QPA with Partial or No Known Crystal Structures (PONKCS) Innovation with Integrity

Classic Rietveld method IF All phases in the mixture are known All phases are crystalline All crystal structures are known THEN ZMV algorithm (1) can be used to relate the Rietveld scale factor to the phase concentration (1) Hill R.J. and Howard C.J. (1987) J. Appl. Crys. Vol 20, 467-474. 14-15/12/2011 Advanced XRD Workshop 2

Classic Rietveld method ZMV calibration constant W α = n S k = 1 α S ( Z MV ) k α ( Z MV ) k W = weight % S = Rietveld scale factor Z = No. of formula units in unit cell M = molecular mass of formula unit V = unit cell volume ZMV is a calibration constant which eliminates the need to measure: instrument calibration constant sample mass absorption If the crystal structure of a phase is not known, then the ZMV calibration constant is unknown 14-15/12/2011 Advanced XRD Workshop 3

Classic Rietveld method internal standard The sample is "spiked" with a known mass of standard material and the QPA normalized accordingly The weight fractions of the crystalline phases present in each sample are estimated using the Rietveld methodology Concentrations to be corrected proportionately according to: Corr( W α ) = W STD STD known measured where Corr(W α ) is the corrected weight percent, STD known the weighted concentration of the standard in the sample and STD measured the analyzed concentration α The amount of amorphous material W amorphous can then be derived from: W = 1 If there is more than one phase with an unknown or only partially known structure, these phases will be considered as a single group = no info on individual phases! amorphous W j j= 1 n Corr( 14-15/12/2011 Advanced XRD Workshop 4 )

PONKCS method An individual unknown or partially unknown crystalline phase can be quantified through PONKCS method, if: the phase is available as pure specimen (or major phase in specimen known impurities that can be quantified with the internal standard method) a synthetic mixture can be prepared in which the amounts of unknown and an internal standard are known the unknown phase in the experimental mixture does not vary too much from the pure specimen, used to derive the intensities Phases with Partial Or No Known Crystal Structure are characterized by measured rather than calculated structure factors For all phases a using empirically derived structure factors ZMV "calibration constants" must be derived, e.g. via an internal standard s 14-15/12/2011 Advanced XRD Workshop 5

PONKCS method crystalline (and even amorphous) phases can be fitted through: (hkl) phase (Pawley or Le Bail) if the phase can be indexed peak phase if the phase cannot be indexed Result: list of peaks described by peak position and intensity This list of peaks replaces str in the Rietveld refinement, while keeping peak position and intensity fixed = group of peaks scaled as a single entity within Rietveld refinement ZMV constant is unknown and has to be provided through calibration 14-15/12/2011 Advanced XRD Workshop 6

PONKCS method Calibration through the preparation of a mixture containing known amounts of unknown (α) and standard material (s) W W α = s S S α s ( Z MV ) ( Z MV ) s α or W W ( ) α s Z MV = ( Z MV ) s α s S S α all known (ZMV) α has NO physical meaning 14-15/12/2011 Advanced XRD Workshop 7

PONKCS Applications in Industry "Nontronite Unknown structure Cement: QA of Bassanite Structure data of C3S do not accurately describe the technical product Cement: QA of blast furnace slag (BFS) Amorphous material (BFS) Cement: QA of C3S polymorphs (M1, M3) Structure data of C3S do not accurately describe the technical product 14-15/12/2011 Advanced XRD Workshop 8

Nontronite 14-15/12/2011 Advanced XRD Workshop 9

#1: Literature Diffraction pattern of nontronite (Co Kα) 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 10

#1: Refinement of standard mixture Sqrt(Counts) 115 110 105 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 2Th Degrees 54 56 58 60 62 64 66 68 70 72 74 76 78 Corundum 49.78 % Nontronite_hkl 50.22 % 80 82 84 86 88 90 11

#1: Results The resultant phase model has been applied to a series of mixtures of known composition 2.0 1.5 Nontronite Corundum 1.0 Bias (wt%) 0.5 0.0-0.5 0 10 20 30 40 50 60 70 80 90 100-1.0-1.5-2.0 Weighed (wt%) 12

Cement 14-15/12/2011 Advanced XRD Workshop 13

PONKCS Example: Quantitative Analysis of Cement Nishi & Takeuchi (1985) C3S structure - misfits due to structural deficiencies Only a "cosmetic" issue for Clinker quantification Counts 8,000 7,500 7,000 6,500 6,000 5,500 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0-500 -1,000-1,500-2,000-2,500-3,000 10 15 20 25 30 35 40 2Th Degrees 45 50 C3S monoclinic (NISHI) 70.69 % C2S beta (MUMME) 8.02 % C3A Na orthorhombic 5.02 % C3A cubic 2.76 % C4AF Colville 10.20 % Periclase 0.63 % Lime 1.24 % Arcanite K2SO4 1.44 % 14-15/12/2011 Advanced XRD Workshop 14 55 60 65

PONKCS Example: Quantitative Analysis of Cement Nishi & Takeuchi (1985) C3S structural deficiencies (intensity misfit) adversly affects Bassanite (CaSO 4 ½H 2 O) quantification due to unevitable peak overlap Degree of overlap depends on Alite chemical composition Counts 3,000 2,500 2,000 1,500 1,000 500 0-500 -1,000 Alite (402) & (310) Bassanite C2S beta (MUMME) 0.00 % C3A cubic 0.00 % C3A Na orthorhombic 0.00 % C4AF Colville 0.00 % Lime 0.00 % Periclase 0.00 % Quartz 0.00 % Arcanite K2SO4 0.00 % Gypsum 0.00 % Bassanite Bezou 0.00 % Anhydrite 0.00 % Calcite 0.00 % Portlandite 0.00 % C3S monoclinic (NISHI) 0.00 % -1,500-2,000-2,500-3,000-3,500 10 12 14 16 18 20 22 24 26 2Th Degrees 28 30 32 34 36 38 40 14-15/12/2011 Advanced XRD Workshop 15

PONKCS Structureless Modelling of Alite Substitution of calculated intensities phase_name C3S_hkl_Phase scale @ 0.1 MVW( 0, 4331.778506, 0) space_group 8 CS_L(cs_c3smonni, 1000 min =50; max =1000;) a a_c3smonni 33.18748 b b_c3smonni 7.05051 c c_c3smonni 18.56319 be be_c3smonni 94.22338 r_bragg 0.006728277788 hkl_is hkl_m_d_th2 0 0 2 2 9.25639439 9.54711437 I 2.05270142e- 010 hkl_m_d_th2 2 0-2 2 8.34458351 10.5931587 I 0.01858485616 hkl_m_d_th2 4 0 0 2 8.27434444 10.683341 I 0.01406196859 hkl_m_d_th2 2 0 2 2 7.8363862 11.2822943 I 0.01484194695 hkl_m_d_th2 4 0-1 2 7.77032948 11.3785229 I 0.0251988569 hkl_m_d_th2 4 0 1 2 7.35505056 12.0232773 I 0.03432590883 hkl_m_d_th2 1 1 0 4 6.89578199 12.8272991 I 0.01149107085 hkl_m_d_th2 1-1 -1 4 6.49472332 13.6230459 I 0.04729996279 hkl_m_d_th2 1 1 1 4 6.42985344 13.7611485 I 0.01753635023 hkl_m_d_th2 4 0-2 2 6.40786076 13.8086081 I 0.03122114305 hkl_m_d_th2 0 0 3 2 6.17092943 14.3415194 I 0.003343324237 hkl_m_d_th2 4 0 2 2 5.9548769 14.8647385 I @ 0.27078 hkl_m_d_th2 3 1 0 4 5.94094992 14.8997831 I @ 1.2198 hkl_m_d_th2 2 0-3 2 5.92666531 14.9358988 I 0.007612890057 hkl_m_d_th2 3-1 -1 4 5.72322845 15.4700003 I 0.003767987738 hkl_m_d_th2 2 0 3 2 5.64745855 15.6788511 I 0.01085298472.. Counts Counts 1,000 900 800 700 600 500 400 300 200 100 1,000 0 10 900 800 700 600 500 400 300 200 100 0 10 11 Classic Rietveld Refinement using the Nishi & Takeuchi (1985) C3S structure 11 12 12 13 13 14 15 16 2Th Degrees 14 15 16 2Th Degrees PONKCS refinement 17 17 18 18 19 19 20 20 14-15/12/2011 Advanced XRD Workshop 16

PONKCS Quantitative Analysis of Bassanite Classic Rietveld Refinement using the Nishi & Takeuchi (1985) C3S structure PONKCS refinement Bassanite: Estimated accuracy ~0.2% Estimated precision <0.1% 14-15/12/2011 Advanced XRD Workshop 17

Slag in Cement 14-15/12/2011 Advanced XRD Workshop 18

PONKCS Quantitative Analysis of BFSlag BFSlag can be described via Pawley / Le Bail or single peak fitting 14-15/12/2011 Advanced XRD Workshop 19

PONKCS Quantitative Analysis of BFSlag E.g. Pawley fit strategy: Chose arbitrary space group and lattice parameters to allow for a couple of peaks Allow for extreme line broadening Refined PONKCS model for BFSlag 14-15/12/2011 Advanced XRD Workshop 20

Round Robin VDZ 2006/7 Quantitative Analysis of BFSlag Preliminary results PONKCS: (5 repeated analyses) 25.1 (2) 67.2 (2) 71.7 (2) 14-15/12/2011 Advanced XRD Workshop 21

Round Robin VDZ 2006/7 Quantitative Analysis of BFSlag Accuracy and Precision (values in wt. %, SD in brackets/1σ) Every sample measured 5 times (D4 ENDEAVOR, LynxEye Detector) Reference values: sample 1: 25,0 wt.% sample 2: 67,0 wt.% sample 3: 72,0 wt.% BFSlag Sample 1 BFSlag Sample 2 BFSlag Sample 3 Measurement 1 25,0 67,2 71,7 Measurement 2 25,1 67,3 71,9 Measurement 3 24,7 67,0 71,6 Measurement 4 25,1 67,3 71,9 Measurement 5 25,3 67,0 71,5 Mean 25,1 67,2 71,7 SD 0,2 0,2 0,2 14-15/12/2011 Advanced XRD Workshop 22

C3S Polymorphs 14-15/12/2011 Advanced XRD Workshop 23

PONCKS Alite Polymorphism The missing link for quality and process control In Cement Clinkers two monoclinic polymorphs of the main Phase Alite (C 3 S) can be found (M1 and M3) Both polymorphs are different in hydraulic properties, like strength development Although this is well know since decades, nobody was able to distinguish and quantify these polymorphs in the last years by XRD Rietveld analysis 14-15/12/2011 Advanced XRD Workshop 24

PONCKS Alite Polymorphism The monoclinic polymorphs M1 and M3 Very similar patterns Available structural data do not accurately describe the technical product Alite polymorph M1 Alite polymorph M3 Lin (Counts) 10000 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 Lin (Counts) 10000 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 50 51 52 53 54 50 51 52 53 54 M1 and M3 can be distinguished using currently available structure data! 14-15/12/2011 Advanced XRD Workshop 25

Alite Polymorphs Monitoring Compressive Strength Observation over months: continuous decrease of the early strength development, although all process control parameters remained unchanged The re-evaluation of the XRD data using TOPAS considering the Alite polymorphs showed a correlated decrease of the M1 polymorph 14-15/12/2011 Advanced XRD Workshop 26

Alite Polymorphs POLYSNAP + TOPAS QA Results PCA: 630 PXRD + M1/M3 alite HS clinker M1 high HS clinker M1 low normal clinker M1 high normal clinker M1 low 14-15/12/2011 Advanced XRD Workshop 27

PONCKS Alite Polymorphism The missing link for quality and process control By using the TOPAS PONKCS capabilities we are able to distinguish the Alite polymorphs M1 and M3 Provides more detailed information about the process Allows to understand and predict the early strength development Allows to control and improve the quality of the product 14-15/12/2011 Advanced XRD Workshop 28

PONKCS Reference Scarlett, N.V.Y. & Madsen, I.C. (2006) Quantification of phases with partial or no known crystal structure Powder Diffraction, 21(4), 278-284 Note: The paper includes an example walk-through on a TOPAS keyword level 14-15/12/2011 Advanced XRD Workshop 29

Conclusions PONKCS allows the quantification of compounds, where the classic Rietveld method fails Materials with partial or no known crystal structure can be quantified with the same accuracy and precision as phases with crystal structure. Often even better, due to the calibration step Crystalline and amorphous phase amounts, accuracy and LLOD < 1% Preferred orientation can be dealt with (if sample is indexed) 14-15/12/2011 Advanced XRD Workshop 30