R. J. Davis, A. M. Clark, A. J. Criddle
1977 Mineralogical magazine  
Natural Histor~J), Cromwell ROad, london S%7 5BD Clark et___.~. (19745 reported on the first part of ~n e~ination of the reltdu~/ concentrates from the gold ~rkiags at Itabire, Mints Gerais, Brazil~ contained in specimen number ~4 1934,72. Pd33.?se30,O or 2[Pe17Se15] , confi~e4 By Gell~ 11962) from the crystal st~cture. Matthias and Geller (1958) , working with chilled melts of composition Pdx_Se found that only the cubic patte~ Ms obtained for x = 1.O, 1.1, or 1.2 ; llke Gronvold and Rost they
more » ... nvold and Rost they found that for _x = 1.3 or more, the powder pattern showed traces of Pd2,8 ~. Schubert et el. (19973 gave cell dimensions for PriSe with the NiAs-type structure, finding it in equilibri~ with Pd1~e13 on one side and With P~Se 2 on the other. These obser~tio~ c~ be reconciled by a phase dish,ram ~ u that sketched in fig 3. The ~it~ell dime~Ions listed in Table IV show that for Pd1~e15 three determinations in two different laboratori~ agree to within 0.002~, well within the experimental errors. However the fsmrtn result (by Schubert et 81., 1937) is significantly l~ger than the other three and agrees well with our re~l~s. We have ind~ed and ~trapolated Gronvold and Host's data independently ~d ha~ checked the factor used to convert kX to ~; all fi&n1~es in Table IV are accurate am claimed and the discrepancy ~st he e~plalned By ass~ing PdlTSe15 is a non-stochi~etric compound exhibiting a small range of c~positione. that, of the real ~d ur,equal solutions Ib/rnlshed by (8) the s~lutlons of a(be) and A(s*y*} to accept~e those obtained from the positive square root, whereas to ~rlve8ta(b*o*) the negative of the square root has to be used. CGZeuZ~tion of 0~. In a~dition to a straight llne iso-d, a straight line iso-Or also p~ses through each point P(X,y) of the quedrilgter~l PIPTP3P h ( fig. i) . The equation of the straight llne leo-Or that pzsses through P(=,y) ~d intersects sides PIP~ ~d P Ph at the points P~{=3,~3) ~d By expressing the coordinates of the points P' ~d P~ as functions of 0~, 0rl3 , OP2~ ~d of the coordinates of the quadrilateral vertices, we arrive at a quadratic fo~ with respect to Or: (33.60 + 0.84) x 1021 within the ~gnetlte. A total of (39.72 -+ 1.14) x 1021 atoms, of which (5.61 -+ 0.16) x 1021 are T~, ~re ca~ated for the magnetite producing a Ti/(Ti + Fe) rat[o of 0.141 ~ 0.004 and leading to a ~lecule contalning 42.3 _+ 1.f2~ ulvCspinel. At the conditions of equilibration of the granular il~nite a change of ~0~ ulvbspinel corresponds to a 29~ change In temperature and leads to a temperature of 930 -+ 3.2~ Because the lines on the 8uddington and Llndsley diagram intersect at a I~ angle,an error on one set of lines influences estlmate~ of error on the other axis
doi:10.1180/minmag.1977.041.320.37 fatcat:rchaa6xpzffzhcherhga4sypsu