br, Rj = 0a^CH2N(C^)4 3 : X = = d/D : R| = OMe, R2 = Me Br, R3 CHjCH^fMefe 4:X = = = Br,R3 CH2N(*-Pr)2 e/E:R, OMe, R2=Et 5 : X = Br, R3 = N(Me)2 x^

2. 30%TFAInCH2aj

R3 1A-5E*

Scheme IV. Solid-Phase Synthesis of Tamoxifen Derivatives from Five Alkynes and Five Aryl Halides via Resin Capture.

16 In 1996 and 1997, Brown and Armstrong at U.C.L.A. published two

22 ' 23 papers regarding the synthesis of Tamoxifen derivatives via a solid-phase

combinatorial approach. Their synthetic method, platinum(O) catalyzed

al24 diboration of alkynes to alkenes, was developed by Ishiyama et (Scheme

IV). They stated, "Using a combinatorial approach, it may be possible to

disorders." develop drugs that are specific for a number of estrogen-mediated

The goal of the research in this thesis is to examine the feasibility of

using the Wittig reaction, one of the three possibilities suggested by Coe and

Scriven, for the C=C formation in Tamoxifen-like molecules. Parallel

Solid-phase combinatorial approach is adopted for the purpose of product

diversity and simplicity in product analysis. Scheme V shows the overall

reaction proposed for this research.

22 D.' Brown S. Armstrong, R. W. Synthesis of Tetrasubstituted Ethylenes on Solid Support via Resin Capture. J. Am. Chem. Soc. 1996, 118, 6331-6332 23 D.' Brown S Armstrong, R. W. Parallel Synthesis of Tamoxifen and Derivatives on Solid Support via Resin Capture'. J. Org. Chem. 1997, 62 (21), 7076-7077 24 A. Platinum(0)-Catalyzed Diboration ofAlkynes. (a) Ishiyama, T.; Matsuda, N.; Miyaura, N.; Suzuki, J. Am. Chem. Soc. 1993, 115, 11018. (b) Ishiyama, T.; Matsuda, N.; Miyaura, N.; Ozawa, R; Suzuki, A. Platinum(0)-CatalyzedPlatinum(0)-uataiyzea Diborationuiooiauuu ofuim^uraAlkynes withwiui Tetrakis(alkoxo)diborons:^uaMj.vaiJvuAu;. An Efficient and Convenient

ApproachAnnmar-h totr. rw-RisChorvDalkenes.cw-Bis(boryl)alkenes. Organometallics, 1996, 15, 713

17 r^N r^N

2. Base ^.^w,PhjP=0

^p Crosslinked Polystyrene

Scheme V. Proposed Solid-Phase Wittig Reaction for Tamoxifen Synthesis

tamoxifen major isomer )={ IB

Boronates C D

59% 57% 48% 54% 52%

55% 46% 45% 64% 68%

54% 59% 57% 64% 68% s 3

27% 13% 19% 19% 31%

59% 64% 54% 61% 44%

Figure 5. Yields of a Five-by-Five Library of Tamoxifen Derivatives Synthesized by Brown and Armstrong. Reasons that make this research challenging and interesting:

Brown and 1) Armstrong report yields ranging from 13% to 68% (Figure 5).

The technique they used is called resin capture25, in which the

intermediates are synthesized in solution, followed by trapping of the

intermediates onto the solid support, and final cleavage of products from

the resin. Our reaction scheme, on the other hand, involves solid-phase

reactions from the very first step. However, it is possible to obtain

reasonable yields through optimization, e.g. employing excess and

concentrated reagents, of each reaction step.

2) Scheme V possesses similar degree of steric hindrance to Scheme III, and

based on the high yields that Coe and Scriven report, stereochemistry is

not an issue needed to be tackled. This research would prove if this is the

case with the Wittig approach.

3) Supposing each reaction step is optimized and the issue of steric hindrance

has been solved, the Wittig reaction, theoretically, introduces a high degree

of stereoselectivity. It has been reported that using sodium amide or

sodium hexamethyldisilylamide as bases gives higher selectivity for

czs-alkenes than is obtained when ylides are prepared with alkyl lithium

reagents. Also, a procedure known as the Schlosser modification can

induce the reaction of unstabilized ylides with aldehydes to yield the more

25 Keating, T. A.; Armstrong, R. W. Postcondensation Modifications of Ugi Four-Component Condensation Products: 1-Isocyanocyclohexene as a Convertible Isocyanide. Mechanism of Conversion, Synthesis of Diverse Structures, and Demonstration of Resin Capture. J. Am. Chem. Soc. 1996, 118, 2574-2583

19 stable /r

halide complex and allowed to react with an aldehyde at low temperature,

at which fragmentation to an alkene and triphenylphosphine oxide does

not occur. Treating the complex with an equivalent of strong base such as

phenyllithium forms a /?-oxido ylide. Addition of /-butyl alcohol

protonates the /?-oxido ylide stereoselective^ to give the more stable

jy-betaine-lithium halide complex. Warming the solution causes the

complex to give the trans-alkene by a ^^-elimination (Scheme VI) .

* O" Li+ i H /'

RCH-CHR' RCH-CR' ,.Buqh X\r'

_p^" " " __~ "\ Li+0~ R H^ Li+CT P+Phj P*Ph3 r/\i K M P+Pha

4th 26 Organic Kluwer Academic/Plenum Carey R A.; Sundberg, R. J. Advanced Chemistry, Ed.; Publishers: New York, 2001; 111-119

20 Chapter 3. Materials and Methods

3.1 Solid-Phase Synthesis of Stilbene Derivatives

As mentioned, the challenge of this research is to overcome possible

low yields caused by (1) employing solid-state chemistry and (2) steric

hindrance of the C=C formation. This research began by optimizing each

step of the solid-phase Wittig mechanism with the synthesis of less sterically

hindered molecules: stilbene derivatives. The general method, adopted from

Bernard and Ford27, is illustrated in Scheme VII. Bernard and Ford

synthesized their own polystyrene-triphenylphosphine resin through

suspension polymerization and examined the yields in terms of degree of

crosslinking. This research, however, does not focus on resin synthesis.

Three commercially available crosslinked polystyrene-triphenylphosphine

resins that have been used in this research are listed in Table I.

27 Cross-Linked Polystyrenes. J. Org. Bernard, M.; Ford, W. T. Wittig Reagents Bound to Chem. 1983, 48, 326-332

21 o II CU X o 01 X a u

1 CO

+ OI

-** B

X CO X OX) o u "^ X BF u o II ^-v wQ 11 CU a) li, s w CL, 3 rj X! Q ~ o o CO i 1- & 01 s- 3 D OI CD + c o Oi s- H Cm CA (A >> ^-4 X o O CM to X 3 u o B 13 + u X c H u II 51 *] "O cu 3 la ou c JB X Vt 1 o CU - | CU ? CO I X I i o> s

u M co

22 Table I. Commercially Available Crosslinked Polystyrene-Triphenylphosphine Resins Used in This Research.

28 29 jn Manufacturer Argonaut NovaBiochem Aldrich Loading 1.0-1.8 1.0-1.5 ca 3 (mmole/g/) % Crosslinking 1 1

Suggested Wittig 1 equiv. (1.42 1 equiv. (1.12 N/A Reaction mmole/g, 3.0g) of mmole/g, lg) of resin Condition by resin, 6 equiv. of swollen in dry 10ml Manufacturer 1-iodobutane in 30ml DME, 5 equiv. of DMF at 65C for 48 benzyl bromide at hrs. Washing with 65C over night. DMF (4 x 40mL), Washing with THF, toluene (4 x 40mL), CH3CN, toluene, ether CH2C12 (4 x 40mL) and hexane and dried and diethyl ether (4 x in vacuo over night. 4 40mL) and dried in equiv. ofNaHMDS in vacuo for 12 h. The THF for lhr at room loading of the dried temperature under Ar. resin was re-examined Washing with THF (5 to be 1 mmole/g. 4 x 4ml ), followed by equiv. ofNaHMDS in 0.5 equiv. (limited) of 42ml THF for lhr at carbonyl compound in

room temperature. 4ml THF at room Washing with THF (5 temperature under Ar x 4ml ), followed by 1 over night. equiv. of carbonyl

compound in 60ml

THF at room temperature for 16 hrs.

28 PS-Triphenylphosphine. Argonaut Technologies. Detailed Overview of http://www.arp;otech.com/PDF/resins/ps triphenvl.pdf. (Accessed October 2002) and Methods. 29 NovaBiochem. Triphenylphosphine polystyrene: Applications http://www.novahiochem.corn/docs/PDS/01-64-0308-000.pdf (Accessed December 2002)

30 Sigma-Aldrich. Triphenylphosphine polymer bound.

http//wwwsigmaaldrich.corr^cm-bir^hsnir^istributecl/HahtShop/HAHTpage/frmCatalogSearchPost?Brand=

April AT DRlCH&ProdNo=366455. (Accessed 2003)

23 3.1.1. Procedure No.l Solid-Phase Synthesis of Stilbene

Ford27 Derivatives Based on Procedures by Bernard and Argonaut2* (Scheme VII) and (Table I, column 1)

resin .0 Argonaut (1 equiv., 1 g) with a loading of 1 .57 mmole/g was swollen in

25 ml DMF, and 5 equiv. of benzyl bromide (1 ml) was added drop wise at room

temperature. The mixture was stirred at 65C for 48 hrs, cooled, filtered, washed

sequentially with DMF (2 x 25 ml), CH3OH (2 x 25 ml), and THF (2 x 25 ml). The

polymer-bound phosphonium salt was swollen in 25 ml THF in and ice bath, and 1

equiv. of 25% NaOCH3/ CH3OH (0.36 ml) was added drop wise. The mixture was

stirred for 2 hrs at room temperature, filtered, washed with THF (2 x 25 ml), and

swollen in 25 ml THF. Benzaldehyde (2/3 equiv, 0.1 ml, limiting reagent) was

added drop wise, and the mixture was stirred at room temperature for 16 hrs. The

filtrate was collected and analyzed by GCMS.

3.1.2. Procedure No.2 Optimized Version of Procedure No. 1

The process of how this procedure is developed from Procedure No. 1

is discussed in detail in Section 4.2:

Argonaut resin (1 equiv., 1 .0 g) with a loading of 1 .57 mmole/g was swollen in

12 ml of DMF, and 15 equiv. of benzyl bromide was added drop wise at room

temperature. The mixture was stirred at 70C for 48 hrs, cooled, filtered, washed

CH3OH (2 x and THF (2 x ml). sequentially with DMF (2 x 30 ml), 30 ml), 30 The

polymer-bound phosphonium salt was swollen in 12 ml THF in an ice bath, and 4

equiv. of butyl lithium (3.9 ml) was added drop wise. The mixture was stirred under

nitrogen for 2 hrs at room temperature, filtered, washed with THF (2 x 30 ml), and

swollen in 12 ml THF. Benzaldehyde (10 equiv., 1.5 ml) was added drop wise, and

24 the mixture was stirred under nitrogen at room temperature for 16 hrs. The filtrate

was collected and analyzed by GCMS.

3.1.3. Procedure No.3 Based on the Procedure Provided by NovaBiochem29 (Table I, column 2)

A decision of switching resins was made after the first stilbene library

was synthesized with Argonaut's resin and Procedure No.2 (see section 4.3

for discussion):

NovaBiochem resin (1 equiv., 1.0 g) with a loading of 1.2 mmole/g was

swollen in 10 ml DMF, and 5 equiv. of benzyl bromide (0.71 ml) was added drop

wise at room temperature. The mixture was stirred at 65C for 48 hrs under

nitrogen, cooled, filtered, washed sequentially with DMF (2 x 25 ml), CH3OH (2 x

25 ml), and THF (2 x 25 ml). The polymer-bound phosphonium salt was swollen in

10ml THF in an ice bath, and 4 equiv. of BuLi (3 ml) was added drop wise. The

mixture was stirred under nitrogen for 2 hrs at room temperature, filtered, washed

with THF (2 x 25 ml), and swollen in 10ml THF Benzaldehyde (10 equiv., 1.2 ml)

was added drop wise, and the mixture was stirred under nitrogen at room

temperature for 16 hrs. The filtrate was collected and analyzed by GCMS.

3.2 Synthesis of Tamoxifen Derivatives

3.2.1 Procedure No. 4 Solution-Phase Triphenylethylene Synthesis

Triphenylphosphine (1 equiv., 0.01 mole) was dissolved in 25ml DMF, and 1

equiv. of bromodiphenylmethane (2.47 g) was added at room temperature. The

25 mixture was stirred at 65C until complete precipitation occurred. The mixture was cooled, filtered, washed with 25 ml DMF, and then added into 25 ml THF contained in an ice bath, and 1 equiv. of BuLi (6.25 ml) was added drop wise. The mixture was stirred for 2 hrs at room temperature, filtered, washed with 25 ml THF, and re-dissolved in 50-75 ml DMF. Benzaldehyde (1/2 equiv., 0.5 ml, limiting reagent) was added drop wise, and the mixture was stirred at 120C for 16 hrs.

Approximately 5 ml of the mixture was transferred into centrifuge tube and centrifuged. The supernatant liquid was collected and analyzed by GCMS.

3.2.2 Procedure No. 5 Solid-Phase Synthesis of Tamoxifen Derivatives (Using Triphenylethylene Synthesis as an Example)

Aldrich resin, (1 equiv, 0.2 g) with a loading of 3.0 mmole/g was swollen in minimum amount of DMF 1 5 equiv. of bromodiphenylmethane (2.2 g) was added at room temperature. The mixture was stirred at 90C for 72 hrs under nitrogen, cooled, filtered, washed sequentially with DMF (2 x 25 ml), CH3OH (2 x 25 ml), and

THF (2 x 25 ml). The polymer-bound phosphonium salt was swollen in minimum

amount of THF in an ice bath, and 4 equiv. of BuLi (1 .5 ml) was added drop wise.

The mixture was stirred under nitrogen for 2 hrs at room temperature, filtered, washed with THF (2 x 25 ml), and swollen in a minimum amount of DMF.

Benzaldehyde (15 equiv, 0.9 ml) was added drop wise, and the mixture was stirred under nitrogen at 70C for 48 hrs. The filtrate was collected and analyzed by GCMS.

26 3.3 Product Analysis

In the final step of all solid-phase procedures, the products cleave from

the resin and are collected in the filtrate. Only the byproduct,

triphenylphosphine oxide, remained linked to the solid support. The products

were analyzed both qualitatively and quantitatively by gas chromatography

mass spectroscopy (GCMS). The instrumental conditions are listed in Table II.

Each peak on the gas chromatogram was identified by its m/z value. The

library installed in the mass spectrometer was tested by commercially

available cis- and trans- stilbenes, and it failed to recognize the difference.

The two stilbene isomers were identified by the difference in their GC

retention time: trans- isomers have higher boiling point and should have

longer retention time. Percent yield of each product was calculated based on

the integration of its GC peak. Taking Procedure No.2 as an example: 10

equiv. of carbonyl compounds were used, therefore the theoretical yield

would be [(product peak area + carbonyl compound peak area) / 10].

In both solid-phase and solution-phase syntheses of Tamoxifen

derivatives, toms'-stilbene in the amount equal to theoretical yield was added

into each filtrate or solution mixture before injecting into GCMS. This new

procedure has made the percent yield calculation more accurate because

trans-stilbene has a structure more similar to the products and is more stable.

Further more, in the cases of using 10 or more equiv. of carbonyl compounds,

27 a 1 equiv. standard in the gas chromatogram has made calculation easier.

28 Table II. GCMS Conditions

GCMS Model Hewlett-Packard 6890 Capillary Column Model HP19091S-936 HP-1MS-Crosslinked Methyl Siloxane Length 60.0 m Diameter 250.00 pm Film Thickness 0.25 pm Mode Constant Pressure

Pressure 26.49 psi

Initial Flow 1.1 ml/min

Average Velocity 28 cm/sec Injector Injection Volume 0.5 uL

Syringe Size 10.0 uL Inlet Model Split Temperature 250 C

Pressure 26.49 psi Split Ratio 13.304: 1

Split Flow 15.0 ml/min

Total Flow 18.6 ml/min Gas Saver Helium

Saver Flow 15.0 ml/min

Saver Time 2.00 min Outlet Model Mass Spectrometric Detector Outlet Pressure Vacuum Oven Initial Temperature 140 C

Initial Time 2.00 min Rate 15 C/min Final Temperature 270 C

Equilibration Time 0.50 min Maximum Temperature 325C Program Tune File ATUNE.U Acquisition Mode Scan

Mass Solvent Delay 5.00 min Spectrometer Lowest Mass Detected 25 amu

Highest Mass Detected 550 amu Percent Report Sort by Retention Time Qualitative Report Library to Search NIST98.L

29 Chapter 4. Results and Discussion

4.1 The Initial Stilbene Synthesis

Six benzyl halides (benzyl bromide; 4-methoxy benzyl chloride; 4-nitro

benzyl bromide; 2-fluoro benzyl bromide; 2-methyl benzyl bromide) and five

aldehydes with one ketone (benzaldehyde; 3,4-dimethoxy benzaldehyde;

ortho-, meta-, para-, anisaldehyde; acetophenone) were chosen to synthesize

30 stilbene derivatives via Procedure No.l (section 3.1.1). Figure 6 illustrates

the structures of both reagents and products of the attempted stilbene library.

This was library expected to answer several preliminary questions, such as the

difference between electron withdrawing groups (-N02) and electron donating

groups' groups (-OCH3), effects of functional positions (ortho-, meta-, and

para-), and steric hindrance (column E and row 6).

The GCMS analysis of the first column showed all six products to be

stilbene (cis : trans = ~ 1 : 1) without any functional groups attached. A

repetition of this column, with GCMS analysis of filtrate after each step

indicated stilbene was formed after the base was added. This meant that the

benzyl groups from the halides were cleaved from the resin (somehow by

NaOCH3) and coupled with each other. The ylides were not formed, and

aldehydes and ketones did not react. Sodium methoxide (pKa = 15.5) was

replaced by a stronger base suggested by Argonaut: sodium

hexamethyldisilazane (NaHMDS; pKa = 26).

30 c

133

o c/3

23 >>

CJ .o

Ol J3

o o S- ft.

o exi

u s -

1/3

s E oi CO ^t- /-> ^O

31 4.2 Optimization: From Procedure No.l to No.2

The undesired coupling was eliminated after switching the bases, but no

product was formed. This indicated the necessity for further procedural

adjustments. All the adjustments leading to Procedure No.2 are listed in Table

III. Results of each adjustment and reasons for determining the next

adjustment are also listed. Because the resin's loading was only 1.0-1.8

mmole/g, it was not possible to see physical evidence for the completion of

each reaction step (except the orange/red color appearance after adding the

base), the entire process of the procedural adjustment was based on trial and

error. Considering the time needed to synthesize and completely evaluate a

parallel six-by-five library, the decision of reducing the library's size to a

five-by-four matrix was made after Adjustment No.2. Figure 7 illustrates the

structures and percent yields of the first stilbene library synthesized with

Procedure No.2. The percent yields reported include both isomers; the cis-,

trans- ratios in column B varied between 1:7 to 13:1, and most products in

columns C and D only had one peak on the chromatogram.

32 33 o

1) 1- 03

> "2 o

OI B w i>

2 ? ^- B it o la 41 o- o B 05 V5 ^S . OI - % o \X 1 9 S- e v\| E/3

OI S- 3 OX E c* m ITi

34 4.3 The First Stilbene Library

4.3.1 Reactivities of the Benzyl Halides

In Figure 7, the yields of each column may be interpreted as the

difference in reactivities of the benzyl halides:

WU >

o/ \_/ ./"\ /

. , ,, , ,, . , 2-ttftlrtHwiyi bromide , ,, , ,, , , 4-nttrobenzyl bromide Mioorohcn/u bromide 4-nwtfhcybeiyl chloride

Based on the significant difference in yields between 4-nitro benzyl bromide

and 4-methoxy benzyl chloride, two possible explanations were studied:

4.3.1.1 Special Study for 4-Methoxy Benzyl Chloride

The first possible explanation would be that bromide ions are better

leaving groups than chloride ions. Three one-step solution-phase reactions

were performed to examine the leaving abilities of the chloride ions and

verify the attack of triphenylphosphine via a SN2 or SN1 mechanism:

(1) 1 equiv. triphenylphosphine with 1 equiv. 4-methoxy benzyl chloride in DMFfor48hrsat65C.

(2) 1 equiv. triphenylphosphine with 1 equiv. 4-methoxy benzyl chloride in CH3OHfor48hrsat65C.

(3) 1 equiv. 4-methoxy benzyl chloride with 1 equiv. silver nitrate in ethanol until precipitation. Evaporate the filtrate on a rotary evaporator. Dissolve the residue in DMF, add 1. equiv. triphenylphosphine and stir for48hrsat65C.

35 The first reaction resembles Procedure No.2 in the solution phase, and no precipitate (triphenylphosphonium chloride) was formed. A more polar solvent, methanol, was used in the second reaction to hopefully increase the stability of the carbocation from the benzyl chloride. Again, no precipitate was formed. These results proved that 4-methoxybenzyl chloride does not

undergo SN2 mechanism with triphenylphosphine. The third reaction was performed trying to force the 4-methoxybenzyl chloride to undergo SN1 mechanism by first forming AgCl as precipitate then reacting

4-mthoxybenzyl nitrate with triphenylphosphine. Still, no precipitation was formed. This indicated 4-methoxybenzyl chloride does not undergo either

Sk2 or SnI in the Wittig reaction. The only solution to keep the methoxy benzyl group in the library was switching the reagent to 3-methoxybenzyl bromide.

4.3.1.2 Electronic Effects

The other possible explanation would be that the Wittig mechanism favors halides with electron withdrawing groups rather than donating ones.

This can be explicated by the mechanism of the triphenylphosphonium salt formation (Figure 8).

36 Figure 8. Formation of a Triphenylphosphonium Salt via Nucleophilic Substitution Reaction between a Triphenylphosphine and a Benzyl Halide with a Functional Group Rt.

If the functional group Ri in Figure 8 is an electron withdrawing group, it would help triphenylphosphine, the nucleophile, attack and stabilize the triphenylphosphonium salt; on the other hand, if Ri is an electron donating group, it would make this formation more difficult. To confirm this explanation, three liquid-phase reactions using different benzyl halides and triphenylphosphine in DMF were conducted. The time needed for

precipitation was recorded.

37 Table IV. Comparison of Triphenylphosphonium Salt Formations from Benzyl Halides with Different Functional Groups.

Time Needed for Triphenylphosphonium Benzyl Halide Salt Formation

3-methoxy benzyl bromide - 25 hrs at 65C

2-methyl benzyl bromide ~ 4 hrs 65C

4-nitro benzyl bromide Less than 5 minutes

4.3.2 The Effects of Functional Groups on Aldehydes

Comparing each row in Figure 7, the five aldehydes can be ranked in

terms of their relative yields:

benzylaldehyde o-aftisaldch>dc m-aiiisatdehydc p-anisaltteliyde

Similar to section 4.3.1.2, electron donating groups were again the cause of

lower yields. This phenomenon may be explained by the mechanism of the

When an ylide reacts with a last step in the Wittig reaction (Figure 9).

carbonyl compound, two possible products can be formed. One is a

which is the four-membered ring compound called an oxaphosphetane,

intermediate to generate an olefin and a triphenylphosphonium oxide (Figure

which can be observed under 9 (a) ). The other one is called a betaine, only

both the carbonyl carbon special (Figure 9 (b) ). For processes,

carbon of the phosphorus has to accept a pair of electrons from the a ylide,

38 group. The and this could be made easier if R2 was an electron withdrawing effects of R4 groups in Figure 9 have already been discussed in Section

4.3.1.2.

39 An Oxaphosphetane

A Betaine

Figure 9. Mechanism of the Last Step in the Wittig Reaction, (a) Formation of an Oxaphosphetane. (b) Formation of a Betaine

4.4 The Second Stilbene Library

4.4.1. Closer Examination of Argonaut's Data

Procedure No.2 (section 3.1.2) involved 15 equiv. of benzyl halide, 4

equiv. ofNaHMDS as base, and 10 equiv. of aldehydes. Another suggested

adjustment would make all the reagents even more concentrated. Five testing

reactions (Figure 7, column E) were performed according to Procedure No.2

with minimum amount (~5ml) of solvent in each step. No improvement in

yield was observed. An re-examination of the data published by Argonaut28,

40 in which they declare more than 85% yields for all products, two things came

into question:

The (1) procedure provided by Argonaut involved 1-iodobutane reacting with

their triphenylphosphine resin. According to our results for halide leaving

abilities, iodide ions would be better leaving groups than bromide ions.

1- Also, the position on butane was better suited for SN2 reactions than the

benzyl carbon with functional groups attached. Their products (Table V,

Entry 1 - 4), however, involved olefins synthesized from benzyl bromide

and carbonyl compounds in the absence or presence of electron

withdrawing functional groups. They replaced benzyl bromide in Entry 5

with 1-iodobutane to react with p-anisaldehyde.

(2) According to Argonaut's procedure (Table I, column 1), the loading of the

resin was re-measured after forming the triphenylphosphonium salt. They

reported only two-thirds of the original loading reacted successfully,

which may imply that the correct yields are only 2/3 of the ones they

reported.

Not being able to obtain Argonaut's procedure with benzyl bromide and the

correct report of their yields, the decision to switch resins was made.

41 Argonaut* Table V. Results Published by (Percent Yield Calculation was unclear according to the procedure provided by Argonaut)

Entry Phosphonium Carbonyl Olefin % Isolated Yield resin compound ' (t i v.l tansl

1 98% (3:1)

Cr^" 2 81% K (5:1)

ri a 9G% 3 ^. H (1:1)

T9 88% 4 j^Xj^wte (2:1)

0 OT~ 94% 5 (2:5)

Wl . 87% 6 MeY>iH (1:3)

OM9 owe

Ms 88% 7 c/* (2:1)

8 o^o 82%

4.4.2. The NovaBiochem Resin

A company, called NovaBiochem, claimed 72% yield of 2,4-dimethoxy

polystyrene-triphenylphosphine resin. Their stilbene using their method

(Table I, column 2), surprisingly, does not ask for conditions as concentrated

42 as reported in Procedure No.2. Procedure No.3 was developed using

NovaBiochem 's method, with which the second stilbene library was synthesized. Figure 10 shows the structures, percent yields, and cis : trans ratios of the Second Stilbene Library. Gas chromatograms are included in the

Appendix.

43 o 42

u >> c O

>~. u

it c

in a c o u ai C/5

03 1-

S S3

.to

a e

B 41 u - 41 CL,

T ^X 41 - % 3

N^N,! CZ3 ^ 4i - 3 6X1 n m IT)

44 4.4.3. Data Interpretation

Attention is focused on the yields of products A4, Bl, Cl and DI. The

unexpectedly high yields of A4 in comparison to A3 and A5, the greater than

100% yield of Bl, and the zero percent yields of Bl and Cl all indicated a

very poor loading homogeneity of the NovaBiochem resin. As mentioned,

loading is defined as the number of reaction sites or functional groups per

gram of resin. This is an average value, and apparently portions of the resin

were loaded unevenly. Generally speaking, higher yields were still obtained

with 4-nitro benzyl bromide, and for columns A, C, and D the yields were still

low but better than the previous library.

4.5 Liquid-Phase Synthesis of Triphenylethylene

Based on data of the first and second stilbene library, all challenges

being faced so far came from the nature of the resin in use. The solid-phase

procedure was optimized to a certain point that using reagents in greater

concentration would loose the practical meaning of this research. With

another commercially available resin (from Aldrich), it was decided to move

on to the synthesis of Tamoxifen derivatives. Before beginning the

solid-phase synthesis, one testing reaction in solution-phase was performed to

obtain product without the difficulties associated with the solid-phase

mechanism. Bromodiphenylmethane and benzaldehyde were chosen to

45 synthesize triphenylethylene with Procedure No.4, which involved one-to-one ratios of all reagents except the limited amount of aldehyde. The temperature of the last step was increased to 120C particularly for this sterically hindered

system (see section 3.2.1 for experimental detail).

The gas chromatogram is shown in Figure 11. A large amount (52.8% of total; solvent excluded) of triphenylphosphine oxide was observed (retention

= time 20.33 min). The other peak, with 15.32 min retention time and 13.1%

of total yield, consisted of both unreacted triphenylphosphine and the product,

triphenylethylene. Its mass spectrum is shown in Figure 12 with those of

triphenylphosphine and triphenylethylene from the GCMS library. Other

peaks in the gas chromatogram are unexpected compounds. This observation

along with the yields of the product and by-product (triphenylphosphine oxide)

indicated that the temperature increase was too much and has decomposed

some of the products. However, the significant quantity of triphenylphosphine

oxide indicated that this testing reaction was a success.

46 Abundance TIC: DYL 041A.D

1.7+07

Retention Time - 15.32 min; Triphenylphosphine, FW=262; 1.58+07 Triphenylethylene, FW=256 (13.1% of total) Retention Time = 20.33 min Triphenylphosphine Oxide, FW=278 (52.8% of total)

6000000

vi \'l n I'lyi'i'iTriyikiYpfJlvyAtriv^jvnt^i u i iti'i'i iti i rmtttt, > , , > ,, , i i i i i n m i i i i i i i i iL | | | | | | ri| | | | | rmie-> 6.00 7.00 6,00 9.00 10.00 11.00 12.00 13.00 14.00 15.00 16.00 17,00 18.00 19.00 20.00 21,00 22.00 23.00 24.00 25.00

Figure 11. Gas Chromatogram of Liquid-Phase Triphenylethylene Synthesis

47 ^bundanca #127727: Benzane. 1.l',1--(1-tin><-2-y*<>n)W- 2*6

eooo (b)

eooo

70O0

eooo

aooo

4000

3000

20OO

1O0O 113 | 152 i

~ '320^ '

-'- 20 40 ao ao ico is6 Viio ite^iMliaisL "306 340 300 380 42C

127828: Phophir>. tripnenyl- 2*2 (c)

133 I 170 I ,63 J. 91 202215228 244 , fc2j.Li ti ,1 I 1 1 ' I ' ' ' I ' " ' I ' ' " I ' ' ' ' I ' ' ' I '' ' I '"'' ' 1 " 140 180 180 200 220 240 260 280 300 320 340 380 380 rn/z~> 20 4Q 8000 60ao 1001 120 400 420

= Figure 12. Mass Spectra of (a) the Peak with Retention Time 15.32 min from Figure 11 (b) Triphenylethylene (c) Triphenylphosphine

48 4.6 Solid-Phase Synthesis of Tamoxifen Derivatives

First Result Second Result

-20% benzylaldehyde No Product Yield

No No p~anisakiehyde Product Product

BromodiphcDvlnKthane No No 4-nitro benzaldehyde ^ Product Product

No No \ Propiophenonc ^ Product Product

Figure 13. Summary of Four Solid-Phase Reactions for Tamoxifen Derivative Synthesis

As shown in Figure 13, four reactions were carried out in solid-phase

with Aldrich resin, employing bromodiphenylmethane with benzaldehyde,

3-methoxy benzaldehyde, 4-nitro benzaldehyde and propiophenone. The first

set of results was obtained by Procedure No.5 (section 3.2.2), which was a

slight modification of Procedure No.2: the quantity of carbonyl compound

was increased to 15 equiv. and the reaction time for the last step was

49 increased to 48 hrs; temperature was maintained at room, since increasing the temperature may disintegrate the products. No product formed. This directly indicated two possibilities:

(1) The Aldrich resin, like the two resins used previously, has some

unexpected difficulties.

(2) The steric hindrance was dominating the entire reaction, even over

the electron-withdrawing effect by the nitro- groups.

The second set of results was obtained after the four reaction mixtures

were put on a Labquake Rotator for three weeks at room temperature. About

20% yield of triphenylethylene was observed (see Figure 14 for gas chromatogram and Figure 1 5 for mass spectrum) in the first mixture

(bromodiphenylmethane and benzaldehyde). The presence of diphenylmethane remained uncertain, but it might come from the formation of triphenylphosphonium salt, since this hydrophobic compound is very stable inside the resin beads and difficult to be washed off. Based on the previous data that 4-nitro benzaldehyde would give higher yields than benzaldehyde in terms of electronic-withdrawing effects; this result further confirmed the dominance of steric effects.

50 TIC: DYL_043A.D 2800000

2700000

2600000

2500000

2400000

2300000

2200000

2100000 frans-Stilbene, FW=180,

2000000 normalized to be 100% yield

1900000

1800000

1700000

1600000

1500000

1400000

1 300000

1200000

1100000 Diphenylmethane

1000000

900000

800000 Triphenylethylene, FW=256 19.78% yield 700000

600000

500000

400000

300000

200000,

100000 L^ JL- 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00 15.00 16.00 17.00 16.00 19.00 20.00 21.00 22.01

Figure 14. Gas Chromatogram of Solid-Phase Synthesized Triphenylethylene

51 Abundance Scan 1702 (15.271 mm): DYL.043A.D 2$6 (a) 9000

8000

7000

6000

5000

4000

3000

2000

1000 !26 ^ 1 1 3i' 5,1 63 V 89 101 139 I 327 341 3S5 415 429 I' > 1 , 1 1 i^ .J. . ,-T, *v4Wfk 1 J.wTdffi|J,X, T"* I T I I | I I > y>z~> 20 40 60 80 100 120 140 160 180 2O0 220 240 260 280 300 320 340 360 380 400 420 Abundance #127727: Benzene, 1.1'.1"-(i-ethenyl-2-yll

9000

240 260 280 300 320 340 360 400 40 60 80 100 120 140 160 160 200 220 . ...420..

Figure 15. Mass Spectra of (a) Solid-Phase Synthesized Triphenylethylene and (b) Standard Triphenylethylene (from GCMS Internal Library)

52 Chapter 5. Conclusion and Suggestions for Future Study

As stated, the challenges faced in this research included experimental

difficulties associated with the solid phase mechanism and steric hindrance of

the forming C=C via the Wittig reaction. This research began with a belief

that through optimization of each reaction step these difficulties can be

overcome. The real obstacle however was the inconsistency among the

commercially available resins in terms of sensitivity to electronically different

functional groups and the loading of the triphenylphosphine reaction sites.

Based upon Bernard and Ford's conclusion about their self-made resins , the

solid-phase Wittig reaction is still believed to be an appropriate approach for

the synthesis of Tamoxifen derivatives:

"Polystyrenes cross-linked with as much as 20% divinylbenzene are suitable supports for Wittig reagents. Even large ketones such as 10-nonadecanone and cholestenone react with polymer-bound

methylene-triarylphosphorane in high yields. The success of the syntheses

depends upon preparation of the reagent via bromination and

phosphination of a cross-linked polystyrene, rather than by a copolymerization using /?-styryldiphenylphosphine, and upon generation

of the phosphorane with a base/solvent system that swells the

network." phosphonium sites in the polymer

Ford27 Margaret Bernard and Warren T.

In order to continue this research, it is necessary to synthesize resins with

desired degree of crosslinking and assured loading of reaction sites.

53 In addition to the resins, this research has also brought up several new questions with respect to the use of combinatorial chemistry. With the combinatorial approach, we have found that the solid-phase Wittig reactions favor electron-withdrawing groups on both halides and carbonyl compounds.

Tamoxifen, however, has an electron-donating group attached. In order to prepare Tamoxifen by the procedure developed in this research, it will be necessary to re-adjust our synthetic approach to functionalization after solid-phase synthesis of triphenylethylene. Controlling the positions of the functional groups will be an additional challenge.

The low yields in our studies were resulted from the possibility that triphenylphosphine becomes a poorer nucleophile when it is hooked to

polystyrene than when it is free in solution. Quantitative analysis of triphenylphosphonium salt formation in comparison to the initial loading of triphenylphosphine on the resin beads will be needed to examine this

possibility.

54 Appendix

Gas Chromatograms of the Second Stilbene Library

55 File C : Operator \HPCHEM\1\DATA\DYL_NBA1 Oavi d y . Lee Acqui red 4 Dec 2002 20:05 sing Acqwethod dyl Ins t rumen R.JTT chemi Al Sampl e Na m1sc info Vial Numb

DYL_NBA1

1 6007

1.9B-07 Benzaldehyde

V4S+07

t.3e-07

1 207

10*07

soooooo

aoooooo Product

7000000

soooooo FW=2lO

ooooooo 3.7% 4000000

3000000

2000000

lOOOOOO ZJ.

F-t 1 e C : \HPCHEM\1\DATA\DV _NBA2. D

Operator Davi d Y . Lee Acqui red 3 Dec 2002 14 : 27 using AcqMethod DYL Instrument RIT ChemH sampl e Nair A2 M 1 s c info vi al Numbe

TIC DYL_N8A2.D

4 Se'07

4.Z07

4a*D7 2,4-dimethoxy 3.8e-'07 benzaldehyde

3.6fl*07

3.20+07

30*07

?.Bo*07

2 Ba*07 Product

FW=270 2.2oQ7

20*07

8OT- 1

T 6*07

B'0~.

+07

8000000 A eoooooo r "^ 4000000

^dL 2000000 16 OO 10 OO e oo a.oo io oo 12.00 LOO 20.00 22.00 -2.A OO File : C operator : \HPCHEM\1\DATA\DYU_NBA3 . D oavi d y . i_ee Acqui red 3 9ec_2002 17 ; 21 instrument: : RIT Chemi using AcqMethod Oyl Samp 1 e Name : Misc info A3 vial Number:

AhMlffKf TIC: DYL_N8A3.D

4.a+07

3-Oo+OT

3.4-K>7

3.2+07

3*07 o-anisaldehyde

2.0e*07

2.6a*07

2.**OT

2.20+07

2+07

i.e+o7

1.6*07

1.40*07 Product

1.2+-07 FW= ie+07 240

aoooooo 1.3%

eoooooo

4000000

2000000

-.- -> 3. 1 OO 12.00 13.00 1-4.OO 1600

File C: \MPCHEM\1\DATA\DYL .NBA4. u

Operator Davl d y . Lee Acqui red A Dec 2002 9:56 us1 ng AcqMethod DYL Instrument RIT Chemi s. amp I e Name Misc Info A4 vial Number

TIC: DYL_NBA4.D

w-anisaldehyde

Product

1.1B*07 FW=240

54.3% ooooooo

eoooooo

7000000 A. (-. A

3000000

2000000

ID,. jL. OO 11.OO 12.00 13.00 14.00 1 S.OO 1 .OO 17.00 10.00 TO s.oa T.OO O.OO 9.00 a.oo A.oo a.oo Fi 1e operator Sii"dcv?M^iDAXANDV1NBAS Acqui red X7:24 Instrument 5-incj AcqMethod OYL s amp 1 e Name Mi sc info A5 vial Number

TIC: QYL_NBAS D

1 Bo+07

1 .70+07

1 So+07

1 - So+07

1 40*07

1 3*07 p-anisaldehyde

1 2a*07

1+07;

SOOOOOO

70OOO00 Product soooooo

soooooo FW=240

aoooooo 1.5% 3000000

2000000

L .^vl-^^JU tOO 5 OO BOO 7.00 BOO Q.OO 10.00 11 OO 1 Z.OO 13 OO 14.OO 1 S OO

File C: \HPCHEM\1\DATA\DYI NBBl. u Operator David Y. Lee Acqui red 3 Dec 2002 12:40 using AcqMethod oyl instrument RIT Cheml Sample Name Mi sc info Bl vial Number

TIC- OYL_NBB1 D

3o+07

4 8-07

4.00+07

4 2js+07

4o*07 Benzaldehyde 3 80+07

3 Oe+07

3.4+07

3 2o+07 3a*07 Product

FW=225

187% 2.2&+D7

2o+OT

1 80+07

1.60+07

1.4*07

1.20+07

10+0 7

SOOOOOO

soooooo

4OOOOOO

4 0 File : C:\HPCHEM\1\DATA\DY L_NBB2 .d Operator David Y. L*e Acqui red : 3 Dec 2002 X5 ; 38 instrument : rit chemi AcqMethod DYL sample Name: Misc info : vi al Number : 1 B2

HC: DYL_NB82 D 5o+07

4.Bo+07

4 eo+07

4 4e+07

4.20+07

40*07

3 0+07 3.80*07 2,4-dimethoxy 3,40+07

3.20+07 benzaldehyde

30+07

2.00+07

2 00+07

2.40+07 Product 2.20+07 20+07 FW=285 1.&0+07

i.eo+07 36.2%

1 .40+07 7^ 1.20+07

1o+07

SOOOOOO

soooooo

4000000

a.OO 20.00 22-OQ 24.OO 26 OO 2S.OO

File : c : \HPCHM\1\DATA\DYL,NBB3 .d

Operator Davi d Y . Lee Acquired : 4 Dec 2002 8:11 using AcqMethod DYL Instrument : RIT chemi B3 Santpl e Name : Misc info : Vial Number: 1

TIC: OYL_NBe3 O ''2'.3*at

2.20+07

2.10+07

2*r+07

1.9O+07

1.SO+07

1 .4o+07

1.30+07

o-anisaldehyde 1.20+07

1 1o*07 Product

1o*07 FW=255 9000000

SOOOOOO 64.9%

7000000

SOOOOOO

8000000

4000000

3000000

2000000

lOOOOOO 'l,r.L.,. *..... 12.00 13 OO acv 10-00 boo e.oo LOO Acquired : 4_Dec 2002 15:56 instrument : RIT chemi using AcqMethod dyl Sample Name: m1 sc info : B4 Vi al Number :

TIC DYL_NB64 D

2.30+07

2.2e+07

2 10-07

20+07

1 Qe+07 ...1

1 8o+07 j 1 7o+07 j 1.60+07 1

130+07

1 .40+07

1.30+07 w-anisaldehyde 1.20+07 Product

1.10+07 FW=255 10+07

SOOOOOO 82.1%

1 7000000 1

SOOOOOO

6000000

3000000

2000000

' 1000000 j

11 OO 12.OO 15 OO 16 OO

File : C: \HPCHEM\1\DATA\DYI NBB5.D Operator : Davi d y. Lee Acquired r 4 Dec 2002 IT': 49 using AcqMethod DYL Instrument : RIT Chemi Sampl e Name : B5 Misc info : Vial Number: 1

TIC: DYL.NBBD.D

2.3O+07

2-20+07

2.10+07

20+07

I.Oo+O?

1 .00+07

1.7*OT

1.6**07

LSo+07

1.40+07

Product 1.20+07

1 l+07 /?-anisaldehyde FW=255

10+07 82.1% SOOOOOO

sooooao

7000000

soooooo

5000000

4000000

3000000

2000000

lOOOOOO

1A.OO 11-OO 12.0O 13.00 9.00 10.00 4.00 O-OO O.00r 7 OO bbo File : C : \HPCHEM\1\DATA\DYL_NBC2 . D operator : David Y. Lee Acqui red : 3 Dec 2002 16 : 10 using instrument : RIT chemi AcqMethod dyl Sampl e Name : C2 Misc info Vial Number:

TIC: DYL_NBC2 D

4.4o+07[

4 2o+07j

4o*07.

3 00+07 j 2,4-dimethoxy 3 60+07 benzaldehyde

3.40+07

3.2+07

3o+07

2.HO + 07

2.80+07

2.40+07

2.2S+07

2e+07

1.8O+07

1 .80+07 Product

1 -4+07 FW=258 1 JZo+07

10+07 l.l%

SOOOOOO

soooooo'

4000000:

2000000I

I OO S.OO S.OO 7.00 S.OO S OO 1QOQ l-OO 12,00 13.DO... 14 OO

ft 1 e ; c : \hpchem\i\data\oyi nbc3 . o Operator : David Y. Lee Acquired : 4 Dec 2002 8:42 using AcqMethod Dyl instrument : RIT chemi C3 Sampl e Name ; Misc info Vial Number: 1

^t/jr-.'-j ro TIC: DYL_NBC3.D

1.SQ+07

1.80+07

1.70+07

1 .00+07

l-So+07

1 .40+07

o-anisaldehyde 1.3w+07

1.20+07

1-10+07

10+07

SOOOOOO

7000000 Product

6000000 FW= 228 SOOOOOO

4000000 4.9%

3000000

2000000

lOOOOOO

^JLil.. I OO 12 OO 13.00 14 OO 1S.OO File Operator - ac4. u Acqui red ; instrument : Sample Name : using a< qMethod oyl M-isc info vial Number -

nc- oy(._moc4.i: s 2.2o*07

2. 1o+07

1 .00+07

1.70+07

1.10+07

1o+07 Product

SOOOOOO

SOOOOOO FW=228 7000000 12.4% SOOOOOO

! SOOOOOO

I 4000000

3000000

\ soooooo

File operator DavDi5?5CCEM>iy>ATA\DY,-_NBC5 Acqui red 4 Dec 2 002 18:56 usi Instrument RIT Chemi ng AcqMethod dyl sample Name Misc mfo C5 Vial Number

TIC: DYL.NBC3 L>

2.4o+07

2.30+07

2J2o*07

2.lo+07

20+07

1.SoQ7

1.8

1.7b+07

1 60+07 /?-anisaldehyde

l.So+07

1 4o*07

i 1 .3o+07

1.20+07

1 . 1 0+07

1o+07

BOOOOOO

SOOOOOO

7QOOOOO

SOOOOOO

4000000

3000000

2000000 J^

lOODOOO

r.OO S OO 9 OO TO OO I .OO 1 2.00 1 3 OO : : \H File C PCHEM\l\OATA\DYI NBD2 . D Operator : David Y. Lee Acqui red i 3 Dec 2002 16:47 using AcqMethod DYL Instrument : RIT Chemi Sampl e Name : D2 Misc Info : Vial Number:

TIC; DYL_NBD2.D So+07

4.60+07

4.40+07

4.2+07

' 40+07

3.80+07 2,4-dimethoxy S.Oo+OT ' benzaldehyde 3.40+07

3 2e+OT ^/.

30+07

2.00+07 Product 2.40+07

2.2o+07 FW=254 2o+07

1.80+07 15.9%

1 .40+07 IS1/ 1 .20+07

10+07

SOOOOOO

4000000

> "i~ ?*-\>K> 2000000 u .. 18 OO 20 OO 22.OO 24 OO

File C:\HPCHEM\1\DATA\DYI Operator David Y. Lee Acqui red 4 Dec 2 002 9:2G usi ng AcqMethod dyl Instrument RIT Chemi D3 Sample Name Misc Info vial Number

TIC- OYL_NDD3 D

10+07

20+07

90+07

0O+O7

70+07;

+07J

So+07

o-anisaldehyde

+07

10+07 Product BOOOOOO

SOOOOOO FW=224

7000000 12.6% SOOOOOO

SOOOOOO

4000000

3000000

2000000

1000000

Aa .a: . . h , Fi le : C:\HPCHEM\I\DATA\DYL .NBD4.D Operator : Davi d Y. Lee Acquired : 4 Dec 2002 16:54 using AcqMethod DYL Instrument : RIT chemi sample Name: Misc Info : D4 vial Number: 1

T(C: DYL_NBD4,[ 2 20+07

* lo+Q7

20+07

i 90+07

' """

1.7o+07

1.6M+07

l.So+07

1 40+07 777-anisaldehyde

1 .30+07

1 .20+07

1.1a+07

Product

FW=224 7000000

6000000 14.2%

SOOOOOO

4000000

3000000

2000000 _A- ~\ 1000000

File C : \MPCHEM\1\DATA\DYL Operator David y. Lee Acqui red 4 Dec 2002 19:29 jsing AcqMethod dyl Instrument RIT chemi Sample Name D5 Misc info vial Number

TIC: DYL_NBD5 D

2.40+07

2.3o+l

2_2o+07

2.10+07

20+07

1.9o+OT

1 So+07

L70+07

1.60+07 p-anisaldehyde 1.So+07

1.40+07

1.30+07

1 .20+07

1 10+07 Product 10+07

9000000 FW=224

SOOOOOO

7000000 4.7%

SOOOOOO

5000000

4000000

3000000

2000000 A

1OO0OO0

.UUL 1 2 OO 1 3.00 OO , 1.4.