Effect of pressure on the crystal structure of L-serine-I and the crystal structure of L-serine-II at 5.4 GPa

The crystal structure of L-serine has been determined at room temperature at pressures between 0.3 and 4.8 GPa. The structure of this phase (hereafter termed L-serine-I), which consists of the molecules in their zwitterionic tautomer, is orthorhombic, space group P2₁2₁2₁. The least compressible cell dimension (c), corresponds to chains of head-to-tail NH⋯carboxylate hydrogen bonds. The most compressible direction is along b, and the pressure-induced distortion in this direction takes the form of closing up voids in the middle of R-type hydrogen-bonded ring motifs. This occurs by a change in the geometry of hydrogen-bonded chains connecting the hydroxyl groups of the —CH₂OH side chains. These hydrogen bonds are the longest conventional hydrogen bonds in the system at ambient pressure, having an O⋯O separation of 2.918 (4) Å and an O⋯O⋯O angle of 148.5 (2)°; at 4.8 GPa these parameters are 2.781 (11) and 158.5 (7)°. Elsewhere in the structure one NH⋯O interaction reaches an N⋯O separation of 2.691 (13) Å at 4.8 GPa. This is amongst the shortest of this type of interaction to have been observed in an amino acid crystal structure. Above 4.8 GPa the structure undergoes a single-crystal-to-single-crystal phase transition to a hitherto uncharacterized polymorph, which we designate L-serine-II. The OH⋯OH hydrogen-bonded chains of L-serine-I are replaced in L-serine-II by shorter OH⋯carboxyl interactions, which have an O⋯O separation of 2.62 (2) Å. This phase transition occurs via a change from a gauche to an anti conformation of the OH group, and a change in the NC_αCO torsion angle from −178.1 (2)° at 4.8 GPa to −156.3 (10)° at 5.4 GPa. Thus, the same topology appears in both crystal forms, which explains why it occurs from one single-crystal form to another. The transition to L-serine-II is also characterized by the closing-up of voids which occur in the centres of other R-type motifs elsewhere in the structure. There is a marked increase in CH⋯O hydrogen bonding in both phases relative to L-serine-I at ambient pressure.

1. Introduction

Molecular crystals display a wide range of intermolecular interactions, from strong ionic and hydrogen-bonding contacts to weak van der Waals contacts. The application of high pressure to organic materials is a very powerful way to probe the nature of these interactions. The magnitudes of the effects which are observed are generally greater than those observed on cooling. Pressure-induced polymorphism occurs in a number of systems. We have characterized, for example, new high-pressure phases in alcohols (Allan & Clark, 1999; Allan et al., 2001, 2002), carboxylic acids (Allan et al., 1998, 2000), acetone (Allan et al., 1999) and, very recently, glycine (Dawson et al., 2005). A different high-pressure phase of glycine has also been recently reported by Boldyreva et al. (2004). The number of high-pressure studies on molecular systems that have actually been carried out is still rather small and systematic trends have yet to emerge. However, this is a rapidly emerging area of structural science and it has been the subject of a number of recent reviews, for example, Boldyreva (2003, 2004a,b,c), Katrusiak (2004) and Hemley & Dera (2000).

Organic compounds crystallize predominantly in low-symmetry crystal systems and the effect of the application of pressure is generally quite anisotropic. The compressibility along different crystallographic directions can occasionally be rationalized in terms of the strengths of hydrogen bonds made along different directions. For example, both Boldyreva et al. (2003) and we (Dawson et al., 2005) have shown that the least compressible lattice direction of α-glycine corresponds to the direction of strongly hydrogen-bonded chains. However, in [Co(NH₃)₅NO₂]Cl₂ some hydrogen bond lengths actually increase with pressure (Boldyreva et al., 1998) and it is clear that the behaviour of hydrogen bonds under high pressure depends not only on the bonds themselves, but also on their relationship to other features of a structure, such as other intermolecular interactions and crystal packing.

The extent to which compressibility can be explained, and how far a structure can be compressed before it undergoes a phase transition, are key issues of current interest in this area of crystallography. In this paper we attempt to address them in a study of the effect of pressure on L-serine. Amino acids have been studied extensively at ambient pressure both by neutron and X-ray diffraction; they are highly crystalline and their structures are dominated by hydrogen bonding (Jeffrey & Maluszynska, 1982). Weak CH⋯O hydrogen bonds occur frequently and play an important role in supporting more familiar medium-strength hydrogen bonds, e.g. NH⋯O (Desiraju & Steiner, 1999; Derewenda et al., 1995). Amino acids therefore make excellent candidates for this kind of study, but we hope that the results will additionally be useful for the development of inter-residue potentials which can be used to model the nature of pressure effects in proteins and other complex systems.

2. Experimental

2.1. Crystal growth

L-Serine (99%) was purchased from Aldrich (catalogue number S2,60-0). One small, block-shaped crystal was obtained directly from the sample bottle and loaded into a diamond anvil cell.

2.2. High-pressure crystallography

High-pressure experiments were carried out using a Merrill–Bassett diamond anvil cell (half-opening angle 40°), equipped with brilliant-cut diamonds with 600 µm culets and a tungsten gasket (Merrill & Bassett, 1974). A 1:1 mixture of n-pentane and isopentane was used as a hydrostatic medium. A small ruby chip was also loaded into the cell as the pressure calibrant, with the ruby fluorescence method utilized to measure the pressure. Measurements were carried out by excitation with a 632.417 nm line from a He–Ne laser, the fluorescence being detected with a Jobin–Yvon LabRam 300 Raman spectrometer.

Diffraction data were collected on a Bruker SMART APEX diffractometer with graphite-monochromated Mo Kα radiation (λ = 0.71073 Å). A hemisphere of data was collected at room temperature using the crystal before it was mounted in the Merrill–Bassett cell. The crystal was orthorhombic and its unit-cell dimensions were a = 8.579 (4), b = 9.349 (4), c = 5.613 (3) Å based on 783 data 8 < 2θ < 45°. The L-serine coordinates of Kistenmacher et al. (1974) were refined against these data to yield a conventional R-factor of 0.029 for 408 data with I > 2σ(I). The aim of this experiment was simply to establish the starting phase of the sample used in this pressure study, and further crystallographic data are not given here.

Data collection and processing procedures for the high-pressure experiments were as described by Dawson et al. (2004). Integrations were carried out using the program SAINT (Bruker AXS, 2003), and absorption corrections with the programs SADABS (Sheldrick, 2004) and SHADE (Parsons, 2004). Data collections were taken in approximately 1.0 GPa steps from 0.3 GPa up to a final pressure of 5.4 GPa. Determinations of the cell constants at 5.4 GPa showed that a single-crystal-to-single-crystal phase transition had occurred to a new polymorph (L-serine-II). The pressure was then reduced back down to ambient pressure and the sample removed from the pressure cell. Once removed, a hemisphere of X-ray diffraction data was collected at room temperature. The phase on return to ambient pressure was identified as L-serine-I on the basis of the unit-cell constants [orthorhombic, a = 8.531 (9), b = 9.249 (10), c = 5.581 (6) Å] and structure refinement of L-serine-I coordinates yielded a conventional R factor of 0.041. This experiment aimed simply to establish the phase of serine after removal from the cell so further data are not given here.

Refinements of the compressed form of L-serine-I were carried out starting from the published coordinates determined at ambient pressure. The structure of the new phase (L-serine-II) was solved by the global minimization method using the program DASH (David et al., 2001). Refinements were carried out against |F|² using all data (CRYSTALS; Betteridge et al., 2003). Owing to the low completeness of the data sets, all 1,2 and 1,3 distances were restrained to the values observed in the ambient pressure structure, and all C, N and O atoms were refined with isotropic displacement parameters.

H atoms attached to carbon and nitrogen were placed geometrically and not refined. At ambient pressure Kistenmacher et al. (1974) showed that the hydroxyl H atom (H7) eclipses C3—H2 with r(OH) = 0.88 Å and <COH = 107°; we have confirmed these results. This feature is ascribable to the formation of intermolecular OH⋯OH hydrogen bonds (see §3). In placing the hydroxyl H atom (H7) in the structures between 0.3 and 4.8 GPa, we initially assumed that the ambient pressure conformation of the CH₂OH side chain was retained and this atom was placed in an ideal position for OH⋯OH hydrogen bonding. However, the positional parameters of H7 were refined subject to the restraints r(O—H) = 0.88 (1) Å and < COH = 107 (1)°, so enabling the HOCC torsion angle to optimize. In all except the 4.8 GPa data set O3—H7 eclipsed C2—H2 as it does at ambient pressure. At 4.8 GPa O3—H7 appeared to adopt a staggered orientation with respect to the neighbouring CH₂ group; refinements in which it was restrained in an eclipsed position failed to converge. Of course, the standard uncertainties on the positional parameters of H7 are so large that the differences between the two models are not statistically significant, but in the 4.8 GPa model presented here H7 is left in its refined position. A definitive statement regarding the position of H7 at 4.8 GPa is not possible from these data, but neutron diffraction experiments would clarify this issue. At 5.4 GPa, H7 was observed in a difference map, but treated during refinement in the same way as at lower pressure. All distances and angles involving H quoted in this paper were calculated after normalizing the H-atom position to mimic those that might be obtained by neutron diffraction [r(C—H) = 1.083, r(N—H) = 1.009, r(O—H) = 0.983 Å].

Listings of crystal and refinement data are given in Table 1.¹

Table 1
Crystallographic data for L-serine at increasing pressures

Weighting scheme: p = P(6)*max(F_o²,0) + (1 − P(6))F_c². Method = SHELXL97 (Sheldrick, 1997).

Pressure	0.3 GPa	1.4 GPa	2.9 GPa
Crystal data
Chemical formula	C3H7NO3	C3H7NO3	C3H7NO3
Mr	105.09	105.09	105.09
Cell setting, space group	Orthorhombic, P212121	Orthorhombic, P212121	Orthorhombic, P212121
a, b, c (Å)	8.5213 (13), 9.172 (2), 5.5847 (8)	8.4365 (10), 8.9506 (19), 5.5512 (6)	8.3702 (10), 8.7699 (19), 5.5103 (6)
V (Å^3)	436.47 (15)	419.18 (11)	404.49 (11)
Z	4	4	4
Dx (Mg m^−3)	1.599	1.665	1.726
Radiation type	Mo Kα	Mo Kα	Mo Kα
No. of reflections for cell parameters	246	270	258
θ range (°)	9–46	9–46	9–47
μ (mm^−1)	0.14	0.15	0.15
Temperature (K)	293	293	293
Crystal form, colour	Block, colourless	Block, colourless	Block, colourless
Crystal size (mm)	0.20 × 0.10 × 0.10	0.20 × 0.10 × 0.10	0.20 × 0.10 × 0.10
Data collection
Diffractometer	Bruker SMART	Bruker SMART	Bruker SMART
Data collection method	ω scans	ω scans	ω scans
Absorption correction	Multi-scan (based on symmetry-related measurements)	Multi-scan (based on symmetry-related measurements)	Multi-scan (based on symmetry-related measurements)
Tmin	0.695	0.552	0.711
Tmax	1.00	1.00	1.00
No. of measured, independent and observed reflections	1169, 151, 93	1006, 146, 100	1112, 134, 100
Criterion for observed reflections	I > 2σ(I)	I > 2σ(I)	I > 2σ(I)
Rint	0.138	0.127	0.119
θmax (°)	23.1	23.3	23.3
Range of h, k, l	−9 → h → 9	−9 → h → 9	−9 → h → 9
	−3 → k → 3	−3 → k → 3	−3 → k → 3
	−6 → l → 6	−6 → l → 6	−6 → l → 6
Refinement
Refinement on	F^2	F^2	F^2
R[F^2 > 2σ(F^2)], wR(F^2), S	0.083, 0.214, 1.07	0.073, 0.164, 1.14	0.065, 0.154, 1.07
No. of reflections	140	135	133
No. of parameters	33	33	33
H-atom treatment	Mixture of independent and constrained refinement	Mixture of independent and constrained refinement	Mixture of independent and constrained refinement
Weighting scheme	w = 1/[σ^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)sin θ]; P(i) are: 0.909E^−1, 2.37, 0.00, 0.00, 0.00, 0.333	w = 1/[σ^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)sin θ]; P(i) are: 0.291E^−1, 2.72, 0.00, 0.00, 0.00, 0.333	w = 1/[σ^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)sin θ]; P(i) are: 0.597E−1, 1.46, 0.00, 0.00, 0.00, 0.333
(Δ/σ)max	<0.0001	<0.0001	<0.0001
Δρmax, Δρmin (e Å^−3)	0.30, −0.23	0.37, −0.27	0.26, −0.24
Extinction method	Larson (1970)	Larson (1970)	Larson (1970)
Extinction coefficient	60 (50)	30 (20)	30 (30)

	4.1 GPa	4.8 GPa	5.4 GPa
Crystal data
Chemical formula	C3H7NO3	C3H7NO3	C3H7NO3
Mr	105.09	105.09	105.09
Cell setting, space group	Orthorhombic, P212121	Orthorhombic, P212121	Orthorhombic, P212121
a, b, c (Å)	8.3266 (13), 8.665 (3), 5.4851 (8)	8.2980 (16), 8.600 (3), 5.4663 (10)	6.9083 (10), 9.644 (3), 5.6166 (8)
V (Å^3)	395.75 (15)	390.09 (17)	374.19 (14)
Z	4	4	4
Dx (Mg m^−3)	1.764	1.789	1.865
Radiation type	Mo Kα	Mo Kα	Mo Kα
No. of reflections for cell parameters	258	247	313
θ range (°)	9–47	9–44	6–46
μ (mm^−1)	0.16	0.16	0.17
Temperature (K)	293	293	293
Crystal form, colour	Block, colourless	Block, colourless	Block, colourless
Crystal size (mm)	0.20 × 0.10 × 0.10	0.20 × 0.10 × 0.10	0.20 × 0.10 × 0.10
Data collection
Diffractometer	Bruker SMART	Bruker SMART	Bruker SMART
Data collection method	ω scans	ω scans	ω scans
Absorption correction	Multi-scan (based on symmetry-related measurements)	Multi-scan (based on symmetry-related measurements)	Multi-scan (based on symmetry-related measurements)
Tmin	0.721	0.739	0.687
Tmax	1.00	1.00	1.00
No. of measured, independent and observed reflections	1112, 130, 103	1068, 129, 99	1990, 140, 90
Criterion for observed reflections	I > 2σ(I)	I > 2σ(I)	I > 2σ(I)
Rint	0.080	0.083	0.081
θmax (°)	23.4	23.2	23.3
Range of h, k, l	−9 → h → 9	−9 → h → 9	−7 → h → 7
	−2 → k → 2	−2 → k → 2	−3 → k → 3
	−6 → l → 6	−6 → l → 6	−6 → l → 6
Refinement
Refinement on	F^2	F^2	F^2
R[F^2 > 2σ(F^2)], wR(F^2), S	0.066, 0.165, 1.06	0.060, 0.122, 1.15	0.048, 0.102, 1.21
No. of reflections	129	128	121
No. of parameters	32	33	32
H-atom treatment	Mixture of independent and constrained refinement	Mixture of independent and constrained refinement	Mixture of independent and constrained refinement
Weighting scheme	w = 1/[σ^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)sin θ]; P(i) are: 0.672E^−1, 1.89, 0.00, 0.00, 0.00, 0.333	w = 1/[σ^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)sin θ]; P(i) are: 0.00, 1.64, 0.00, 0.00, 0.00, 0.333	W = 1/[σ^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)sin θ] P(i) are: 0.219E^−1, 1.16, 0.00, 0.00, 0.00, 0.333
(Δ/σ)max	<0.0001	<0.0001	<0.0001
Δρmax, Δρmin (e Å^−3)	0.25, −0.20	0.19, −0.21	0.25, −0.20
Extinction method	None	Larson (1970)	None
Extinction coefficient	–	44 (19)	–

Crystal structures were visualized using the programs CAMERON (Watkin et al., 1993) and MERCURY (Bruno et al., 2002). Analyses were carried out using PLATON (Spek, 2004), as incorporated in the WIN-GX suite (Farrugia, 1999). Searches of the Cambridge Database (Allen, 2002; Allen & Motherwell, 2002) utilized the program CONQUEST and Version 5.25 of the database with updates up to April 2004.

The numbering scheme used is the same as in the CSD refcode LSERIN01 (Kistenmacher et al., 1974). In macromolecular structures our C2, C3 and O3 would be designated CA, CB and OG, respectively. The settings of the structures reported here are the same as used in LSERIN01; that used for L-serine-II was chosen to facilitate the comparison with L-serine-I.

3. Results and discussion

3.1. Structure of L-serine-I at ambient pressure

Prior to this work there was only one known crystalline form of anhydrous L-serine and this crystallizes with one molecule in the asymmetric unit in the space group P2₁2₁2₁. We refer to this form as L-serine-I. The structure was determined using X-ray diffraction by Benedetti et al. (1973) and then later by Kistenmacher et al. (1974). The serine molecule is in its zwitterionic form [see (I)], with the CH₂OH side-chain in the gauche conformation with respect to the ammonium and carboxyl groups (χ₁ = 61.5°). This conformation is observed under all conditions investigated during this work.

The structure of L-serine-I is dominated by hydrogen bonding (Figs. 1a–3a show projections of the structure along a, b and c, respectively). Many amino-acid crystal structures have one cell dimension of ca 5.5 Å and this is associated with a head-to-tail chain motif formed by NH⋯OOC interactions. This is observed in L-serine-I, where the molecules form a chain via lattice repeats along the crystallographic c direction through three-centre N1H5⋯O1/2 interactions (Fig. 1a). Jeffrey & Maluszynska (1982) have shown that such interactions can take on varying degrees of asymmetry and that observed here is relatively symmetrical, with distances of 1.91 (H5⋯O2) and 2.29 Å (H5⋯O1). If the weaker hydrogen bond is ignored, the graph-set descriptor of this chain is C(5) (Bernstein et al., 1995).

Figure 1 Effect of pressure on the crystal structure of L-serine as viewed along a: (a) L-serine-I at ambient pressure; (b) L-serine-I at 4.8 GPa; (c) L-serine-II at 5.4 GPa. This layer is referred to as the A-layer in the text. Colour scheme: red oxygen, blue nitrogen, green carbon and white hydrogen. The orientations of all diagrams are the same; the scale is the same as that used in Figs. 2 and 3. The bifurcation of the N1H5⋯O1/2 and CH⋯O bonding are shown only in the bottom-left fragment of the diagram.

Figure 2 Effect of pressure on the crystal structure of L-serine as viewed along b: (a) L-serine-I at ambient pressure; (b) L-serine-I at 4.8 GPa; (c) L-serine-II at 5.4 GPa. This layer is referred to as the B-layer in the text. The orientations of all the diagrams are the same; the scale is the same as that used in Figs. 1 and 3. The bifurcation of the N1H5⋯O1/2 and CH⋯O bonding are shown only in the bottom-left fragment of the diagram. The colour scheme is as shown in Fig. 1.

Figure 3 Effect of pressure on the crystal structure of L-serine as viewed along c: (a) L-serine-I at ambient pressure; (b) L-serine-I at 4.8 GPa; (c) L-serine-II at 5.4 GPa. The A-layers run vertically, the B-layers horizontally. The orientations of all diagrams are the same; the scale is the same as that used in Figs. 1 and 2. The CH⋯O bonding is shown only in the bottom-left fragment of the diagram. Colour scheme: as shown in Fig. 1.

A second C(5) chain, generated by the 2₁ about c, is linked to the first via N1H6⋯O1 hydrogen bonds [N1⋯O1 2.840 (4) Å)] to form a ribbon. The N1H6⋯O1 interactions also generate primary level C(5) chains along the ribbon. Along the length of the ribbon the combination of the two C(5) chains forms secondary-level R₃³(11) ring motifs. The CH₂OH side chains are distributed along the outside edges of the ribbons and these interact via C(2)⋯O3H7⋯O3H7 hydrogen bonds to link the ribbons into layers. The hydrogen bonds between the hydroxyl groups are quite weak, with O3⋯O3 measuring 2.918 (4) Å under ambient conditions. The combination of the C(2) and C(5) N1H5…⋯O2 chains generates secondary-level R₃³(13) ring motifs (Fig. 1a).

The layers are stacked along a, having a sinusoidal appearance when viewed in projection onto (001) (Fig. 3a). The layers are linked by N1H4⋯O2 interactions which form yet another primary level C(5) chain which runs along a (Fig. 2a). The intersection of the N1H5⋯O2 and N1H4⋯O2 C(5) chains along a and c builds a third set of secondary-level ring motifs, these having the descriptor R³₄(14).

The N1H5⋯O2 and N1H4⋯O2 interactions actually build another layer which is parallel to the ac plane (Fig. 2a). Overall then, the structure consists of two sets of layers: one stacks along a and contains R₃³(11) and R₃³(13) ring motifs, the other is more planar, stacks along b and contains R³₄(14) rings. We shall refer to these as the A and B layers, respectively. The N1H4⋯O2 interactions which occur within the B layers can also be viewed as interactions between the A layers; similarly, the N1H6⋯O1 and O3H7⋯O3H7 interactions within the A layers serve to link the B layers (Fig. 3a). The N1H5⋯O2 hydrogen bonds are common to both layers.

Plots of the cell dimensions and volume of L-serine as a function of pressure are given in Fig. 4. L-Serine-I is stable to 4.8 GPa (48 kbar). Above this pressure it undergoes a single-crystal-to-single-crystal phase transition to a new phase, which we designate L-serine-II. We prefer this I-, II-, etc. phase nomenclature to the α-, β-, etc. nomenclature for amino acids even though the polymorphs of glycine are denoted α, β, γ, etc. because the symbols α and β are used for other purposes in amino-acid chemistry.

Figure 4 Variation of the lattice parameters (a–c, Å) and volume (d, Å^3) of L-serine as a function of pressure (GPa).

3.2. Response of L-serine-I to pressure up to 4.8 GPa

The response of the unit-cell dimensions of L-serine-I to high pressure is anisotropic (Fig. 4), although, since the crystal system is orthorhombic, the principal axes of the strain tensor must be coincident with the crystallographic axes. The largest reduction occurs in the b axis (6.2%), while the a and c axes change by 2.6 and 2.1%, respectively. The volume changes most rapidly between 0.2 and 2.9 GPa, and then the trend flattens-off up to 4.8 GPa; above 4.8 GPa all three axis lengths change suddenly during the transition to L-serine-II. Fig. 5 shows the superposition of the structures of L-serine-I at ambient pressure and at 4.8 GPa, where the molecules are represented by their inertial tensors. The orientations of the molecules change slightly, but the large compression along b is readily apparent.

Figure 5 Comparison of packing in the structures of L-serine-I at ambient pressure (blue) and 4.8 GPa (red). Molecules are represented by their inertial tensor ellipsoids. The unit cell shown corresponds to the ambient pressure phase. Note that the molecules do not greatly change orientation.

The variation of hydrogen-bonding parameters in L-serine-I between 0.3 and 4.8 GPa is presented in Table 2. N1H4⋯O2 shortens from N⋯O 2.887 (4) Å at ambient pressure to 2.691 (13) Å at 4.8 GPa (Fig. 2b). A search of the Cambridge Database reveals that there are only three amino acid structures (out of 213) in which NH⋯O interactions are shorter than this, the shortest, 2.661 Å, being observed in L-arginine L-glutamate trihydrate (DUSMAF; Suresh et al., 1986). The shortening of this distance occurs quite smoothly between 0.3 and 4.8 GPa. N1H6⋯O1 is formed approximately along the b* axis (Fig. 1b) and reflection data along this direction of reciprocal space were severely shaded by the pressure cell. This distance is therefore not very precisely determined in the present study, but it also shortens from 2.840 (4) to 2.72 (3) Å.

Table 2
Hydrogen-bonding parameters (Å, °) in L-serine-I

The distances to hydrogen have been normalized to neutron values (see §2.2).

Pressure (GPa)	0	0.3	1.4	2.9	4.1	4.8
N1H5⋯O2^i
H5⋯O2	1.91	1.94	1.89	1.86	1.86	1.86
N1⋯O2	2.871 (3)	2.862 (15)	2.814 (14)	2.786 (12)	2.775 (12)	2.775 (13)
∠N1H5O2	158 (2)	153	152	153	150	150
N1H5⋯O1^i
H5⋯O1	2.29	2.30	2.28	2.24	2.22	2.19
N1⋯O1	3.118 (3)	3.106 (15)	3.070 (14)	3.028 (11)	3.010 (12)	2.981 (12)
∠N1H5O1	139 (2)	138	136	136	136	136
N1H4⋯O2^ii
H4⋯O2	1.90	1.89	1.86	1.82	1.80	1.78
N1⋯O2	2.887 (4)	2.825 (16)	2.796 (15)	2.751 (12)	2.709 (13)	2.691 (13)
∠N1H4O2	167 (2)	153	153	153	150	149
N1H6⋯O1^iii
H6⋯O1	1.87	1.86	1.81	1.80	1.79	1.76
N1⋯O1	2.840 (4)	2.81 (3)	2.76 (3)	2.75 (2)	2.75 (3)	2.72 (3)
∠N1H6O1	162 (2)	158	157	157	159	158
O3H7⋯O3^iv
H7⋯O3	2.02	1.98	1.98	2.00	1.92	2.08
O3⋯O3	2.918 (4)	2.882 (16)	2.842 (14)	2.807 (12)	2.790 (12)	2.781 (11)
∠O3H7O3	153 (3)	153 (14)	147 (7)	139 (6)	147 (6)	129 (7)
∠O3⋯O3⋯O3	148.5 (2)	151.2 (14)	155.4 (12)	158.1 (8)	158.9 (7)	158.5 (7)
C2H1⋯O1^v
H1⋯O1	2.75	2.67	2.56	2.48	2.45	2.44
C2⋯O1	3.368 (4)	3.307 (16)	3.222 (15)	3.153 (13)	3.113 (13)	3.106 (13)
∠C2H1O1	118.3(13)	118	120	121	120	120
C3H2⋯O3^iv
H2⋯O1	2.56	2.57	2.54	2.51	2.50	2.48
C3⋯O1	3.214 (4)	3.182 (17)	3.147 (16)	3.099 (14)	3.076 (13)	3.062 (13)
∠C3H2O3	117.6 (12)	116	116	115	114	114
C3H3⋯O1^v
H2⋯O1	2.55	2.48	2.46	2.44	2.43	2.42
C3⋯O1	3.053 (4)	3.005 (14)	2.942 (13)	2.905 (11)	2.876 (11)	2.871 (11)
∠C3H3O1	109.4 (18)	110	107	107	105	105
Symmetry codes: (i) x,y,1+z; (ii) ; (iii) ; (iv) ; (v) .

The least compressible interaction is the head-to-tail, bifurcated, N1H5⋯O1/2 chain-forming interaction along c. The O⋯O interaction should just have one set which constitutes the three-centre hydrogen bond along this direction which decreases from O⋯O = 2.871 (3) to 2.775 (13) Å between ambient pressure and 4.8 GPa. The longer bond in this bifurcated motif decreases from 3.118 (3) to 2.981 (12) Å, which corresponds to an enhancement of the bifurcated character of this hydrogen bond. The crystal structure of α-glycine also contains head-to-tail chains of molecules and also in that structure an increase in bifurcation is also observed with pressure (Dawson et al., 2005). The crystallographic direction parallel to this chain in α-glycine is the least compressible in the system (Boldyreva et al., 2003), as it is here.

The hydrogen bonds in the OH⋯OH⋯OH chain formed by the side groups of the serine molecules are longer than those formed between the ammonium and carboxylate groups, and the O⋯O distances measure 2.918 (4) Å at ambient pressure. These interactions also decrease in length to 2.781 (11) Å. Brock & Duncan (1994) quote a range of 2.55–3.05 Å for the O⋯O distances in this type of interaction at ambient pressure, with an average of 2.79 (1) Å. The angles subtended at O3 in these chains increases from 148.5 (2)° at ambient pressure to 158.5 (7)° at 4.8 GPa.

We have recently described the crystal structure of α-glycine at 6.2 GPa (Dawson et al., 2005). The structure consists of a stack of hydrogen-bonded bi-layers which interact via CH⋯O hydrogen bonds. Within the layers sets of C(5) chains intersect to form R₄⁴(16) ring motifs. The effect of pressure was described in terms of the closing-up of holes in the middle of the R₄⁴(16) rings and the formation or shortening of CH…⋯O hydrogen bonds both across the R₄⁴(16) rings and between the bilayers. It is interesting to investigate whether similar features can be observed in the compression of L-serine-I.

Inspection of space-filling plots (Fig. 6a–f) shows that there are holes in the centres of each of the R₃³(11), R₃³(13) and R₃⁴(14) ring motifs formed in L-serine-I. The shortening of the NH⋯O hydrogen bonds (even N1H4⋯O2, which became very short) is not enough to close up the holes in the middle of the R₃³(11) and R³₄(14) rings, which occur in the A and B layers, respectively. Closure of the hole in the centre of the R₃³(13) ring does occur though, and as this closure occurs along the b-axis direction, the greater compressibility of the b axis compared with the a and c axes is understandable. The closure occurs not only by a shortening of the O3⋯O3 distance from 2.918 (4) to 2.781 (11) Å, but also an increase in the O3⋯O3⋯O3 angles made along the chains of hydroxyl groups from 148.5 (2) to 158.5 (7)° (the angle made by the vector between the central O3 and the midpoint of the two flanking O3s and [010] is 15.4°; Fig. 1b). As these appear to be comparatively weak hydrogen bonds this is presumably a rather `soft' parameter, which deforms easily under pressure.

Figure 6 Space-filling plots showing R-type graph sets which occur in L-serine phases I and II as a function of pressure. The top, middle and bottom rows correspond to the L-serine-I at ambient pressure, L-serine-I at 4.8 GPa and L-serine-II at 5.4 GPa, respectively. On the left of the diagram (a), (d) and (g) are all R3^3(11) motifs which occur in the A-layers (cf. Fig. 1); the hole in this small ring does not become very much smaller with increasing pressure. In the middle column of the diagram (b) and (e) show R^33(13) motifs, which also occur in the A-layers (cf. Fig. 1). Note that the hole in the middle of the ring is significantly smaller at 4.8 GPa than at ambient pressure. At 5.4 GPa (h) the R^33(13) motif of form I has been converted into R^23(13) in form II. On the right of the figure (c), (f) and (i) show the R^34(14) rings. Although the hole in the middle of this ring does not close-up in form I between ambient (c) and 4.8 GPa (f), it does in L-serine-II at 5.4 GPa (i).

CH⋯O interactions occur frequently in the structures of amino acids and in proteins (Desiraju & Steiner, 1999). A survey of amino-acid crystal structures determined by neutron diffraction showed that the most common H⋯O distances are around 2.4 Å, with a minimum of 2.15 Å (Jeffrey & Maluszynska, 1982). Generally, it is the H atom attached to the α-C atom which is involved in this type of interaction, as this is activated by the neighbouring ammonium and carboxy­late groups (Derewenda et al., 1995; Desiraju & Steiner, 1999); the H atoms of side chains are involved less frequently. L-Serine under ambient conditions does not conform to this general trend and under ambient conditions the strongest CH⋯O interactions are formed by the CH₂ group to the O atoms of neighbouring hydroxyl and carboxylate groups at normalized distances of 2.55 and 2.56 Å for C3H3⋯O1 and C3H2⋯O3, respectively. The shortest CH⋯O contact made by the αH atom is 2.75 Å at ambient pressure. CH⋯O hydrogen bonding from the C_αH group becomes more significant at high pressure and the normalized C2H1⋯O1 distance becomes 2.44 Å at 4.8 GPa. Most of the shortening in this interaction occurs between 0.3 and 2.9 GPa. The C3H3⋯O1 interaction shortens to 2.42 Å and together these interactions form a pair of contacts to the same O1 atom across the R³₄(14) rings in the B layers. These CH⋯O interactions can also be considered to support the N1H4⋯O1 interactions formed between the A layers.

Therefore, while the compression of α-glycine was characterized by the closing up of voids in R motifs with concomitant shortening of weak CH⋯O hydrogen bonds, that in L-serine-I is associated with the deformation and shortening of rather weak OH⋯OH hydrogen bonds. Although different interactions are involved in the two amino acids, both might be considered easily deformable. The formation of CH⋯O hydrogen bonds between the layers in the α-glycine structure is paralleled in L-serine-I by CH⋯O bond formation between the A layers (Fig. 3b).

One interesting conclusion of the, admittedly limited, research that has been carried out on hydrogen-bonded molecular systems is that super-short hydrogen bonds are not formed by the application of pressures below ca 10 GPa. The lower distance limits for such interactions which apply at ambient pressure also seem to apply at high pressure. In serine at 4.8 GPa at least one N⋯O distance (N1H4⋯O2) approaches the lower limit for this kind of interaction observed in the Cambridge Database. Above this pressure a phase change occurs to a hitherto uncharacterized phase, L-serine-II.

3.3. L-Serine-II at 5.4 GPa

The transition from L-serine-I to L-serine-II occurs with a marked reduction in the volume of the unit cell (Fig. 4). The volume per non-H atom in phase II is only 13.4 ų. Remarkably, the transition proceeds from one single crystal of L-serine-I to a single crystal of L-serine-II, and this transition is fully reversible.

The observation that this transition occurs from one single crystal to another strongly implies that the overall topologies of phases I and II are similar to each other. This proves to be the case, and the structure also consists of two sets of layers which are stacked along the a and b directions (Figs. 1c and 2c; hydrogen-bonding information is presented in Table 3). In terms of hydrogen bonds formed, the structure of the B layers is the same as in L-serine-I. Chains are formed by lattice repeats along c in which the molecules interact via three-centre N1H5⋯O1/2 bonds. In a reversal of the trend established during the compression of L-serine-I, these bonds are less symmetrical than the equivalent ones in L-serine-I at 4.8 GPa.

Table 3
Table of hydrogen-bonding parameters (Å, °) for L-serine-II at 54 GPa

H-atom coordinates have been normalized to neutron values.

	H⋯A	D⋯A	∠D—H⋯A
N1H5⋯O2^i	1.86	2.810 (14)	155
N1H5⋯O1^i	2.30	3.145 (11)	141
N1H4⋯O2^ii	1.89	2.850 (10)	159
N1H6⋯O1^iii	1.76	2.64 (2)	143
O3H7⋯O2^iii	1.71	2.62 (2)	152
C2H1⋯O1^v	2.50	3.059 (13)	111
C2H1⋯O2^v	2.37	3.411 (10)	162
C3H2⋯O2^i	2.40	3.149 (13)	125
C3H2⋯O3^iv	2.45	3.276 (15)	132
C3H3⋯O3^vi	2.37	3.207 (15)	133
C3H3⋯O1^v	2.42	2.982 (10)	111
Symmetry codes: (i) x,y,1+z; (ii) ; (iii) ; (iv) ; (v) ; (vi) .

Neighbouring chains are linked into a B-layer by N1H4⋯O2 hydrogen bonds, to build R³₄(14) motifs (Fig. 2c). These are completely analogous to those in the B-layers of L-serine-I, but the rings in L-serine-II are longer and thinner (cf. Figs. 2a–c). This change occurs by

• (i) opposite displacements of the molecules in successive N1H5⋯O2 chains along the c direction (i.e. the chains slide across each other), and

• (ii) compression of the distance between the chains along the a direction.

Thse changes enable compression of the R³₄(14) rings without further shortening the N1H4⋯O1 hydrogen bond. Inspection of the space-filling plots shows that the voids which occur in the centre of the R³₄(14) rings in L-serine-I even at 4.8 GPa close up during the phase transition (cf. Figs. 6c, f and i). The dimensions of the ring are equal to a/2 and c, and so the transition effects the lengths of the a and c unit-cell axes, the former sharply decreasing and the latter increasing slightly relative to L-serine-I.

As in L-serine-I, the structure of the A-layers consists of ribbons in which C(5) chains formed by N1H5⋯O2 hydrogen bonds are linked by N1H6⋯O1 hydrogen bonds (Fig. 1c). The N1⋯O1 distances R₃³(11) rings so-formed appear to be somewhat shorter than in L-serine-I at 4.8 GPa, although the standard uncertainties are high. There is a substantial change in the way the ribbons are connected into a layer. In L-serine-I this was achieved through OH⋯OH interactions, but in L-serine-II these are replaced by much stronger O3H7⋯O2 hydroxyl to carboxylate hydrogen bonds. This generates R²₃(13) motifs which replace the R₃³(13) motifs. The hydroxyl H7 atom was located in a difference-Fourier map and is clearly attached primarily to O3. Its position was also optimized in a plane-wave DFT calculation, with results consistent with those implied by the difference map (details of these calculations will be described in another publication). The length of the new hydrogen bond is very short, with an O3⋯O2 distance of 2.62 (2) Å, although the (normalized) H7…⋯O2 distance is 1.71 Å. In order to accommodate this interaction the N1—C2—C1—O2 torsion angle changes from −178.1 (2)° at 4.8 GPa to −156.3 (10)° at 5.4 GPa. The orientation of the O3H7 group changes from being gauche to anti with respect to C2—C3. This movement of the H atom implies that the C(5) chains which run along c must move apart slightly, with the result that the b axis is actually ca 0.5 Å longer in L-serine-II than in L-serine-I.

The formation of L-serine-II is also characterized by a marked increase in CH⋯O hydrogen bonding, with each H atom making two interactions. These occur within the A-layers between the CH₂ groups and the hydroxyl and carboxylate groups of neighbouring molecules involved in the new R²₃(13) ring motifs. In the B-layers they are formed across the R³₄(14) rings in the direction of the a axis; they can therefore be considered to stabilize the compression of these rings.

4. Conclusions

We have described the effect of high pressure on the crystal structure of L-serine. The structure can be considered to consist of two sets of layers, which stack along the a and b axes of the unit cell, and which have been referred to above as the A- and B-layers. The A-layers contain NH⋯O and OH…⋯OH interactions which combine to give R₃³(11) and R₃³(13) ring motifs; NH⋯O interactions in the B-layers form R³₄(14). The R₃⁴(14) motifs within the B-layers can also be viewed as connections between the A-layers. This structure remains stable up to 4.8 GPa. It undergoes anisotropic compression in which the principal structural effect is to compress voids in the middle of the R₃³(13) rings of the A-layers by deforming the rather `soft' hydrogen-bonded hydroxyl chains. The stacking distance between the A-layers also decreased, with a shortening of NH⋯O hydrogen bonds supported by the formation of CH⋯O hydrogen bonds. This latter effect continued until, at 4.8 GPa, the length of the NH⋯O hydrogen bonds approached the minimum value observed for this kind of interaction. Above 4.8 GPa a single-crystal-to-single-crystal phase change to L-serine-II occurs.

The phase change from L-serine-I to L-serine-II is accomplished by the change in two torsion angles and small positional displacements, and there are no major changes in the orientations of the molecules. The observation that the transformation occurs from one single crystalline form to another is therefore readily understood. In the new phase the hydrogen-bonded links in the A-layers between OH⋯OH groups are replaced by stronger, shorter OH⋯carboxyl interactions. The layers also move closer together by closing-up voids which occur in the centres of the R³₄(14) rings. All three H atoms which are attached to carbon take part in two CH⋯O interactions. The b and c axes are longer in L-serine-II than in L-serine-I, but the a axis is substantially shorter and the overall effect is a reduction in the volume of the unit cell. This reduction is ascribable to the closing-up of the voids in the R³₄(14) rings.

In L-serine high pressure closes up voids which occur in R motifs and decreases the interactions between layers by CH⋯O hydrogen-bond formation. Similar comments apply to the behaviour of glycine under pressure. We are currently investigating the effect of pressure on other α-amino acids and it will be interesting to discover to what extent these same effects apply in those systems.

Supporting information

S1. Refinement

022_ALERT_3_A Ratio Unique / Expected Reflections too Low ···. 0.39 027_ALERT_3_A _diffrn_reflns_theta_full (too) Low ············ 17.81 Deg. 029_ALERT_3_A _diffrn_measured_fraction_theta_full Low ······. 0.43 089_ALERT_3_B Poor Data / Parameter Ratio (Zmax. LT. 18) ···.. 4.24 201_ALERT_2_B Isotropic non-H Atoms in Main Residue(s) ······. 7 210_ALERT_3_B No Anisotropic ADP's Found in CIF ············.. ? 340_ALERT_3_B Low Bond Precision on C—C bonds (x 1000) Ang ··· 21 911_ALERT_3_B Missing FCF Refl. Between TH(Min) & STH/L=0.6.. 244 910_ALERT_3_C Missing FCF Reflections Below TH(Min) ·········. 1

The volume of reciprocal space accessible was severely limited by the pressure cell. This is the reason for all the alerts above. The non-H atoms were refined isotropically to minimize the number of parameters needed to model the data.

023_ALERT_3_B Resolution (too) Low [sin(th)/Lambda < 0.6]···.. 23.12 Deg. 020_ALERT_3_C The value of R_int is greater than 0.10 ········· 0.14

The resolution limit chosen for integration was based on inspection of the data collection images: there were no data beyond this limit. The data were quite weak because the size of the gasket hole meant that only a small crystal could be used; in addition the direct beam is attenuated by the Be and diamond anvils of the cell. Merging weak data yields high R_int values.

031_ALERT_4_B Refined Extinction Parameter within Range ······ 1.20 Sigma

Noted, but no action taken.

152_ALERT_1_C Supplied and Calc Volume s.u. Inconsistent ···.. ?

Reported 436.47 (15) Calculated 436.49 (13) Rounding error?

432_ALERT_2_C Short Inter X···Y Contact O1.. C3.. 3.01 A ng.

This is a CH···O contact.

731_ALERT_1_A Bond Calc 0.87 (13), Rep 0.881 (9) ······ 9.90 su-Rat O3 –H7 1.555 1.555 732_ALERT_1_B Angle Calc 106 (7), Rep 106.1 (9) ······ 7.78 su-Rat C3 –O3 –H7 1.555 1.555 1.555 731_ALERT_1_C Bond Calc 1.24 (2), Rep 1.234 (8) ······ 2.50 su-Rat C1 –O1 1.555 1.555 731_ALERT_1_C Bond Calc 1.246 (18), Rep 1.245 (8) ······ 2.25 su-Rat C1 –O2 1.555 1.555 731_ALERT_1_C Bond Calc 1.51 (3), Rep 1.507 (9) ······ 3.33 su-Rat C2 –C3 1.555 1.555 731_ALERT_1_C Bond Calc 1.484 (17), Rep 1.487 (8) ······ 2.13 su-Rat C2 –N1 1.555 1.555 732_ALERT_1_C Angle Calc 116.1 (14), Rep 116.1 (6) ······ 2.33 su-Rat C2 –C1 –O2 1.555 1.555 1.555

Checkcif uses only variances to calculate these values; the reported ones use the full variance-covariance matrix. Note that all these primary distances were subject to restraints.

Computing details

For all compounds, data collection: SMART (Siemens, 1993); cell refinement: SAINT (Siemens ,1995); data reduction: SAINT (Siemens ,1995); program(s) used to solve structure: USER DEFINED STRUCTURE SOLUTION; program(s) used to refine structure: CRYSTALS (Betteridge et al. 2003); molecular graphics: CAMERON (Watkin et al. 1996); software used to prepare material for publication: CRYSTALS (Betteridge et al. 2003).

Figures

(ser103) L-Serine

Crystal data

C3H7NO3	F(000) = 224
Mr = 105.09	Dx = 1.599 Mg m^−3
Orthorhombic, P212121	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: P 2ac 2ab	Cell parameters from 246 reflections
a = 8.5213 (13) Å	θ = 9–46°
b = 9.172 (2) Å	µ = 0.14 mm^−1
c = 5.5847 (8) Å	T = 293 K
V = 436.47 (15) Å^3	Block, colourless
Z = 4	0.20 × 0.10 × 0.10 mm

Data collection

Bruker SMART diffractometer	93 reflections with I > 2σ(I)
Graphite monochromator	Rint = 0.138
ω scans	θmax = 23.1°, θmin = 4.3°
Absorption correction: multi-scan SADABS (Siemens, 1996)	h = −9→9
Tmin = 0.695, Tmax = 1.00	k = −3→3
1169 measured reflections	l = −6→6
151 independent reflections

Refinement

Refinement on F^2	Hydrogen site location: inferred from neighbouring sites
Least-squares matrix: full	H atoms treated by a mixture of independent and constrained refinement
R[F^2 > 2σ(F^2)] = 0.083	P = P(6)*max(Fo^2,0) + (1-P(6))Fc^2 Method = SHELXL 97 (Sheldrick, 1997) W = 1. / [Sigma^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)Sin(theta)] P(i) are: 0.909E-01 2.37 0.00 0.00 0.00 0.333
wR(F^2) = 0.214	(Δ/σ)max = 0.000013
S = 1.07	Δρmax = 0.30 e Å^−3
140 reflections	Δρmin = −0.23 e Å^−3
33 parameters	Extinction correction: Larson 1970 Crystallographic Computing eq 22
28 restraints	Extinction coefficient: 60 (50)
Primary atom site location: structure-invariant direct methods

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	0.1122 (15)	0.195 (2)	−0.1399 (14)	0.038 (2)*
C2	0.0795 (11)	0.243 (2)	0.1165 (15)	0.038 (2)*
C3	0.0764 (11)	0.407 (2)	0.138 (2)	0.039 (3)*
N1	0.2010 (12)	0.184 (3)	0.2804 (17)	0.037 (3)*
O1	0.2254 (12)	0.114 (2)	−0.1768 (16)	0.039 (3)*
O2	0.0237 (10)	0.246 (3)	−0.2964 (16)	0.039 (3)*
O3	0.2262 (11)	0.468 (2)	0.076 (2)	0.042 (3)*
H7	0.260 (13)	0.51 (2)	0.205 (15)	0.0500*
H1	−0.0249	0.2028	0.1655	0.0461*
H2	0.0486	0.4329	0.3064	0.0476*
H3	−0.0044	0.4460	0.0263	0.0476*
H6	0.2033	0.0848	0.2676	0.0453*
H4	0.2953	0.2194	0.2406	0.0453*
H5	0.1775	0.2077	0.4319	0.0453*

Geometric parameters (Å, º)

C1—C2	1.524 (8)	C3—H2	0.999
C1—O1	1.234 (8)	C3—H3	0.995
C1—O2	1.245 (8)	N1—H6	0.909
C2—C3	1.507 (9)	N1—H4	0.896
C2—N1	1.487 (8)	N1—H5	0.897
C2—H1	1.003	O3—H7	0.881 (9)
C3—O3	1.434 (8)
C2—C1—O1	118.3 (6)	O3—C3—H2	110.171
C2—C1—O2	116.1 (6)	C2—C3—H3	108.635
O1—C1—O2	125.6 (6)	O3—C3—H3	109.114
C1—C2—C3	111.5 (8)	H2—C3—H3	109.948
C1—C2—N1	110.2 (6)	C2—N1—H6	109.508
C3—C2—N1	109.3 (8)	C2—N1—H4	109.677
C1—C2—H1	108.126	H6—N1—H4	109.023
C3—C2—H1	109.411	C2—N1—H5	109.541
N1—C2—H1	108.268	H6—N1—H5	108.973
C2—C3—O3	110.6 (8)	H4—N1—H5	110.101
C2—C3—H2	108.317	C3—O3—H7	106.1 (9)

(ser114) L-Serine

Crystal data

C3H7NO3	F(000) = 224
Mr = 105.09	Dx = 1.665 Mg m^−3
Orthorhombic, P212121	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: P 2ac 2ab	Cell parameters from 270 reflections
a = 8.4365 (10) Å	θ = 9–46°
b = 8.9506 (19) Å	µ = 0.15 mm^−1
c = 5.5512 (6) Å	T = 293 K
V = 419.18 (11) Å^3	Block, colourless
Z = 4	0.20 × 0.10 × 0.10 mm

Data collection

Bruker SMART diffractometer	100 reflections with I > 2σ(I)
Graphite monochromator	Rint = 0.127
ω scans	θmax = 23.3°, θmin = 4.3°
Absorption correction: multi-scan SADABS (Siemens, 1996)	h = −9→9
Tmin = 0.552, Tmax = 1.00	k = −3→3
1006 measured reflections	l = −6→6
146 independent reflections

Refinement

Refinement on F^2	Hydrogen site location: inferred from neighbouring sites
Least-squares matrix: full	H atoms treated by a mixture of independent and constrained refinement
R[F^2 > 2σ(F^2)] = 0.073	P = P(6)*max(Fo^2,0) + (1-P(6))Fc^2 Method = SHELXL 97 (Sheldrick, 1997) W = 1. / [Sigma^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)Sin(theta)] P(i) are: 0.291E-01 2.72 0.00 0.00 0.00 0.333
wR(F^2) = 0.164	(Δ/σ)max = 0.000008
S = 1.14	Δρmax = 0.37 e Å^−3
135 reflections	Δρmin = −0.27 e Å^−3
33 parameters	Extinction correction: Larson 1970 Crystallographic Computing eq 22
28 restraints	Extinction coefficient: 30 (20)
Primary atom site location: structure-invariant direct methods

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	0.1103 (13)	0.196 (3)	−0.1428 (13)	0.033 (2)*
C2	0.0759 (10)	0.246 (2)	0.1144 (14)	0.033 (2)*
C3	0.0724 (10)	0.414 (2)	0.135 (2)	0.034 (2)*
N1	0.1967 (11)	0.184 (3)	0.2815 (17)	0.033 (2)*
O1	0.2262 (11)	0.115 (2)	−0.1787 (15)	0.034 (2)*
O2	0.0192 (10)	0.245 (3)	−0.3007 (15)	0.034 (2)*
O3	0.2244 (10)	0.476 (2)	0.0760 (18)	0.036 (2)*
H7	0.279 (10)	0.48 (2)	0.211 (9)	0.0433*
H1	−0.0302	0.2067	0.1626	0.0389*
H2	0.0437	0.4431	0.3030	0.0402*
H3	−0.0087	0.4550	0.0204	0.0402*
H6	0.1984	0.0841	0.2689	0.0388*
H4	0.2928	0.2212	0.2432	0.0388*
H5	0.1724	0.2103	0.4338	0.0388*

Geometric parameters (Å, º)

C1—C2	1.525 (8)	C3—H2	0.999
C1—O1	1.235 (8)	C3—H3	1.002
C1—O2	1.246 (8)	N1—H6	0.896
C2—C3	1.510 (9)	N1—H4	0.902
C2—N1	1.485 (8)	N1—H5	0.901
C2—H1	0.998	O3—H7	0.881 (9)
C3—O3	1.436 (8)
C2—C1—O1	118.4 (6)	O3—C3—H2	109.102
C2—C1—O2	116.0 (6)	C2—C3—H3	109.306
O1—C1—O2	125.7 (6)	O3—C3—H3	108.932
C1—C2—C3	111.5 (8)	H2—C3—H3	109.385
C1—C2—N1	110.1 (6)	C2—N1—H6	109.552
C3—C2—N1	109.8 (8)	C2—N1—H4	109.379
C1—C2—H1	108.565	H6—N1—H4	109.604
C3—C2—H1	108.240	C2—N1—H5	109.405
N1—C2—H1	108.482	H6—N1—H5	109.725
C2—C3—O3	110.6 (8)	H4—N1—H5	109.162
C2—C3—H2	109.452	C3—O3—H7	106.1 (9)

(ser129) L-Serine

Crystal data

C3H7NO3	F(000) = 224
Mr = 105.09	Dx = 1.726 Mg m^−3
Orthorhombic, P212121	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: P 2ac 2ab	Cell parameters from 258 reflections
a = 8.3702 (10) Å	θ = 9–47°
b = 8.7699 (19) Å	µ = 0.15 mm^−1
c = 5.5103 (6) Å	T = 293 K
V = 404.49 (11) Å^3	Block, colourless
Z = 4	0.20 × 0.10 × 0.10 mm

Data collection

Bruker SMART diffractometer	100 reflections with I > 2σ(I)
Graphite monochromator	Rint = 0.119
ω scans	θmax = 23.3°, θmin = 4.4°
Absorption correction: multi-scan SADABS (Siemens, 1996)	h = −9→9
Tmin = 0.711, Tmax = 1.00	k = −3→3
1112 measured reflections	l = −6→6
134 independent reflections

Refinement

Refinement on F^2	Hydrogen site location: inferred from neighbouring sites
Least-squares matrix: full	H atoms treated by a mixture of independent and constrained refinement
R[F^2 > 2σ(F^2)] = 0.065	P = P(6)*max(Fo^2,0) + (1-P(6))Fc^2 Method = SHELXL 97 (Sheldrick, 1997) W = 1. / [Sigma^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)Sin(theta)] P(i) are: 0.597E-01 1.46 0.00 0.00 0.00 0.333
wR(F^2) = 0.154	(Δ/σ)max = 0.000009
S = 1.07	Δρmax = 0.26 e Å^−3
133 reflections	Δρmin = −0.24 e Å^−3
33 parameters	Extinction correction: Larson 1970 Crystallographic Computing eq 22
28 restraints	Extinction coefficient: 30 (30)
Primary atom site location: structure-invariant direct methods

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	0.1083 (11)	0.198 (2)	−0.1444 (12)	0.0310 (18)*
C2	0.0738 (9)	0.2493 (19)	0.1142 (13)	0.0301 (18)*
C3	0.0699 (9)	0.4206 (19)	0.1348 (19)	0.031 (2)*
N1	0.1952 (9)	0.186 (2)	0.2827 (14)	0.030 (2)*
O1	0.2265 (9)	0.118 (2)	−0.1806 (12)	0.033 (2)*
O2	0.0167 (8)	0.250 (3)	−0.3028 (13)	0.033 (2)*
O3	0.2226 (8)	0.484 (2)	0.0752 (15)	0.034 (2)*
H7	0.283 (7)	0.472 (19)	0.204 (12)	0.0414*
H1	−0.0329	0.2086	0.1621	0.0365*
H2	0.0411	0.4494	0.3045	0.0378*
H3	−0.0125	0.4615	0.0203	0.0378*
H6	0.1970	0.0834	0.2691	0.0363*
H4	0.2921	0.2235	0.2451	0.0363*
H5	0.1701	0.2116	0.4359	0.0363*

Geometric parameters (Å, º)

C1—C2	1.520 (7)	C3—H2	0.998
C1—O1	1.233 (8)	C3—H3	1.001
C1—O2	1.247 (8)	N1—H6	0.905
C2—C3	1.507 (9)	N1—H4	0.899
C2—N1	1.483 (8)	N1—H5	0.898
C2—H1	0.998	O3—H7	0.881 (9)
C3—O3	1.434 (8)
C2—C1—O1	118.2 (5)	O3—C3—H2	109.335
C2—C1—O2	115.7 (6)	C2—C3—H3	108.968
O1—C1—O2	126.0 (6)	O3—C3—H3	109.244
C1—C2—C3	111.6 (7)	H2—C3—H3	109.499
C1—C2—N1	110.3 (5)	C2—N1—H6	109.355
C3—C2—N1	109.8 (7)	C2—N1—H4	109.778
C1—C2—H1	108.259	H6—N1—H4	109.169
C3—C2—H1	108.467	C2—N1—H5	109.601
N1—C2—H1	108.305	H6—N1—H5	109.187
C2—C3—O3	110.6 (8)	H4—N1—H5	109.734
C2—C3—H2	109.148	C3—O3—H7	106.1 (9)

(ser141) L-Serine

Crystal data

C3H7NO3	F(000) = 224
Mr = 105.09	Dx = 1.764 Mg m^−3
Orthorhombic, P212121	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: P 2ac 2ab	Cell parameters from 258 reflections
a = 8.3266 (13) Å	θ = 9–47°
b = 8.665 (3) Å	µ = 0.16 mm^−1
c = 5.4851 (8) Å	T = 293 K
V = 395.75 (15) Å^3	Block, colourless
Z = 4	0.20 × 0.10 × 0.10 mm

Data collection

Bruker SMART diffractometer	103 reflections with I > 2σ(I)
Graphite monochromator	Rint = 0.080
ω scans	θmax = 23.4°, θmin = 4.4°
Absorption correction: multi-scan SADABS (Siemens, 1996)	h = −9→9
Tmin = 0.721, Tmax = 1.00	k = −2→2
1112 measured reflections	l = −6→6
130 independent reflections

Refinement

Refinement on F^2	Primary atom site location: structure-invariant direct methods
Least-squares matrix: full	Hydrogen site location: inferred from neighbouring sites
R[F^2 > 2σ(F^2)] = 0.066	H atoms treated by a mixture of independent and constrained refinement
wR(F^2) = 0.165	P = P(6)*max(Fo^2,0) + (1-P(6))Fc^2 Method = SHELXL 97 (Sheldrick, 1997) W = 1. / [Sigma^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)Sin(theta)] P(i) are: 0.672E-01 1.89 0.00 0.00 0.00 0.333
S = 1.06	(Δ/σ)max = 0.000012
129 reflections	Δρmax = 0.25 e Å^−3
32 parameters	Δρmin = −0.20 e Å^−3
28 restraints

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	0.1074 (11)	0.200 (2)	−0.1452 (11)	0.0304 (16)*
N1	0.1962 (10)	0.189 (3)	0.2816 (14)	0.0298 (19)*
O1	0.2270 (9)	0.119 (2)	−0.1830 (12)	0.0327 (18)*
C2	0.0726 (9)	0.251 (2)	0.1148 (13)	0.0295 (16)*
O2	0.0149 (8)	0.250 (3)	−0.3050 (12)	0.0315 (18)*
C3	0.0678 (9)	0.425 (2)	0.1345 (19)	0.0306 (18)*
O3	0.2211 (8)	0.490 (2)	0.0747 (15)	0.034 (2)*
H7	0.278 (8)	0.49 (2)	0.210 (9)	0.0410*
H1	−0.0340	0.2089	0.1647	0.0359*
H2	0.0387	0.4539	0.3051	0.0371*
H3	−0.0139	0.4657	0.0189	0.0371*
H6	0.1994	0.0856	0.2695	0.0359*
H4	0.2932	0.2281	0.2406	0.0359*
H5	0.1728	0.2162	0.4356	0.0359*

Geometric parameters (Å, º)

C1—O1	1.234 (8)	C2—C3	1.508 (9)
C1—C2	1.522 (7)	C2—H1	0.999
C1—O2	1.245 (8)	C3—O3	1.434 (8)
N1—C2	1.479 (8)	C3—H2	0.999
N1—H6	0.898	C3—H3	0.996
N1—H4	0.904	O3—H7	0.881 (9)
N1—H5	0.899
O1—C1—C2	118.5 (5)	N1—C2—C3	109.8 (7)
O1—C1—O2	125.3 (6)	C1—C2—H1	108.500
C2—C1—O2	116.0 (6)	N1—C2—H1	108.331
C2—N1—H6	109.818	C3—C2—H1	108.852
C2—N1—H4	109.277	C2—C3—O3	110.7 (8)
H6—N1—H4	109.294	C2—C3—H2	108.981
C2—N1—H5	109.518	O3—C3—H2	109.298
H6—N1—H5	109.766	C2—C3—H3	109.128
H4—N1—H5	109.152	O3—C3—H3	108.788
C1—C2—N1	109.9 (5)	H2—C3—H3	109.902
C1—C2—C3	111.4 (7)	C3—O3—H7	106.1 (9)

(ser148) L-Serine

Crystal data

C3H7NO3	F(000) = 224
Mr = 105.09	Dx = 1.789 Mg m^−3
Orthorhombic, P212121	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: P 2ac 2ab	Cell parameters from 247 reflections
a = 8.2980 (16) Å	θ = 9–44°
b = 8.600 (3) Å	µ = 0.16 mm^−1
c = 5.4663 (10) Å	T = 293 K
V = 390.09 (17) Å^3	Block, colourless
Z = 4	0.20 × 0.10 × 0.10 mm

Data collection

Bruker SMART diffractometer	99 reflections with I > 2σ(I)
Graphite monochromator	Rint = 0.083
ω scans	θmax = 23.2°, θmin = 4.4°
Absorption correction: multi-scan SADABS (Siemens, 1996)	h = −9→9
Tmin = 0.739, Tmax = 1.00	k = −2→2
1068 measured reflections	l = −6→6
129 independent reflections

Refinement

Refinement on F^2	Hydrogen site location: inferred from neighbouring sites
Least-squares matrix: full	H atoms treated by a mixture of independent and constrained refinement
R[F^2 > 2σ(F^2)] = 0.060	P = P(6)*max(Fo^2,0) + (1-P(6))Fc^2 Method = SHELXL 97 (Sheldrick, 1997) W = 1. / [Sigma^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)Sin(theta)] P(i) are: 0.00 1.64 0.00 0.00 0.00 0.333
wR(F^2) = 0.122	(Δ/σ)max = 0.000012
S = 1.16	Δρmax = 0.19 e Å^−3
128 reflections	Δρmin = −0.21 e Å^−3
33 parameters	Extinction correction: Larson 1970 Crystallographic Computing eq 22
28 restraints	Extinction coefficient: 44 (19)
Primary atom site location: structure-invariant direct methods

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	0.1058 (11)	0.200 (2)	−0.1471 (11)	0.0348 (18)*
C2	0.0718 (9)	0.251 (2)	0.1136 (13)	0.0344 (18)*
C3	0.0672 (9)	0.425 (2)	0.1345 (19)	0.0354 (19)*
N1	0.1961 (10)	0.187 (3)	0.2811 (14)	0.034 (2)*
O1	0.2259 (9)	0.119 (2)	−0.1860 (13)	0.037 (2)*
O2	0.0141 (8)	0.254 (3)	−0.3062 (12)	0.035 (2)*
O3	0.2207 (8)	0.490 (2)	0.0730 (14)	0.038 (2)*
H7	0.287 (5)	0.460 (17)	0.190 (16)	0.0459*
H1	−0.0355	0.2084	0.1638	0.0412*
H2	0.0390	0.4544	0.3061	0.0424*
H3	−0.0161	0.4669	0.0200	0.0424*
H6	0.1989	0.0833	0.2686	0.0411*
H4	0.2932	0.2269	0.2407	0.0411*
H5	0.1722	0.2145	0.4361	0.0411*

Geometric parameters (Å, º)

C1—C2	1.517 (7)	C3—H2	1.000
C1—O1	1.230 (8)	C3—H3	1.000
C1—O2	1.246 (7)	N1—H6	0.898
C2—C3	1.504 (9)	N1—H4	0.901
C2—N1	1.482 (8)	N1—H5	0.901
C2—H1	0.999	O3—H7	0.882 (9)
C3—O3	1.430 (8)
C2—C1—O1	118.3 (5)	O3—C3—H2	109.256
C2—C1—O2	115.7 (6)	C2—C3—H3	109.281
O1—C1—O2	125.7 (6)	O3—C3—H3	109.147
C1—C2—C3	111.3 (7)	H2—C3—H3	109.486
C1—C2—N1	110.2 (5)	C2—N1—H6	109.571
C3—C2—N1	109.6 (7)	C2—N1—H4	109.451
C1—C2—H1	108.608	H6—N1—H4	109.503
C3—C2—H1	108.510	C2—N1—H5	109.461
N1—C2—H1	108.546	H6—N1—H5	109.551
C2—C3—O3	110.3 (8)	H4—N1—H5	109.291
C2—C3—H2	109.328	C3—O3—H7	106.2 (9)

(ser254) L-Serine

Crystal data

C3H7NO3	F(000) = 224
Mr = 105.09	Dx = 1.865 Mg m^−3
Orthorhombic, P212121	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: P 2ac 2ab	Cell parameters from 313 reflections
a = 6.9083 (10) Å	θ = 6–46°
b = 9.644 (3) Å	µ = 0.17 mm^−1
c = 5.6166 (8) Å	T = 293 K
V = 374.19 (14) Å^3	Block, colourless
Z = 4	0.20 × 0.10 × 0.10 mm

Data collection

Bruker SMART diffractometer	90 reflections with I > 2σ(I)
Graphite monochromator	Rint = 0.081
ω scans	θmax = 23.3°, θmin = 4.2°
Absorption correction: multi-scan SADABS (Siemens, 1996)	h = −7→7
Tmin = 0.687, Tmax = 1.00	k = −3→3
1990 measured reflections	l = −6→6
140 independent reflections

Refinement

Refinement on F^2	Primary atom site location: structure-invariant direct methods
Least-squares matrix: full	Hydrogen site location: inferred from neighbouring sites
R[F^2 > 2σ(F^2)] = 0.048	H atoms treated by a mixture of independent and constrained refinement
wR(F^2) = 0.102	P = P(6)*max(Fo^2,0) + (1-P(6))Fc^2 Method = SHELXL 97 (Sheldrick, 1997) W = 1. / [Sigma^2(F^2) + (P(1)p)^2 + P(2)p + P(4) + P(5)Sin(theta)] P(i) are: 0.219E-01 1.16 0.00 0.00 0.00 0.333
S = 1.21	(Δ/σ)max = 0.000024
121 reflections	Δρmax = 0.25 e Å^−3
32 parameters	Δρmin = −0.20 e Å^−3
15 restraints

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	0.1434 (14)	0.2142 (12)	−0.0410 (11)	0.023 (3)*
C2	0.0920 (11)	0.2511 (13)	0.2148 (13)	0.019 (3)*
C3	0.1205 (10)	0.4038 (14)	0.2677 (17)	0.021 (3)*
N1	0.2191 (11)	0.1688 (18)	0.3767 (13)	0.024 (3)*
O1	0.2179 (10)	0.0996 (14)	−0.0761 (11)	0.030 (3)*
O2	0.1037 (9)	0.3009 (14)	−0.2006 (13)	0.027 (2)*
O3	0.3257 (8)	0.4328 (18)	0.2748 (14)	0.034 (3)*
H1	−0.0463	0.2258	0.2453	0.0235*
H2	0.0609	0.4269	0.4249	0.0251*
H3	0.0582	0.4606	0.1400	0.0251*
H4	0.3436	0.1925	0.3505	0.0299*
H5	0.1876	0.1867	0.5282	0.0299*
H6	0.2034	0.0780	0.3458	0.0299*
H7	0.3367 (19)	0.519 (5)	0.33 (2)	0.0414*

Geometric parameters (Å, º)

C1—C2	1.522 (7)	C3—H2	0.999
C1—O1	1.234 (9)	C3—H3	1.000
C1—O2	1.256 (9)	N1—H4	0.902
C2—C3	1.514 (9)	N1—H5	0.895
C2—N1	1.492 (8)	N1—H6	0.900
C2—H1	1.001	O3—H7	0.881 (6)
C3—O3	1.446 (8)
C2—C1—O1	117.2 (6)	O3—C3—H2	109.645
C2—C1—O2	117.8 (6)	C2—C3—H3	109.680
O1—C1—O2	124.9 (6)	O3—C3—H3	109.591
C1—C2—C3	112.5 (7)	H2—C3—H3	109.516
C1—C2—N1	108.3 (6)	C2—N1—H4	109.079
C3—C2—N1	108.7 (7)	C2—N1—H5	109.511
C1—C2—H1	109.067	H4—N1—H5	109.759
C3—C2—H1	109.131	C2—N1—H6	109.227
N1—C2—H1	109.127	H4—N1—H6	109.330
C2—C3—O3	108.7 (8)	H5—N1—H6	109.917
C2—C3—H2	109.668	C3—O3—H7	106.0 (3)

 

Footnotes

¹Supplementary data for this paper are available from the IUCr electronic archives (Reference: WS5017 ). Services for accessing these data are described at the back of the journal.