The synthesis of a novel palladium-phosphine catalyst and its Sonogashira cross-coupling activity. ________________________________________

Chemical catalysis was investigated through the synthesis of a palladium catalyst (Figure 1) containing the novel phosphine ligand 1-naphthyl(diphenyl)phosphine. A Sonogashira cross-coupling reaction catalyzed by the new palladium catalyst was carried out and compared to the results observed using a common catalyst, Palladium dichlorobis(triphenylphosphine). The results of relevant catalytic methods were characterized by analytical tools and a prospective mechanism of the observed cross-coupling suggested.

Figure 1: [PdCl2(PPh2Ar)2] (Ar = 1-naphthyl)
________________________________________

Catalysis is an increasingly integral part of synthetic organic chemistry due to its ability to facilitate such large improvements in reaction rate and selectivity control. The careful design of organometallic palladium compounds with incorporation of ligands of varying reactivity, lability and steric bulk allows the production of increasingly effective catalysts.

Phosphines such as triphenylphosphine are categorically important ligands in inorganic chemistry mainly due to applications in asymmetric synthesis, however another important utility is noticeable in palladium-catalysed coupling reactions. In 1975 a method appeared for preparing symmetrically substituted alkynes, afterwards named the Sonogashira cross-coupling reaction1. This method utilises a copper salt and palladium complex for catalysis to produce the alkyne under mild conditions relative to other methods (such as Heck Coupling2). The typical catalyst used in this process is [PdCl2(PPh2Ar)2](3) and the performance of this catalyst will be compared with that of the prepared catalyst pictured in Fig 1. The usefulness of the Sonogashira cross coupling reaction in carbon–carbon bond forming events has underpinned a vast array of applicability in the pharmaceutical industry and research synthesis.

Due to their usefulness, synthetic routes to produce tertiary phosphines are an important research area. Unfortunately trialkylphosphines are very sensitive to oxidation, so synthetic methods can be problematic and must be performed in air free conditions. There are a number of possible synthetic routes to tertiary phosphines with chiral elements, such as inclusion type crystalline complexation, metal complexation and chiral pool synthesis4. The following experimentation simply mono-substitutes the readily available triphenylphosphine with a 1-naphthyl group.

Another similar coupling reaction that forms carbon-carbon double bonds (instead of triple bonds in the Sonogashira reaction) between activated alkenes and aryl/vinyl halides is the Heck Reaction2. This coupling occurs in a basic environment and provides exceptional trans alkene selectivity at the β-hydride elimination step. A general mechanism is shown below.

Figure 2: Heck cross-coupling mechanism

In step A the organic halide oxidatively adds across the palladium(0) catalyst forming a trans Pd-complex that subsequently produces a pi-complex with the alkene in step B. From here insertion occurs, where the alkene bonds to the Pd, forming complex shown after process C. This undergoes β-hydride elimination (D) to afford the cross coupling product, subsequently reductively eliminating HX to regenerate the Palladium(0) catalyst that can re-enter the cycle.

Results & Discussion
Magnesium and 1-bromonaphthylene are refluxed in diethyl ether to produce a Grignard reagent that is combined with chlorodiphenylphosphine to produce 13.9461g (83%) of 1-naphthyl(diphenyl)phosphine (1) as buff, crystalline needles with a melting point of 104°C–106°C (lit: 124°C5).

Figure 3: Synthesis of 1-naphthyldiphenylphosphine
Table 1.1: 1-Naphthyldiphenylphosphine reactant table

Reactant/product | MW | Quantities used or obtained (g) | Moles used or obtained | Molar ratio(theoretical/actual) | 1-bromonaphthalene | 207.07 | 11.100 | 0.053605 | 1 / 1 | PPh2Cl | 220.64 | 13.520 | 0.061276 | 1 / 1.14 | (1) | 312.11 | 13.946 | 0.044683 | 1 / 0.83 |
The following spectral data tables (Tables 1.2-1.6) compare the experimentally obtained values of the 1-naphthyldiphenylphosphine produced with reference sources.

The 31P{1H} NMR peak at -14.20 (Table 1.3) directly correlates to the literature obtained value of -14.21. This additionally confirms that there is no starting material PPh2Cl residing in the product, as it would produce a peak at 81.56. As seen in table 1.4 the melting point obtained for product (1) is approximately 20°C lower than the accepted literature value. This is possibly due to impurities in the product. The observed infrared peaks are highly characteristic of the desired product and match up to the literature values. The 1432cm-1 peak correlates to the phosphorus-phenyl bond stretching. The presence of the aromatic rings is confirmed in the ring stretch frequency at 1588cm-1 and aromatic C-H stretching at 3054cm-1.

Table 1.2: 1H NMR (δH) | (1) x | Reference a | 8.34-8.31 (m, 1H) | 8.44-8.41 (t, 1H) | 7.84-7.66 (t, 2H) | 7.81-7.76 (t, 2H) | 7.67-7.55 (m, 2H) | 7.43-7.35 (m, 2 H) | 7.29-7.20 (m, 11H) | 7.32-7.24 (m, 11 H) | 6.97-6.89 (m, 1H) | 7.07-6.99 (m, 1 H) |

Table 1.3: 31P{1H} NMR (δH) | | Table 1.5: Infrared (cm-1) | (1) | Reference a | | (1) | Reference a | -14.20 | -14.21 | | 3378 | 3390 | | | | 3054 | 3053 | Table 1.4: Melting Pt. (°C) | | 1812 | 1711 | (1) | Reference a | | 1432 | 1435 | 104-106 | 124 | | 1177 | 1186 | | | | 1157 | 1118 | x 1H NMR recorded at 400MHz in CDCl3. a Obtained from Ref. 5, 1H NMR recorded at 400MHz in CDCl3, 31P{1H} NMR recorded at 162MHz in CDCl3

Table 1.6: Mass Spectrum | m/z | Fragment | 312.1 | C22H17P | 233.0 | C16H12P | 183.0 | C12H10P | 157.0 | C10H7P | 127.0 | C10H8 | 108.0 | C6H5P |

The key fragments recognised in the mass spectrum of (1) (noted in Table 1.6) are displayed in the Figure 4 scheme.

Figure 4: 1-Naphthyldiphenylphosphine mass fragmentation pattern

A solution of Li2[PdCl4] was prepared and reacted with 2 equivalents of the tertiary phosphine (1). This reaction (outlined in Figure 5) afforded 0.5731g (62%) of bis(naphthalen-1-yldiphenylphosphine) palladium chloride (2) as buff, crystalline needles with a melting point of over 275°C (lit: 290°C8).

Figure 5: Synthesis of [PdCl2(PPh2Ar)2]

Table 2.1: [PdCl2(PPh2Ar)2] reactant table

No experimental data has been reported for compound (2) in the literature, so spectral and experimental comparison has been drawn with the analogous compound bis(triphenylphosphine) palladium chloride. The following spectral tables (Tables 2.2-2.5) compare the experimentally obtained values of the [PdCl2(PPh2Ar)2] produced with literature findings for [PdCl2(PPh3)2].

Analysis of Table 2.3 and 2.4 show that both 31P{1H} NMR and melting point values are comparable with the literature values for [PdCl2(PPh3)2]. Examination of infrared data shows that absorbances around 1480cm-1 (ring stretch), 1095cm-1 and 690cm-1 are common in both spectra. A peak at 3053cm-1 signifies the aromatic C-H bonds, while absorbance at 1433 cm-1 suggests the P-Ph bond. The 1HNMR spectrum of the new palladium complex can be expected to be different to that observed for [PdCl2(PPh3)2] due to the larger number of dissimilar protons. This is seen through a larger number of shifts as noticed in Table 2.2, and it is noted that the number of protons observed correlates directly to the number of protons contained in the target molecule.

Table 2.2: 1H NMR (δH) | (2) x | Referenceb | 9.05-8.58 (dd, 1H) | 7.73 - 7.69 (m, 9H) | 8.06-8.04 (d, 1H) | 7.45 - 7.36 (m, 6H) | 7.95-7.91 (t, 2H) | - | 7.74-7.35 (m, 13H) | - |

Table 2.3: 31P{1H} NMR (δH) | | Table 2.5: Infrared (cm-1) | (2) | Referenceb | | (2) | Reference7 | 19.26 | 24.46 | | 3053 | - | | | | 1479 | 1729 | Table 2.4: Melting Pt. (°C) | | 1433 | 1480 | (2) | Reference8 | | 1096 | 1095 | >270 | 290 | | 691 | 690 | x 1H NMR recorded at 400MHz in CDCl3. b [PdCl2(PPh3)2] data obtained from Ref. 7, 1H NMR recorded at 400MHz in CDCl3, 31P{1H} NMR recorded at 162MHz in CDCl3

Reactant/ product | MW | Quantities used or obtained (g) | Moles used or obtained | Molar ratio(theoretical/actual) | (1) | 312.11 | 0.72000 | 0.0023069 | 2 / 2 | PdCl2 | 177.33 | 0.22260 | 0.0012553 | 1 / 1.09 | (2) | 802.00 | 0.57310 | 0.00071459 | 1 / 0.62 |
The Sonogashira cross-coupling of 4-iodonitrobenzene and 2-methylbut-3-yn-2-ol in triethylamine was performed as shown in Scheme 3a, catalysed by [PdCl2(PPh3)2]. This afforded 0.2266g (48%) of proposed 1-ethynyl-4-nitrobenzene (3a) as buff, crystalline needles with a melting point of 105-106°C (lit: 150°C -151°C9). The Sonogashira cross-coupling of 4-iodonitrobenzene and 2-methylbut-3-yn-2-ol in triethylamine was performed as shown in Scheme 3b, catalysed by [PdCl2(PPh2Ar)2]. This afforded 0.2768g (59%) of proposed 1-ethynyl-4-nitrobenzene (3b) as buff, crystalline needles with a melting point of 171°C-173°C (lit: 150°C -151°C9).

Figure 6: Synthesis of 1-ethynyl-4-nitrobenzene

Table 3.1a: 1-ethynyl-4-nitrobenzene (3a) reactant table Reactant/ product | MW | Quantities used or obtained (g) | Moles used or obtained | Molar ratio(theoretical/actual) | 4-iodonitro- benzene | 249.01 | 0.7941 | 0.0031890 | 1 / 1 | 2-methylbut-
3-yn-2-ol | 84.12 | 0.302 | 0.0035901 | 1 / 1.13 | 1-ethynyl-4- nitrobenzene (3a) | 147.13 | 0.2266 | 0.0015401 | 1 / 0.48 |

Table 3.1b: 1-ethynyl-4-nitrobenzene (3b) reactant table

Reactant/ product | MW | Quantities used or obtained (g) | Moles used or obtained | Molar ratio(theoretical/actual) | 4-iodonitro- benzene | 249.01 | 0.80000 | 0.0032127 | 1 / 1 | 2-methylbut-
3-yn-2-ol | 84.12 | 0.30200 | 0.0035901 | 1 / 1.12 | 1-ethynyl-
4-nitrobenzene
(3b) | 147.13 | 0.27680 | 0.0018813 | 1 / 0.59 |
The following spectral tables (Tables 3.2-3.4) compare the experimentally obtained values of the 1-ethynyl-4-nitrobenzene products (3a) and (3b) produced with reference sources.

Table 3.2: 1H NMR (δH) | (3a) | (3b) | Referencec | 8.20-8.16 (m, 2H) | 7.96-7.90 (m, 4H) | 8.19 (d, 2H) | 7.96-7.90 (m, 1H) | - | 7.63 (d, 2H) | 7.57-7.54 (m, 2H) | - | 3.36 (s, 1H) | 2.02 (s, 1H) | - | - | 1.55 (s, 6H) | - | - |
Initial inspection of the melting points of the products returns some telling information. The literature value for the expected product is 150°C-151°C. The observed products have melting points of 105°C-107°C for (3a) and 171°C-172°C for (3b). Literature search for the melting points of starting reagents and intermediate compounds reveals that the unprotected final product, 2-methyl-4-(4-nitrophenyl)but-3-yn-2-ol (compound (c) in Fig. 7) has a melting point of 100°C-102°C10, while the starting reagent 4-iodonitrobenzene has a melting point of 171°C-173°C11. From this we can hypothesize that product (3a) was not de-protected properly, and that product (3b) did not undergo cross coupling, resulting in reproduction of the starting material.

The 1HNMR spectrum for (3a) attests to the identification as the unprotected molecule, with indicative singlets at 1.55ppm and 2.02ppm showing the 6 methyl protons and hydroxyl proton respectively. The spectrum given by (3b) returns only one significant multiplet correlating to the 4 ring protons seen in 4-iodonitrobenzene.

Table 3.3: Melting Pt. (°C) | (3a) | (3b) | Reference9 | 105-107 | 171-172 | 150-151 |

Table 3.4: Infrared (cm-1) | (3a) | (3b) | Reference9 | 3163 | 3088 | 3252 | 2227 | - | 2102 | 1512 | 1504 | 1513 | 1340 | 1350 | 1344 | x 1H NMR recorded at 400MHz in CDCl3. c Obtained from Ref. 9, 1H NMR recorded at 500MHz in CDCl3.

The 1500cm-1 and 1300cm-1 peaks in the two synthesized products and reference product IR spectra are representative of the nitro-group anti-symmetric stretch and symmetric stretch respectively. Examination of the reference spectrum shows that a weak absorbance at ~2100cm-1 signifies the mono-substituted alkyne stretching frequency present in the desired structure. This peak is seen at 2227cm-1 in the (3a) spectrum, this red-shifted absorbance is indicative of the di-substituted alkyne found in 2-methyl-4-(4-nitrophenyl)but-3-yn-2-ol. Additionally, a broad peak in the 3300cm-1 region of the (3a) spectrum signifies the presence of an OH group, confirming that the product is 2-methyl-4-(4-nitrophenyl)but-3-yn-2-ol.

Thus, Scheme 3a in fact afforded 0.2266g (35%) of 2-methyl-4-(4-nitrophenyl)but-3-yn-2-ol (3a) as buff powder with a melting point of 105°C-106°C (lit: 100°C-102°C10). While scheme 3b afforded 0.2768g (35%) of 4-iodonitrobenzene (3b) as buff, crystalline needles with a melting point of 171°C-173°C (lit: 171°C-173°C11). Updated reactant tables are shown in Table 3.5a and 3.5b.
Table 3.5a: 2-methyl-4-(4-nitrophenyl) but-3-yn-2-ol (3a) reactant table Reactant/ product | MW | Quantities used or obtained (g) | Moles used or obtained | Molar ratio(theoretical/actual) | 4-iodonitro- benzene | 249.01 | 0.7941 | 0.0031890 | 1 / 1 | 2-methylbut-
3-yn-2-ol | 84.12 | 0.302 | 0.0035901 | 1 / 1.13 | 2-methyl-4-
(4-nitrophen
yl)but-3-yn-
2-ol (3a) | 205.21 | 0.2266 | 0.0011042 | 1 / 0.35 |

Reactant/ product | MW | Quantities used or obtained (g) | Moles used or obtained | Molar ratio(theoretical/actual) | 4-iodonitro- benzene | 249.01 | 0.80000 | 0.0032127 | 1 / 1 | 2-methylbut-
3-yn-2-ol | 84.12 | 0.30200 | 0.0035901 | 1 / 1.12 | 4-iodonitro- benzene (3b) | 249.01 | 0.27680 | 0.0011116 | 1 / 0.35 | Table 3.5b: 4-iodonitrobenzene (3b) reactant table

The lack of de-protection evident in the synthesis of product (3a) is likely due to insufficient reflux time after addition of KOH. Laboratory time restrictions did not permit the ideal time of 30 minutes reflux, so less time was used, resulting in the subsequent crystallisation of the protected product.

No experimental flaws were noted after inspection of the experimental process of scheme 2 (Fig. 5), so the lack of cross coupling using product (2) as catalyst may be put down to the fact that the novel molecule simple does not catalyse the Sonogashira reaction.

The originally hypothesized reaction (Fig 7.) commences with the reduction of the palladium catalyst to create a Pd0 species that can oxidatively add to reactant a. The copper salt undergoes its catalytic cycle that results in the formation of the copper-acetylene complex from triethylamine-facilitated addition of reactant b to CuI. This copper-acetylene complex undergoes transmetallation with the palladium complex, regenerating the copper catalyst. Reductive elimination then affords product c (the precursor to 1-ethynyl-4-nitrobenzene) and regenerates the Pd0 species that can re-enter the cycle. It is worthy to note that transmetallation will in fact produce the trans isomer, and thus cis-trans isomerisation must occur before reductive elimination.

Figure 7: Sonogashira cross-coupling mechanism

Conclusions
It can be seen through analysis of products that the new catalyst (2) was insufficient in catalysing the Sonogashira cross-coupling reaction, as the product afforded was the starting reagent 4-iodonitrobenzene. The process using the common [PdCl2(PPh3)2] catalyst resulted in correct cross-coupling, but insufficient de-protection procedure resulted in the production of the protected molecule instead of the desired 1-ethynyl-4—nitrobenzene. Thus it can be concluded that [PdCl2(PPh3)2] is a much more effective Sonogashira catalyst than [PdCl2(PPh2Ar)2] as the latter does not catalyse the reaction.

The reported experimentation provides insight into the role of tertiary phosphines in catalysis and the synthetic routes available to chemists to produce phosphines and related palladium catalysts. The continued development of catalytic methods such as those that drive Sonogashira cross-couplings will be integral to the future of synthetic organic chemistry.

Experimental Procedure
Compound (1): 1-naphthyl(diphenyl)phosphine
Magnesium turnings (1.3450 g, 55 mmol) and a magnetic stirrer bar are placed in a dry 250mL three-necked round bottom flask and the system degassed.
Dry diethyl ether (20 mL) is added and 1-bromonaphthylene (7.5mL, 11.1, 54mmol) is added dropwise while heating and stirring to initiate reaction. The remaining 1-bromonaphthylene is added to the reaction vessel at such a rate so as to maintain a gentle reflux of the solution whilst also adding more diethyl ether (100mL). The solution is refluxed for 15 minutes using a warm water bath, cooled, and chlorodiphenylphosphine (11mL, 13.2 g, 60mmol) added. Solution is cooled and a saturated solution of ammonium chloride (20 mL) followed by degassed water (30 mL) added. The aqueous phase is removed from the reaction vessel, and the organic phase dried with MgSO4.The dried organic phase is added to a degassed 250 mL two-necked round bottom flask, and the solvent removed with a solvent trap. The crude tertiary phosphine is recrystallised from hot ethanol, washed with cold ethanol (5 mL) and dried in air producing 13.9461g (83%) of 1-naphthyl(diphenyl)phosphine as yellow crystalline needles. mp: 104°C–106°C (lit5: 124°C); 1H NMR (400MHz, CDCl3) δ 8.34-8.31 (t, 1H), 7.84–7.66 (t, 2H), 7.67–7.55 (m, 2H), 7.29–7.20 (m, 11H), 6.97–6.89 (m, 1H); 31P{1H} NMR δ 14.20; IR 3378cm-1, 3054cm-1, 1812cm-1, 1432cm-1, 1177cm-1, 1157cm-1.

Compound (2): dichlorobis(naphthalen-1-yldiphenylphosphine) Palladium
A solution of Li2[PdCl4] in methanol is prepared by reacting PdCl2 (0.20g) with excess anhydrous LiCl (0.5 g) in methanol (30 mL). The mixture was stirred for 1 hour. The solution was filtered and a solution of (1) (0.72g, 2.3 mmol) in methanol (10 mL) added to the red-brown filtrate. The solution was stirred for 15 minutes and cooled on and ice bath. The product was washed with diethyl ether and dried by suction, affording 0.5731g (62%) of (2) as an off-white, crystalline solid. mp: >275°C (lit8 290°C); 1H NMR (400MHz, CDCl3) δ 9.05-8.58 (dd, 1H), 8.06-8.04 (d, 1H), 7.95-7.91 (t, 2H), 7.74-7.35 (m, 13H); 31P{1H} NMR δ 19.26; IR 3053cm-1, 1479cm-1, 1096cm-1, 691cm-1.

Compound (3a): 2-methyl-4-(4-nitrophenyl)but-3-yn-2-ol
Set up a reaction vessel consisting of a 250 mL two-neck round-bottom flask containing an adaptor-with-a-tap, a glass stopper and a stirrer bar. Attach a nitrogen line to the adaptor with- a-tap and evacuate the flask. Allow nitrogen back into the flask. Triethylamine (25 mL) and 4-iodonitrobenzene (0.8 g, 3.2 mmol) are added to a degassed 250 mL round-bottom flask stirred. The solution is degassed 3 times by subjecting to vacuum for 20 seconds and back filling with N2. [PdCl2(PPh3)2] (0.003g, 0.004mmol) and CuI (0.005 g, 0.026 mmol) are added and the mixture degassed. 2-methylbut-3-yn-2-ol (0.35 mL, 3.3 mmol) is added and the reaction mixture stirred for 2 hours.
The solution is filtered and the solid residue washed with toluene (5 mL). The solid residue is discarded, the solvent removed from filtrate on a rotary evaporator. The resulting solid is re-dissolved in toluene (5 mL), filtered a pad of silica (230-400 mesh) (~2 cm) and eluted with toluene/DCM (5:1, 100 mL). The solvent is removed on a rotary evaporator and the solid re-dissolved in toluene (25 mL) 0.76g (13.4 mmol) of finely ground KOH is added to the solution and refluxed at 80°C for 15 minutes. The solution is cooled, filtered and the solvent removed from the filtrate. The crude 1-ethynyl-4-nitrobenzene is recrystallized in hot cyclohexane. The product was collected, washed with cold cyclohexane and dried by suction, affording 0.2266g (35%) of 2-methyl-4-(4-nitrophenyl)but-3-yn-2-ol (3a) as a buff powder. mp: 105°C –106°C (lit10: 100°C–102°C ); 1H NMR (400MHz, CDCl3) δ 8.20-8.16 (m, 2H), 7.96-7.90 (m, 1H), 7.57-7.54 (m, 2H), 2.02 (s, 1H), 1.55 (s, 6H); IR 3163cm-1, 2227cm-1, 1512cm-1, 1340cm-1.

Compound (3b): 1-ethynyl-4-nitrobenzene
Synthesis of 3b followed the same method as 3a, using the palladium complex (2) (0.0035g, 0.004 mmol) instead of [PdCl2(PPh3)2] and refluxing at 80°C for 30 minutes instead of 15. This produced 0.2768g (35%) of 4-iodonitrobenzene (3b) as a yellow, crystalline solid. mp: 171°C–173°C (lit11: 171°C–173°C ); 1H NMR (400MHz, CDCl3) δ 7.96-7.90 (m, 4H); IR 3088cm-1, 1504cm-1, 1350cm-1.

References
1. Sonogashira, K., "Development of Pd-Cu catalyzed cross-coupling of terminal acetylenes with sp2-carbon halides" J. Organomet. Chem., 2002, 653, 46.

2. Heck, R. F.; Nolley, Jr., J. P. "Palladium-catalyzed vinylic hydrogen substitution reactions with aryl, benzyl, and styryl halides". J. Org. Chem. 1972, 37(14) 2320–2322.

3. Valentine, D., Hillhouse, J., “Electron-Rich Phosphines in Organic Synthesis II. Catalytic Applications”, Synthesis, 2003, 16: 2437

4. Salem, G, “CHEM3206 Catalysis in Chemistry – Transition Metal-Based Asymmetric Catalysis”, Australian National University, 2014,

5. Sun, Meng, “Nickel-Catalyzed CP Cross-Coupling by C-CN Bond Cleavage”, Chemistry, 2011, 17(35) 9566-70

6. O. Kühl, "Phosphorus-31 NMR Spectroscopy", Springer, Berlin, 2008, 10.

7. Pfeiffer, Hendrik, “Sonogashira and "Click" reactions for the N-terminal and side chain functionalization of peptides with [Mn(tpm)(CO)3]+-based CO releasing molecules (tpm = tris(pyrazolyl)methane)”, Royal Soc. Chem., 2009, Supplementary Information

8. Pandey, R, “Spectral studies on hydridophosphine complexes of Pt-group metals”, Journal of Ultra Chemistry, 2009, 5(3) 327-332

9. Yuan, Wang Zhang, “Direct Polymerization of Highly Polar Acetylene Derivatives and Facile Fabrication of Nanoparticle-Decorated Carbon Nanotubes”, Macromolecules, 2009, 42(1) 52–61

10. Bo-Nan, L, “Sonogashira Reaction of Aryl and Heteroaryl Halides with Terminal Alkynes Catalyzed by a Highly Efficient and Recyclable Nanosized MCM-41 Anchored Palladium Bipyridyl Complex”, Molecules, 2010, 15(12), 9157-9173

11. Yungfang, L, “Reduction of nitroarenes to azoxybenzenes by potassium borohydride in water”, Molecules, 2011, 16 3563-3568

Supporting Information
For risk assessment sheets and annotated spectra for compounds 1, 2, 3a and 3b see ‘Supporting Data’ document.