§ These peaks have noticeable tails due to axial divergence: the peak positions given are the "true" value.
* This peak has a noticeable tail on the high-angle (i.e. right-hand) side.

The tails should not influence the positions of the reflections, I took the values as the accurate values. Seeing the pattern I would say that the compound crystalises in a lower symmetry crystal system (quite complex set of reflections), with relatively larger size of unit cell (the first reflection at only 7.5° (2q).

Because zone indexing is a method that works well for low-symmetry systems since it makes no prior assumptions about degrees of symmetry, and it has also the advantage of lower sensitivity to impurity reflections in the powder pattern than other methods, the program FZON seems to be good choise.

The first step in indexing process is the determination of the unit-cell geometry. For the first indexing trial I choose the set of first 20 reflections (no. 1 - 20) since the indexing is more useful for lower 2q values- even though they provide less accurate lattice parameters. (if erroneous a-parameter is used, this results in a very large position error for peaks at high scattering angles and this position error can lead to erroneous indexing. Further, at lower 2 theta there are lower number of reflections, and so the possibility of erroneous indexing decreases). The result is here (1st run). The indexing with this input file failed: the M₂₀ values are too low for all cases, combining with the low number of indexed reflections.

Then the strategy I accepted was to exclude gradually one reflection, starting from the higher 2q values and going to lower 2q values. The idea was that there could be an impurity peak(s) or the position of low intensity reflections has not been accurately estimated (the FZON program is told to be suited for this type of work). The results are here ( 2nd run, excluded is the reflection No.17), here (3rd run, excluded is the reflection No.13), and here (4th run, excluded is the reflection No.9).

The far best result (20 indexed reflections with M₂₀ larger then 17) has been observed with the excluded reflection at 20 .84°(2q)(see also the fifth column in the table). The material seems to crystalise in primitive monoclinic lattice (parameters a = 23.743 A, b = 6.181 A, c = 11.031 A, a = 90°, b = 97.83°, g = 90°, (if I understood well the standard setting has been used, an axis of order 2 parallel to the b direction so that a = g = 90°) V = 1603 A³)). The value of beta is not too close to 90°, so the tetsts for possible orthorombic description due to the experimental error can be ommited. The remaining two outputs gave markedly worse results.

The next step I did was the generation of hkl indeces by the hkl-www program using the input data received above (P lattice, a = 23.743 A, b = 6.181 A, c = 11.031 A, a = 90°, b = 97.83°, g = 90°). The output is here, and the 2q(°) and hkl indeces obtained are in the 6th and 7th column of the table. Comparison between observed and calculated Q values for the angles up to approx. 17° is given in the next figure.

Many calculated reflections were not included in the indexing (were not observed), the  computed reflections account for all the of observed ones. Comparing the 2q values, I can see that the precision of indexing is acceptable to the angle about 15-16° (2q). Afterthat the differencies between calculated and observed lines grow, and at angles above 25°(2q) problems reveal to assign unambiguously the hkl indeces to observed 2q values. For the to determination of the absence or presence of lattice reflection conditions only the reflections up to 16° have been considered.

The next step is to find out whether any reflection condition exist: firstly for centred lattices, then for glide planes, and lastly for screw axes. Seeing the table of 230 space groups, the monoclinic system contains 1 Laue class, 3 crystal classes, each with P and C centering, and 13 space groups. The Laue class of our material is 2/m and knowing that in the monoclinic system only two types of lattice are possible I can easily test the pattern for centering.  When I used the C centering in the hkl-www generation program, I obtain this output. Comparing hkl output data with observed ones I can see that the material can not be C-centred because the observed reflection at 16.64°(2q, indexed as 011) breakes the general reflection condition for C centering hkl: h+k = 2n.
The reflections, which break the reflection condition for centering (either A, C, or I) are  marked with star in the 8th column of the table.
The choise of space groups narrows from 13 to 8 possibilities: P2, P2₁, Pm, Pc, P2/m, P2₁/m, P2/c, and P2₁/c.

The next step is to look for glide planes. As the previous search, it is important to consider reflections that do not have all even integer Miller indices. Presence of reflection 011 positively tells me that the lattice cannot have b-glide, c-glide, or d-glide plane perpedicular to a since the reflection condition 0kl: k = 2n, l = 2n, k+l = 4n, are broken, respectively. The presence of reflection 001 positively tells me that the lattice cannot have n-glide perpedicular to a,or c-glide perpedicular to b since the reflection condition 0kl: k + l = 2n, and h0l: l = 2n are broken, respectively. The presence of reflection 102 positively tells me that the lattice cannot have a-glide n-glide, and d-glide planes perpedicular to b, or n-glide plane perpedicular to b since the reflection conditions h0l: h = 2n, h0l: h +l = 2n, h0l: h + l = 4n are broken, respectively. The presence of reflection 110 positively tells me that the lattice cannot have a-glide, b-glide, or d-glide plane perpedicular to c since the reflection condition hk0: h = 2n, hk0: hk0: h + k = 4n, , are broken, respectively. The presence of reflection 410 positively tells me that he lattice cannot have n-glide plane perpedicular to c since the reflection condition hk0: h +k = 2n is broken.
This result narrows the number of possible space group to 5: P2, P2₁, Pm, P2/m, and P2₁/m.

 Using the first 13 reflections (including 011- 2q position 16.44°), I obtained this output from www program. Using the last 16 reflections I obtained this output.

Because a routine powder diffraction pattern was  collected using a well-aligned Bragg - Brentano diffractometer and Cu Ka1 radiation (l = 1.54056 A) it seems to me that the the value of 2q zero offset is still acceptable. It could be explained partially by the effect of (high) transparency of the sample (organic compound with higher content of ligth elements) which could cause that reflections will be seen by the detector at a slightly smaller 2q angle. which should cause the sytematic shift of positions of the reflections. Nevertheless, this is not observed for this paricular case. Taking into account usage of low angle input data, relative errors in lattice parameters seem to me also acceptable. Normally (using both low, and high angle data) the ESD of a, b and c should be one order lower. (When I used the last 16 reflections the ESD of lattice parameters  has been lower.)
 

Estimated standard deviations are usually obtained by including higher angle data. I think that there are the following reasons why this is difficult for this particular data set:

1. the width of the Bragg reflections at larger scattering angles is generally increasing with increasing angle, while at the same time the intensity of the peaks is diminishing due to the effects of X-ray atomic factors and thermal vibration. The result may be that the positions of the high-angle peaks are determined less precisely than those at lower angles.
2. the result of indexing of reflections for the higher angle reflections is particularly unwarranted (assignement hkl triplets to high-angle data based on poor values of the lattice parameters leads to increased probability of assigning incorrect values; also the density of reflections lower from its maximum at 2q = 90° towords to 0° and 180° so that low-angle reflections are easier to index).

   ---