Crystal structures of five halido gold complexes involving piperidine or pyrrolidine as ligands or (protonated) as cations

The structures of five gold complexes involving piperidine or pyrrolidine, either as neutral ligands or protonated counter-cations, are presented.


Chemical context
According to the well-known classification of metal ions and ligands introduced by Pearson (1963), gold(I), and to a lesser extent gold(III), are regarded as archetypal 'soft' metal centres and, as such, would be expected to form stable complexes with soft ligands (typically with donor atoms such as sulfur and phosphorus) rather than hard ligands (with e.g.nitrogen or oxygen donors).As a general trend this is true, but even gold(I) nevertheless forms a wide variety of complexes with nitrogen ligands such as amines (in which we include azaaromatics).We have studied these extensively, particularly with regard to their structural aspects, in the series of papers 'Gold complexes with amine ligands', of which this forms the latest part; part 1 appeared in 1997 (Jones & Ahrens, 1997) and the previous part (part 11) in 2018 (Do ¨ring & Jones, 2018b).However, our first (unnumbered) paper on the subject appeared almost half a century ago (Guy et al., 1977), and concerned complexes of the cyclic secondary amine piperidine (henceforth 'pip' in formulae).Crystals of [AuCl(pip)] were obtained in small quantities in an attempt to crystallize the complex [Au(pip) 2 ]Cl; the structure was determined, and consisted, predictably, of molecules with linear coordination at gold.Quite unpredictable at the time was the fact that the molecules associated to form tetramers (Fig. 1) based on an approximately square quadrilateral of gold atoms with short Au� � �Au contacts of 3.301 (5) A ˚, Au� � �Au� � �Au angles of 88.3 � and deviations from the plane of �0.29 A ˚.A literature survey 'X-Ray structural investigations of gold compounds' by one of us (Jones, 1981) presented numerous examples of structures with such Au I � � �Au I contacts, later termed 'aurophilic contacts' by Schmidbaur, who published extensively on the subject (see e.g.Schmidbaur & Schier, 2008, 2012).
Two additional features of the [AuCl(pip)] structure remained unnoticed (or at least were not commented on) at the time.First, the structure contains N-H� � �Cl hydrogen bonds between adjacent molecules of the tetramer, with H� � �Cl 2.57 and N-H� � �Cl 136 � (and a possible weaker branch of a three-centre hydrogen bond, with H� � �Cl 2.91 and N-H� � �Cl 127 � ); hydrogen bonds involving halides bonded to metals are now an established concept, thanks to extensive research by Brammer and others (see e.g. Brammer, 2003).Secondly, the substituents Z at a piperidine ring may adopt an axial or an equatorial position, with C-C-N-Z torsion angles of approximately 180 � , or an axial position, with values of approximately 60 � .Because the equatorial positions are sterically more favourable, these would tend to be occupied preferentially, and this is indeed the case for [AuCl(pip)], with C-C-N-Au torsion angles of �176 � .Clearly a modern redetermination of the structure of [AuCl(pip)] would be worthwhile, for instance to determine directly the positions of the NH hydrogen atom (which had been positioned geometrically, as was normal at the time, rather than directly located and refined).However, despite repeated attempts, we have never again succeeded in synthesizing or crystallizing the complex.
Much later, we succeeded in determining the structure of [Au(pip) 2 ]Cl (Ahrens et al., 1999), which consists of inversionsymmetric dimers with NH� � �Cl À � � �HN linkages.The Au� � �Au distance of 4.085 (2) A ˚within the dimers is too long to be considered an interaction.
Our initial studies involved derivatives of gold(I) chloride, because the easiest access to the amine complexes is the reaction of the amine L (generally as a neat liquid, because many of the products are only stable in the presence of excess amine) with AuCl complexes containing easily displaced ligands such as tetrahydrothiophene (tht) or dimethyl sulfide.However, the products are not easily predictable and can be of various types such as [AuClL], [AuL 2 ]Cl or [AuL 2 ][AuCl 2 ] (see Fig. 2, which presents an overview of various product Structure of the tetrameric unit of [AuCl(pip)] (Guy et al., 1977), which displays crystallographic 4 symmetry.Radii are arbitrary.Dashed lines indicate short Au� � �Au contacts or H� � �Cl hydrogen bonds.Throughout this paper, hydrogen atoms of the CH 2 groups are omitted from the packing diagrams for clarity.

Figure 2
Schematic summary of various structure types for derivatives of amines with gold halides.*X 2 in the case of chlorine refers to the chlorinating agent PhICl 2 rather than elemental chlorine, the use of which has practical difficulties.
types that have been established during our investigations) or in rare cases a mixture such as Au 3 Cl 3 L 4 , as mentioned above for pyrrolidine.
We have since extended the studies to bromide, cyanide (Do ¨ring & Jones, 2013) and thiocyanate (Strey et al., 2018) complexes of gold(I).A further extension has been the attempt to oxidize the gold(I) derivatives to gold(III) analogues.Our studies have however unfortunately shown that these reactions (typically using the oxidising agents PhICl 2 or elemental bromine) often lead to intractable mixtures of products, and that some reactions are extremely sensitive to traces of H + (arising perhaps from adventitious water or by reactions with the solvent), leading to salts of the protonated amine with [AuX 4 ] À and X À anions.Crystallization processes tend to be slow and can lead to decomposition products rather than the intended [AuX 3 L] (L = amine, X = Cl or Br).Nevertheless, the structures thus obtained display some interesting features, which compensate to some extent for the disappointing lack of synthetic efficiency.Here we present the structures of the Au I derivative [AuBr(pyr)]•[Au(pyr) 2 ]Br (2) (�Au 2 Br 2 (pyr) 3 ), together with the Au III derivatives [AuCl 3 (pip)] ( 3), (pipH) 2 Cl[AuCl 4 ] ( 4), the closely related pyrrolidine complex (pyrH) 2 Br[AuBr 4 ] (6), and [AuBr 3 (pip)] as its dichloromethane solvate 7 (see below).For the sake of completeness, we also make brief reference to the structures of [AuBr(pip)]•[Au(pip) 2 ]Br (1) and (pipH) 2 Br[AuBr 4 ] (5), which were however of poor quality.The preparative pathways to these compounds are summarized in Fig. 3.

Structural commentary
At the outset it should be noted that, for the structures 2, 4 and 6, which contain more than one residue in the asymmetric unit, the distinction between the categories Structural commentary (which generally refers to the asymmetric unit) and Supramolecular features becomes blurred.
The formula unit of compound 2 is shown in Fig. 4. For reasons discussed below, the structure is only of moderate quality, but it provides important information on this structure type, which differs in stoichiometry from the chlorido complexes in our previous publications.The compound is formally a 1:1 adduct of [AuBr(pyr)] and [Au(pyr) 2 ]Br (types II and I respectively according to our arbitrary classification).The neutral molecule and the cation display bond lengths and angles that may be regarded as normal (Table 1).The Preparative pathways to the compounds discussed in this paper.

Figure 4
The asymmetric unit of compound 2 with ellipsoids at the 30% probability level.The minor disorder component is not shown.The dashed lines represent the Au� � �Au contact and the hydrogen bond.
Au-N-C-C all lie in the range 157-173 � .The gold atoms are connected by a short aurophilic contact Au1� � �Au2 3.1478 (6) A ˚; the residues lie with the coordination axes at the gold atoms almost perpendicular to each other, so that the torsion angles about the Au1� � �Au2 contact are all roughly 90 � .The NH group at N21 makes a classical hydrogen bond to the bromide anion Br2 (Table 2).The unsatisfactory structure of compound 1 is effectively isotypic to 2; in particular, the absolute torsion angles Au-N-C-C all lie in the narrow range 176-180 � , so that the gold atoms lie equatorially with respect to the piperidine rings.
The molecular structure of compound 3 (type XI) is shown in Fig. 5.The coordination geometry at the gold atom is, as expected, square planar; the r.m.s.deviation of Au1 and its bonded atoms from the plane that they define is only 0.014 A ˚. Bond lengths and angles (Table 3) are normal; the Au1-Cl1 bond length (trans to N11) is not markedly different from the other Au-Cl bond lengths.The Au-N-C-C torsion angles are close to �180 � , so that the AuCl 3 moiety lies in an equatorial position with respect to the piperidine ring.The short intramolecular contact H01� � �Cl3 of 2.57 (3) A ˚might represent a hydrogen bond (Table 4), although the angle is necessarily far from linear at 114 (2) � ; the synperiplanar disposition of H01, N11, Au1 and Cl3, torsion angle À 8 (2) � , would be consistent with this interpretation.
The structure determination of compound 6 in space group Ibam proved to be relatively straightforward (but see section 6 below); the formula unit is shown in Fig. 6.The atoms Au1, Br2 and Br3 lie in the mirror plane x, y, 0; the bromide ions Br4 and Br5 lie on the special positions 0, 0.5, 0 and 0, 0.5, 0.25 respectively, with site symmetry 2/m.The pyrrolidinium ion lies on general positions.The [AuBr 4 ] À anion displays the expected square-planar geometry (Table 5).The NH 2 group forms hydrogen bonds via H01 to the bromide ion Br5, and via H02 to form a three-centre system involving bromide Br4 and metal-bonded Br2 (Table 6); the H02� � �Br2 distance of The asymmetric unit of compound 3 with ellipsoids at the 50% probability level.

Figure 6
The asymmetric unit of compound 6, extended by the symmetryequivalent atom Br1 0 (unlabelled), with ellipsoids at the 50% probability level.The dashed lines indicate hydrogen bonds.
2.90 (6) A ˚is long, but three-centre distances and hydrogen bonds to metal-bonded halogens tend to be longer than conventional hydrogen bonds.
The pseudosymmetric structure 4, closely related to 6, but crystallizing in space group Iba2, is shown in Fig. 7, with the dimensions of the anion in Table 7.Because of the reduced symmetry, all atoms occupy general positions apart from the chloride anions Cl5 and Cl6, which lie on the twofold axes 0.5, 0.5, z, and there are two independent cations.The significant shifts with respect to 6 mean that there is no longer a hydrogen bond from either NH 2 group to a metal-bonded chlorine (Table 8).
The compound [AuBr 3 (pip)] crystallizes as its dichloromethane monosolvate 7 (Fig. 8); the solvent molecule is wellordered.The Au-Br1 bond trans to N11 is slightly shorter than the other Au-Br bonds (Table 9), but somewhat longer than in the [AuBr(pyr)] component of compound 2. The r.m.s.deviation of Au1 and its bonded atoms from the plane that they define is 0.024 A ˚.In contrast to 3, there is no intramolecular hydrogen bond (Table 10) and the torsion angle H01-N11-Au1-Br2 is 31 (2) � .The antiperiplanar Au-N-C-C torsion angles again correspond to an equatorial position of the AuBr 3 group at the piperidine ring.

Supramolecular features
For the compounds consisting of more than one residue, supramolecular features within the asymmetric units have already been discussed in the Structural commentary.
For compound 2, there is a second aurophilic contact Au2� � �Au1 (at 1 2 À x, À 1 2 + y, z) 3.1549 (6) A ˚, so that the extended structure involves an infinite chain of alternating gold atoms Au1/Au2 (and thus of the corresponding residues), involving the b glide operator, with overall direction parallel to the b axis.The chain is almost linear, with Au� � �Au� � �Au angles of 163.07 (2) and 167.78 (2) � .The chains are crosslinked by N-H� � �Br2 hydrogen bonds (Table 2) and the borderline contacts Au1� � �Br2 [3.855 (1) A ˚], forming a layer structure (Fig. 9) parallel to the ab plane in the region z ' 0.25 (with a second such layer at z ' 0.75).

Figure 7
The asymmetric unit of compound 4 with ellipsoids at the 50% probability level.Because of the pseudosymmetry (see text), the carbon and hydrogen atoms had to be refined isotropically and are shown as circles of arbitrary radius.The dashed lines indicate hydrogen bonds.

Figure 8
The asymmetric unit of compound 7 with ellipsoids at the 50% probability level.
For a compound such as 3, the packing may involve a variety of secondary interactions, such as classical or 'weak' hydrogen bonds, Cl� � �Cl contacts (which represent a type of 'halogen bond'; see e.g.Metrangolo et al., 2008), or short Au� � �Cl contacts, sometimes leading to stacking of AuCl 3 units, as observed for [AuCl 3 (tht)] (for which we determined the structures of four different forms; Upmann et al., 2017) and for the primary amine complex [AuCl 3 (isopropylamine)] (Do ¨ring & Jones, 2018a).It is not always straightforward to establish objectively which contacts are more important, because packing patterns are determined by the energy of various interactions, whereas conventional structure determination only gives distances between atoms.For 3 we subjectively assess the important effects to be classical hydrogen bonds and Au� � �Cl interactions; there are, however, several borderline C-H� � �Cl contacts that we do not consider (if only for the sake of simplicity).The hydrogen bond N11-H01� � �Cl3 (Table 4) is quite long, but acceptably linear, and the contact Au1� � �Cl3 (at 1 À x, 1 2 + y, 3 2 À z) at 3.3365 (6) A ˚is short.Both contacts involve c glide operators, and the overall effect is to form a layer structure parallel to the bc plane (Fig. 10).In each layer, the AuCl 3 and NH 2 units lie in the region x ' 0.5, and the rings project outwards from the layer, thus occupying the regions at x ' 0 and 1.

Figure 10
Packing  Similar substructures are present for compound 4 as for 6, but there are significant differences.The chain of piperidinium and chloride ions in 4 is closely analogous to the pyrrolidinium/bromide chain of 6.The tetrachloridoaurate/chloride substructure is topologically closely similar to the tetrabromidoaurate/bromide system of 6, but the distances differ appreciably; thus the gold� � �chloride contacts Au1� � �Cl6(À 1 2 + x, 3 2 À y, z) = 3.8135 (4) A ˚and Au1� � �Cl1( 1 2 À x, À 1 2 + y, z) = 3.995 (3) A ˚of 4 are, counterintuitively, much longer than the corresponding Au� � �Br distances in 6 (and are probably too long to represent appreciable interactions), whereas the interanionic Cl3� � �Cl3(À x, 1 À y, z) contact of 3.085 (5) A ˚is much shorter than its Br� � �Br counterpart in 6; the associated Au-Cl� � �Cl angle is wider than its counterpart in 6 at 168.1 (2) � .Qualitatively, the packing diagrams are the same as those presented in Figs.11 and 12 for 6, and so we do not present analogous diagrams for 4. Another significant difference, as noted in Section 2, is that there are no N-H� � �Cl interactions involving a metal-bonded chloride; this is shown in projections of structures 6 and 4 parallel to their b axes (Figs. 13 and 14; these figures also show clearly the presence of mirror planes, perpendicular to the c axis, that relate pairs of cations in 6, whereas this symmetry element is missing for the corresponding cation pairs in 4).Instead, there are some short C-H� � �Cl contacts that may reasonably be considered as hydrogen bonds (Table 8).Compound 5 seems to be isotypic to 4, but the pseudosymmetry proved too severe to refine the light atoms reliably.We observed similar effects in the struc-tures of two closely related organic compounds, whereby a toluenesulfonyl derivative crystallized in P2 1 /c with Z 0 = 1 (Elgemeie et al., 2013), whereas its benzenesulfonyl analogue crystallized with a closely similar cell and structure in Pc with Z 0 = 2 (Elgemeie et al., 1998).For compound 7, classical hydrogen bonds (Table 10) connect the molecules via a c glide operator to form chains parallel to the c axis (Fig. 15).These are reinforced by offset stacking of neighbouring AuBr 3 units, with Au1� � �Br3(x, 1 2 À y, Finally, adjacent ribbons are connected by the contact Br2� � �Br3(1 + x, 1 2 À y, 1 2 + z) = 3.3817 (4) A ˚to form a layer structure parallel to the ac plane at y ' 0.25.Another such layer lies at y ' 0.75.The dichloromethane molecule is omitted from Fig. 15 for clarity; it forms a weak hydrogen bond to Br1 within the asymmetric unit and also displays a Cl1� � �Cl1 contact of 3.562 (2) A ˚to an adjacent solvent molecule at À x, 1 À y, 1 À z.

Database survey
The searches employed the routine ConQuest (Bruno et al., 2002), part of Version 2022.3.0 of the Cambridge Database (Groom et al., 2016).
A search for all complexes of gold with unsubstituted piperidine or pyrrolidine ligands gave five hits for each; all of these were from our group.For the piperidine structures, one unexpected feature (that we failed to draw attention to at the time of the original publications) was a marked tendency for the gold moiety to lie axially with respect to the piperidine rings.Thus for [Au(pip)(SCN)] (refcode DIXBAQ; Strey et al., 2018), [Au(pip) 2 ][AgCl 2 ] (DUHQUS/DUQHUS01; Ahrens et al., 2000, corrected in Ahrens et al., 2003), [Au(CN)(pip)] (FIMSOL; Do ¨ring & Jones, 2013) and [Au(pip) 2 ]Cl (GOGFEN/GOGFEN01; Ahrens et al., 1999), all the absolute C-C-N-Au torsion angles lie in the range 65-72 � .A possible explanation might be that the low coordination number of gold alleviates the steric disadvantages somewhat.A more extensive search for piperidine complexes of any transition metal gave 193 hits, for most of which the metal residue lay equatorial to the piperidine ring.Almost all of the 35 exceptions belonged to the subset of 64 hits for coinage metals, with their generally low coordination numbers.For the pyrrolidine complexes, the situation was more clear-cut; 60 of the 63 hits had absolute TM-N-C-C torsion angles in the range 140-180 � .
Searches for short intermolecular Cl� � �Cl or Br� � �Br contacts in neutral complexes of the form [AuCl 3 L] or [AuBr 3 L] (as in compound 7) were conducted.The Cl� � �Cl search gave 51 hits with distances up to 3.5 A ˚, twice the maximum (CCDC-defined) van der Waals radius, of which 24 were shorter than 3.4 A ˚.The shortest were 3.086 and 3.191 A between cis (to L) chlorines in two carbene complexes (HOLGUM and HOKJIC; Teci et al., 2017 andToma ´s-Mendivil et al., 2013).The Br� � �Br search gave 28 hits up to 3.7 A ˚; 11 were shorter than 3.6 A ˚.The shortest was 3.260 A between cis bromines in a phosphine sulfide complex (BOKQUQ; Upmann et al., 2019).

Synthesis and crystallization
Syntheses were performed under an atmosphere of dry nitrogen; the small-scale crystallization experiments were performed in laboratory air.
Bromido(piperidine)gold(I) bis(piperidine)gold(I) bromide ( 1): 90 mg (0.247 mmol) [AuBr(tht)] were dissolved in 2 mL piperidine.The solution was divided into five small test tubes and overlayered with five different precipitants.The test-tubes were stoppered and stored in a refrigerator at 278 K for 1 day.Crystals in the form of colourless laths were obtained in small quantities using petroleum ether as precipitant, despite considerable decomposition that led to a gold mirror.
Trichlorido(piperidine)gold(III) (3) and bis(piperidinium) chloride tetrachloridoaurate(III) ( 4): 120 mg (0.374 mmol) [AuCl(tht)] were dissolved in a mixture of 4 mL of piperidine and 4 mL of dichloromethane.The solution was overlayered with n-pentane in a 100 mL round-bottomed flask and transferred to the refrigerator overnight.The supernatant was pipetted off and the solid residue (presumed to be [Au(pip) 2 ] Cl) dried in vacuo (148.7 mg, 0.369 mmol).The solid was divided into two parts; each was dissolved in 2 mL of dichloromethane, and 50.7 mg (0.184 mmol) of PhICl 2 in 2 mL of dichloromethane was added to each, causing the solutions to turn first red and then orange.After 16 days at 278 K, crystals of 3 (yellow blocks and laths, 91% yield) were obtained using n-heptane as precipitant, and of 4 (a few yellow laths) using petroleum ether.
Bis(piperidinium) bromide tetrabromidoaurate(III) ( 5): 49.6 mg (0.136 mmol) [AuBr(tht)] were dissolved in 2 mL piperidine and overlayered with n-pentane in a test-tube, which was stoppered and transferred to the refrigerator overnight.The supernatant was pipetted off and the solid residue dried in vacuo to give 541 mg (0.067 mmol, 98%) of compound 1.This was dissolved in 2 mL of dichloromethane, and 2 drops of elemental bromine were added.The solution was overlayered with diisopropyl ether and stored at 278 K for 4 days, leading to crystals in the form of red plates (78% yield).Elemental analysis [%]: calc.: C 15.62, H 3.15, N 3.64; found: C 15.68, H 3.32, N 3.86.
Bis(pyrrolidinium) bromide tetrabromidoaurate(III) (6): 135.7 mg (0.372 mmol) of [AuBr(tht)] were dissolved in 2 mL of pyrrolidine.Diisopropyl ether was added until a permanent turbidity was observed, and the mixture was transferred to the refrigerator overnight.The supernatant was pipetted off and the solid dark-grey residue was taken up in 4 mL of dichloromethane.After filtration, two drops of elemental bromine were added, leading to a dark-red solution with a dark-red solid residue (not identified).After this had settled, the clear solution was pipetted off into five test-tubes and overlayered as above for 1.After 1 day at 278 K, crystals in the form of dark-red tablets and laths (yield not determined) were obtained with diisopropyl ether as precipitant.
Tribromido(piperidine)gold(III) dichloromethane solvate (7): 127 mg (0.157 mmol) of compound 1 were dissolved in 4 mL of dichloromethane, and two drops of elemental bromine were added. 2 mL of the solution were subjected to five different precipitants as described above for 1; after 10 days at 278 K, crystals in the form of orange-red needles (yield not determined) were obtained using n-heptane as precipitant.Elemental analysis [%]: calculated (including the solvent content): C 11.88, H 2.16, N 2.31; found: C 11.83, H 2.19, N 3.15.

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 11.Structures were refined anisotropically on F 2 .Hydrogen atoms of the NH groups were  Special details and exceptions: The structure of compound 2 was difficult to refine satisfactorily; the data are weak and the absorption coefficient is high.Hydrogen atoms of the NH groups were located in difference syntheses and refined freely, but with N-H distances restrained to be approximately equal (command 'SADI'); the positions of freely refined hydrogen atoms in heavy-atom structures should of course be interpreted with caution, but seem to be acceptable for 2 and for the other structures presented here.The ring at N31 is disordered, with atoms C33 and C34 occupying alternative sites with occupation factors 0.64 (2) and 0.36 (2); atoms of the minor site were refined isotropically.Appropriate restraints were employed to improve refinement stability, but the dimensions of disordered groups should always be interpreted with caution.The residual electron density near the gold atom was high, which is probably attributable to residual absorption errors; for poor data, errors are likely to be reflected in this way.Nevertheless we believe that the refinement provides at least a qualitatively reliable picture of the structure.Data for compound 1 were also collected, showing that it is effectively isotypic to 2, but the refinement was highly unsatisfactory, with two very large difference peaks that did not lie close to the gold atom.Despite considerable efforts, we were unable either to explain these peaks (e.g. by detecting additional weak reflections corresponding to a larger cell or indicating twinning effects) or to collect better data from other crystals of 1 and 2.
The structure of compound 4, which is pseudosymmetric, was refined as a two-component inversion twin; the relative volume of the smaller component refined to 0.45 (3).Originally the structure was refined in space group Ibam, whereby the gold atom and four of the five chlorine atoms lay in mirror planes; the piperidinium cations were disordered.However, it can be refined with ordered cations in Iba2, so we prefer this model.Because of the well-known difficulties of refining an almost centrosymmetric structure in a non-centrosymmetric space group, the light atoms (carbon and nitrogen) had to remain isotropic, and many restraints were necessary to improve refinement stability.The dimensions of the cations should therefore be interpreted with caution.There is also the danger that the refinement results may represent a false minimum (although these often involve chemically implausible structures).The hydrogen atoms, in particular those of the NH 2 groups, could not be located in difference syntheses and were therefore included using a riding model starting from calculated positions (with N-H 0.91 A ˚).The closely related structure of compound 6, however, was successfully refined in Ibam without disorder.The hydrogen atoms of the NH 2 group were refined freely, but with N-H distances restrained to be approximately equal (command 'SADI').We also recorded a dataset for 5, which appears to be isotypic to 4, but for which the pseudosymmetry proved too severe to allow satisfactory refinement.
We note also that, for rings of the form [(CH 2 ) n NH 2 ] + , it may be difficult to distinguish between the carbon and nitrogen atoms in the presence of heavy atoms (especially for pseudosymmetric structures such as 4).Our assignments of these atoms were based on U values (although these are somewhat irregular, e.g. the low value of 0.014 A ˚2 for C14 of 4) and, more importantly, on the hydrogen-bonding patterns of the corresponding hydrogen atoms; thus only the hydrogen atoms of the chosen nitrogen sites are involved in the short hydrogen-halide contacts of the cation/halide chains of 4 and 6.However, the H� � �Br distances for hydrogen bonds N-H� � �Br and C-H� � �Br are unlikely to differ greatly, so some degree of C/N disorder for 6 cannot be ruled out.
For compound 7, checkCIF suggested a smaller cell, generated by halving the b axis.However, careful inspection of the data shows that the reported cell is correct.Reflections with k odd are weaker, but definitely present.We note that the gold atom and two of the three bromine atoms have y coordinates of approximately 0.25; this is probably the factor responsible for the systematically weak reflections.

Figure 9
Figure 9 Packing diagram of compound 2 viewed parallel to the c axis in the region z ' 0.25.For all packing diagrams, hydrogen atoms bonded to carbon are omitted for clarity.Thin dashed lines indicate hydrogen bonds or Au1� � �Br2 interactions; thick dashed lines indicate aurophilic interactions.Note that the [AuBr(pyr)] molecules (involving Au1) are viewed end-on.Labelled atoms belong to the asymmetric unit.
diagram of compound 3 viewed perpendicular to the bc plane.Hydrogen atoms bonded to carbon are omitted for clarity.Thick dashed lines indicate hydrogen bonds; thin dashed lines indicate Au1� � �Cl3 interactions.The labelled atom belongs to the asymmetric unit.

Figure 11
Figure 11Packing of compound 6, first substructure; the bromide and pyrrolidinium ions combine to form a chain parallel to the c axis.The view direction is parallel to the a axis.Dashed lines indicate hydrogen bonds.Labelled atoms belong to the asymmetric unit.
Figure 12Packing of compound 6, second substructure; the bromide ions Br4 and the tetrabromidoaurate(III) ions (seen edge-on) combine to form a layer structure parallel to the ab plane.The view direction is parallel to the c axis, and the region shown is at z ' 0. Dashed lines indicate (thin) Au� � �Br or (thick) Br� � �Br interactions.Labelled atoms belong to the asymmetric unit.

Figure 13
Figure 13 Projection of the structure of compound 6 parallel to the b axis.Thick dashed lines indicate the hydrogen bonds of the type N-H� � �Br-Au, thin dashed lines indicate other hydrogen bonds; other contacts are not explicitly included.

Figure 14
Figure 14Projection of the structure of compound 4 parallel to the b axis.Dashed lines indicate hydrogen bonds; other contacts are not explicitly included.Note the absence of hydrogen bonds of the type N-H� � �Cl-Au (cf.Fig.13).

Figure 15
Figure 15 Packing of compound 7. Solvent molecules are omitted.The view direction is parallel to the b axis (so that the molecules are seen approximately end-on), and the region shown is at y ' 0.25.Thick dashed lines indicate hydrogen bonds; thin dashed lines indicate Au� � �Br or Br� � �Br interactions.Labelled atoms belong to the asymmetric unit.

Table 11
Experimental details.Methylene hydrogens were included at calculated positions and refined using a riding model with C-H = 0.99 A ˚and H-C-H = 109.5� .Isotropic U(H) values were fixed at 1.2 � U eq of the parent carbon atom (or nitrogen, see below).