6.2 A Simultaneous Measurement of the Strong Coupling Constant and the Colour Factors 95 6.2. Corrections for the Four-Jet Observables Background plus Hadronization Corrections The theoretical NLO prediction as obtained from DEBRECEN should be corrected for background and hadronization effects. The correction procedure was detailed in Section 5.3.. In Fig. 6.9 the background plus hadronization corrections as obtained from PYTHIA and HERWIG are shown. The background plus hadronization corrections from the HERWIG simulation show large discrepancies with respect to the ones coming from PYTHIA. The showering and hadronization parameters used, both in HERWIG and PYTHIA, are the standard ones, which were obtained by the tuning of MC simulations starting from q μq configurations. Therefore this may be an indication of the non-universality of these parameters. If a better tuning cannot be found, then the reason for the discrepancies could be due to the implementation of the showering and hadronization processes in the MC programs. More detailed studies about the four-parton Monte Carlo programs are presented in Chapter 7. Detector Corrections The theoretical prediction for the four-jet angular correlations, already corrected for hadronization effects, has to be corrected further to include detector effects before being compared to ALEPH data. This is done by computing these observables from the MC before and after simulation, and the corrections are found in Fig. 6.. In the case of cos ff 34 the detector corrections calculated from the simulation with the PYTHIA fourparton option passed through the detector simulation is also shown. This, as explained in Section 5.3.2, is the one used for the systematic uncertainty estimation. Total Corrections Taking into account the hadronization and detector corrections as explained in the previous chapter, the total corrections for each four-jet observable can be constructed as C tot (i bin )=C bck+had (i bin ) C det (i bin ) : (6.2) Figure 6. shows the total bin-by-bin corrections for the angular correlations when using PYTHIA for the hadronization corrections. Typically such corrections are found within the 5-% range. The corrections for the four-jet rate can be found in Fig. 6.3.
96 Measurements of the Strong Coupling Constant and the Colour Factors Bck + Had Correction Bck + Had Correction.5..5.95.9 Pythia 6. Herwig 6. 5.2.4.6.5..5.95.9 Cosχ BZ 5.2.4.6 CosΘ NR Bck + Had Correction Bck + Had Correction.5..5.95.9 5 - -.5.5.5..5.95.9 CosΦ KSW 5 - -.5.5 Cosα 34 Figure 6.9: Background and hadronization corrections for the four-jet angular correlations. 6.2.2 Results An experimental covariance matrix is calculated to take into account the statistical error of the data, the statistical errors of the detector and hadronization corrections, and the binby-bin statistical correlations among the different observables as well as the correlations between the bins of a single observable. Tables 6.3-6.6 show thedata distributions for the four-jet angular correlations from ALEPH. Results for the four-jet rate can be found in Table 6.. The correlations among the angular abservables are displayed in Fig. 6.2, where the correlations reach values higher than 5% for some bins. Then a χ 2 minimization is performed with respect to, x and y, using statistical errors only. The fit range is selected by requiring the total corrections for each observable
6.2 A Simultaneous Measurement of the Strong Coupling Constant and the Colour Factors 97 C det C det.2.5..5.95.9 5.2.4.6 Cosχ BZ.2.5..5.95.9 5.2.4.6 CosΘ NR C det C det.2 ALEPH Preliminary.5..5.95.9 5 - -.5.5 CosΦ KSW.2.5 qq _ full MC / HL..5.95.9 5 4p full MC / HL PYTHIA-4p - -.5.5 Cosα 34 Figure 6.: Detector corrections for the four-jet angular correlations. to be smaller than %. The results are seen in Table 6.7, and show goodagreement withboth QCD expectations and previous results [2] [55]. However, an important reduction of the statistical error is achieved. In the previous ALEPH analysis, the normalization of the four-jet angular correlations was also fitted, which prevented from achieving a better statistical precision. The fact that different observables were used should also be taken into account when comparing to the results in this thesis. The comparison with the OPAL results should be done after considering the smaller amount of MC used in that case for the hadronization and detector corrections, as well as the differences in the fit ranges for the angular correlations.
98 Measurements of the Strong Coupling Constant and the Colour Factors.2.5..5.95.9 5.2.4.6 C tot C tot Cosχ BZ.2.5..5.95.9 5.2.4.6 CosΘ NR C tot C tot.2 ALEPH Preliminary.5..5.95.9 5 - -.5.5.4.3.2..9 CosΦ KSW - -.5.5 Cosα 34 Figure 6.: Total corrections for the four-jet angular correlations. The dashed lines show the maximum allowed corrections used for the fit. The fitted distributions can be seen in Figs. 6.3 and 6.4. In the case of cos ff 34 a significant discrepancy in the central region of the distributions is observed. This disagreement was already seen in [55]. Its origin is not understood. For the four-jet rate, the disagreement between the fitted predictions at small values of y cut are again observed. 6.2.3 Systematic Studies Tables 6.8-6.22 show the systematic uncertainties that have been studied. A brief description of each of them can be found in the following paragraphs.
6.2 A Simultaneous Measurement of the Strong Coupling Constant and the Colour Factors 99 Center of bin N events (ALEPH data) N events (full MC) N events (at HL).25 582 ± 74:86 427 ± :45 79225 ± 277:4.75 5639 ± 73:78 492 ± :77 7947 ± 277:37.25 5554 ± 73:24 898 ± 2:64 8675 ± 279:48.75 5665 ± 73:95 77 ± 2:6 822 ± 282:4.225 5795 ± 74:76 8 ± 3:63 8447 ± 285:23.275 5872 ± 75:24 34 ± 4:63 856 ± 287:6.325 5938 ± 75:64 679 ± 6:2 88343 ± 292:.375 5926 ± 75:57 2235 ± 8:52 9475 ± 296:95.425 679 ± 76:5 228 ± 8:7 94282 ± 3:3.475 697 ± 77:2 2655 ± :29 97842 ± 36:7.525 667 ± 79:62 346 ± 3:38 35 ± 34:36.575 699 ± 8:8 3846 ± 5:5 825 ± 32:92.625 7297 ± 83:49 475 ± 8:54 3282 ± 328:98.675 7742 ± 85:87 567 ± 2:94 255 ± 34:3.725 822 ± 87:85 6768 ± 26:2 359 ± 35:86.775 8929 ± 9:87 8295 ± 3:4 4982 ± 366: 25 9862 ± 96:26 23 ± 38: 55978 ± 382:6 75 967 ± :4 2295 ± 43:83 756 ± 44:29.925 283 ± 8:72 26429 ± 55:85 22268 ± 43:45.975 2538 ± 45:8 523 ± 27:48 42553 ± 592:89 Table 6.3: Number of events per bin for the j cos χbzj distribution from the ALEPH data. The events at detector level (DL) from the full MC simulation and at hadron level (HL) are also given since these are the distributions used to calculate the detector corrections. Fit Range The sensitivity ofthe measurement to the fit range is checked by repeating the analysis with the requirement of a total correction per bin smaller than 2% (it was % in the standard analysis). The systematic variation due to this new fit range, see Table 6.8, is smaller than the statistical errors of the measurement showing that the range chosen for the nominal fit does not introduce any important biasinthemeasurement. Selection Criteria All cuts imposed in the selection of hadronic events have been moved in order to evaluate the effect on the measurement. The new values for the selection cuts on track parameters
Measurements of the Strong Coupling Constant and the Colour Factors Center of bin N events (ALEPH data) N events (full MC) N events (at HL) -.95 846 ± 4:8 23557 ± 47:84 2284 ± 44: -5 7938 ± 86:9 584 ± 22:67 289 ± 348:88 -.75 7527 ± 84:73 533 ± 2:77 296 ± 34:73 -.65 7539 ± 84:8 535 ± 2:4 98 ± 336:94 -.55 754 ± 84:8 59 ± 2:35 7738 ± 335:7 -.45 749 ± 84:54 5342 ± 2:92 62 ± 332:88 -.35 7768 ± 86: 4966 ± 9:5 3898 ± 329:83 -.25 7698 ± 85:64 478 ± 8:79 499 ± 326:5 -.5 7868 ± 86:53 549 ± 9:8 24 ± 324:57 -.5 85 ± 87:48 5335 ± 2:89 262 ± 327:28.5 722 ± 83:7 3799 ± 4:96 295 ± 3:37.5 6775 ± 8:58 268 ± :4 96427 ± 34:57.25 6537 ± 79:2 257 ± 9:67 9534 ± 32:92.35 62 ± 76:7 2638 ± :22 9527 ± 32:8.45 6325 ± 77:97 2646 ± :26 96998 ± 35:43.55 6286 ± 77:74 2832 ± :3 762 ± 3:6.65 672 ± 8:27 3662 ± 4:4 679 ± 38:82.75 734 ± 83:73 5448 ± 2:3 832 ± 335:85 5 8942 ± 9:93 862 ± 32:5 44837 ± 369:55.95 9439 ± 3:84 497 ± 89:73 3287 ± 529:43 Table 6.4: Number of events per bin for the cos Φ KSW distribution from the ALEPH data. The events at detector level (DL) from the full MC simulation and at hadron level (HL) are also given since these are the distributions used to calculate the detector corrections. are found by changing them until the number of selected events per unit luminosity isthe same in data and MC [54]. The analysis has been repeated by introducing the following changes (only one at a time): at least six measured space coordinates from the TPC; a polar angle at the origin in the range 2 ffi < < 6 ffi both for charged and neutral tracks; transverse momentum p t > :25 GeV/c; d =:867cm; z =6:64cm; at least 8 selected charged tracks; minimum charged energy 22 GeV; j cos Sph j < :85; and fraction of electromagnetic energy < 2%. The observed changes when modifying the selection cuts are in general small and in many cases even negligible. The largest are at the % level for x and 2% for y (always below %for ).
6.2 A Simultaneous Measurement of the Strong Coupling Constant and the Colour Factors Center of bin N events (ALEPH data) N events (full MC) N events (at HL).25 62 ± 77:23 843 ± 6:83 89238 ± 293:43.75 69 ± 77:8 936 ± 7:24 894 ± 293:2.25 63 ± 76:8 282 ± 8:29 8984 ± 294:38.75 6238 ± 77:45 267 ± 7:8 938 ± 296:7.225 642 ± 78:53 242 ± 9:27 9322 ± 299:5.275 649 ± 78:52 2944 ± :49 95685 ± 33:44.325 669 ± 79:63 3228 ± 2:66 99272 ± 38:85.375 6742 ± 8:39 365 ± 4:22 2727 ± 33:96.425 726 ± 8:99 487 ± 6:49 732 ± 32:33.475 749 ± 84: 474 ± 8:64 269 ± 328:6.525 7622 ± 85:24 5627 ± 2:98 865 ± 336:3.575 7929 ± 86:85 5862 ± 22:85 2642 ± 346:56.625 847 ± 89:29 6927 ± 26:69 33769 ± 355:97.675 885 ± 9:3 7958 ± 3:27 4353 ± 367:4.725 9349 ± 93:88 8976 ± 33:69 52675 ± 378:8.775 993 ± 96:57 24 ± 37:7 64 ± 388:47 25 258 ± 98:4 2967 ± 4:7 6872 ± 396:87 75 76 ± :25 2398 ± 4:4 7895 ± 399:23.925 99 ± :24 2229 ± 43:9 7728 ± 45:8.975 359 ± :34 27399 ± 58:43 2577 ± 443:77 Table 6.5: Number of events per bin for the j cos NRj distribution from the ALEPH data. The events at detector level (DL) from the full MC simulation and at hadron level (HL) are also given since these are the distributions used to calculate the detector corrections. The total experimental systematic uncertainty is the quadratic sum of all contributions in Table 6.9, where individual contributions are calculated in the Bayesian approach. It results in.2 for, aswell as.2 and. for the color factor ratios, x and y. Hadronization and Background Corrections The hadronization uncertainty is taken as the change in the fitted parameters when the corrections are calculated with HERWIG. The values can be found in Table 6.2 and show large systematic variations, up to 8.5% for y. The strong increase in the χ 2 when using
2 Measurements of the Strong Coupling Constant and the Colour Factors Center of bin N events (ALEPH data) N events (full MC) N events (at HL) -.95 9 ± 5:3 26388 ± 55:74 94722 ± 424: -5 876 ± :75 239 ± 46:57 76953 ± 45:73 -.75 8 ± 97:42 245 ± 4:43 6684 ± 393:97 -.65 946 ± 94:4 9943 ± 36:84 564 ± 382:68 -.55 969 ± 92:54 8834 ± 33:22 46238 ± 37:23 -.45 8778 ± 9:3 7892 ± 3:4 38956 ± 362:42 -.35 8663 ± 9:57 683 ± 26:35 32395 ± 354:24 -.25 8472 ± 89:62 6366 ± 24:68 27638 ± 348:6 -.5 8536 ± 89:94 645 ± 23:52 23478 ± 342:74 -.5 8553 ± 9:2 632 ± 23:47 2247 ± 34:32.5 8654 ± 9:52 5858 ± 22:83 25 ± 34:2.5 8573 ± 9:2 5873 ± 22:89 2253 ± 34:46.25 87 ± 9:75 6337 ± 24:58 2552 ± 345:39.35 876 ± 9:78 644 ± 24:82 26296 ± 346:42.45 86 ± 9:26 6535 ± 25:29 294 ± 35:.55 8376 ± 89:4 6359 ± 24:66 3377 ± 35:68.65 7634 ± 85:3 5844 ± 22:78 26767 ± 347:4.75 694 ± 77:9 335 ± 3:6 759 ± 325:46 5 296 ± 53:92 6727 ± 8:7 5689 ± 235:68.95 37 ± :7 358± 8:9 292 ± 54: Table 6.6: Number of events per bin for the cos ff34 distribution from the ALEPH data. The events at detector level (DL) from the full MC simulation and at hadron level (HL) are also given since these are the distributions used to calculate the detector corrections. (M Z ) x y χ 2 =N dof :255 ± :3 2:7 ± :6 :37± :2 76:8=8 ρ y ρ xy -.45 45 Table 6.7: Results for the combined fit using ALEPH data.
6.2 A Simultaneous Measurement of the Strong Coupling Constant and the Colour Factors 3 Correlation.6.4.2 8 7 6 Bin number 5 4 3 2 8 7 6 5 4 3 2 Bin number Figure 6.2: Bin-by-bin statistical correlations for the four-jet angular correlations. The numbers in the axis are the bin number: bins from to 2 correspond to j cos χbzj, binsfrom 2 to 4 to cos Φ KSW, bins from 4 to 6 to j cos NRj and bins from 6 to 8 to cos ff34. (M Z ) x y χ 2 =N dof tot.corr. < 2% :2565 ± :2 2:9 ± :56 :387 ± :9 89:=88 Range sys. = : x = :2 y = :2 ρ y ρ xy.. Table 6.8: Results when changing the fit range. The total corrections are chosen to be smaller than 2% instead of %.
4 Measurements of the Strong Coupling Constant and the Colour Factors N/ bin N 4 at Detector Level 3 2.5 2.5.2..9.2.4.6 Cosχ BZ Data/Fitted dist. N/ bin N 4 at Detector Level.4.2.6.4.2..9 - -.5.5 CosΦ KSW Data/Fitted dist. ALEPH Preliminary.5 Fit Range..2.3.4.5.6.7.9 Cosχ BZ.2 Fit Range - - -.6 -.4 -.2.2.4.6 CosΦ KSW N/ bin N 4 at Detector Level 2.8.6.4.2.6.4.2.2..9.2.4.6 CosΘ NR Data/Fitted dist. Fit Range N/ bin N 4 at Detector Level.7.6.5.4.3.2..2..9 - -.5.5 Cosα 34 Data/Fitted dist. Fit Range..2.3.4.5.6.7.9 CosΘ NR - - -.6 -.4 -.2.2.4.6 Cosα 34 Figure 6.3: Comparison of ALEPH data and fit results for the angular correlations in four-jet events. The curves are obtained at detector level. Full dots correspond to ALEPH data. The solid lines show the fitted distributions while dashed lines correspond to their statistical uncertainty. The ratio of data with respect to fitted distributions is shown in the small inserts. the corrections from HERWIG for the angular observables is an indication for unresolved issues in this new four-parton option. Detector Corrections An estimation of the systematic uncertainty due to the detector corrections has been obtained by repeating the analysis using charged tracks only. This change in the measurement procedure leads to systematic deviations on the parameters of about % for and x, andabout3%for y.
6.2 A Simultaneous Measurement of the Strong Coupling Constant and the Colour Factors 5 R 4 at Detector Level.35.3.25.2.5 Data/Fitted dist.2..9 ALEPH preliminary -7-6 -5-4 -ln(y cut )..5 Fit Range -7-6.5-6 -5.5-5 -4.5 -ln(y cut ) Figure 6.4: Comparison of ALEPH data and fit results for the four-jet rate. The curves are obtained at detector level. As in the previous figure, full dots correspond to ALEPH data and the solid line to the fitted distribution. The dashed lines are also plotted, however they are indistinguishable from the solid line for most of the y cut range. Again, the ratio of data with respect to fitted distributions is shown in the small insert. Another estimate is obtained by means of the four-parton full MC simulation described in Section 5.3.2. As explained, this new simulation is only used to correct cos ff 34. It results in deviations similar to using charged tracks only. The final uncertainty due to detector corrections, quoted in Table 6.2, is calculated by taking into account the two sources described above. Theoretical Predictions The lack of knowledge of higher order perturbative QCD is estimated by varying the renormalization scale in the theoretical predictions. The scale is varied from x μ =.5
6 Measurements of the Strong Coupling Constant and the Colour Factors (M Z ) x y χ 2 =N dof Sphericity cut :2559 ± :3 2:72 ± :64 :373 ± :2 7:9=8 TPC cut :254 ± :3 2:66 ± :62 :362 ± :2 76:3=8 N ch cut :2562 ± :3 2:5 ± :62 :363 ± :2 84:6=8 E ch cut :255 ± :3 2:79 ± :63 :373 ± :2 75:9=8 ch cut :2568 ± :3 2:67 ± :62 :376 ± :2 75:8=8 nt cut :2556 ± :3 2:55 ± :62 :366 ± :2 8:=8 Fraction of e.m. energy cut :2553 ± :3 2:63 ± :62 :366 ± :2 77:3=8 z cut :2553 ± :3 2:68 ± :62 :369 ± :2 78:=8 d cut :2553 ± :3 2:68 ± :62 :369 ± :2 76:9=8 p t cut :2549 ± :3 2:69 ± :62 :368 ± :2 79:=8 Experimental sys. = :2 x = :2 y = : ρ y ρ xy.766.532 Table 6.9: Systematic uncertainties due to the selection cuts used in the analysis. (M Z ) x y χ 2 =N dof HERWIG - all :2592 ± :33 2:27 ± :72 :428 ± :23 432:=8 HERWIG - angles, PYTHIA - R 4 :258 ± :32 2:225 ± :7 :37 ± :23 42:=8 PYTHIA - angles, HERWIG - R 4 :2639 ± :33 2:35 ± :64 :47 ± :2 79:=8 Background & Hadronization Sys. = :6 x = :2 y = :3 ρ y ρ xy.797.58 Table 6.2: Systematic uncertainties due to the background and hadronization corrections. to x μ =2, and the largest difference to the value found for x μ = is taken as systematic uncertainty. As the theoretical predictions for R 4 and for the angular correlations are known at different accuracy, the scale uncertainty is estimated by varying x μ separately for each of the two kind of observables. The resulting uncertainty is 4% for, 2% for x and 3% for y. It is the dominant one for the first and third parameters.
6.2 A Simultaneous Measurement of the Strong Coupling Constant and the Colour Factors 7 (M Z ) x y χ 2 =N dof Charged Only :2577 ± :3 2:43 ± :62 :359 ± :2 82:5=8 4-partons Full MC :2583 ± :3 2:89 ± :6 :346 ± :2 :=8 Detector sys. = : x = :2 y = : ρ y ρ xy -96.99 Table 6.2: Systematic uncertainties due to detector effects. The experimentally optimized scale method, which was used for the measurement of the strong coupling from the four-jet rate, is not used for the combined measurement for several reasons. First, the normalized angular correlations have been used due to the lack of the resummation of large logarithms. The normalized observables are expected to have a small scale dependence, as this is cancelled out in the normalization by the NLO four-jet cross section. Then, at least two different scales should be fitted, one for the four-jet rate and another for the angular observables. Second, since the colour factors and x μ are highly correlated, the addition of this new variable introduces instabilities in the fit. Finally, even if the predictions for all four-jet angular correlations are known at NLO, the optimized scale for each observable can take ondifferent values. An evaluation of mass effects, which are not included in the theoretical predictions, is attempted by using the FOURJPHACT MC program [58]. As the parameters for PYTHIA were optimized for massless partons, the hadronization and background corrections with massive partons for the angular observables are calculated as follows, C bck+had (i bin )= cos Xpart 4j (i bin ) cos X had py (i bin ) cos X part py (i bin ) cos X part py (i bin ) ; (6.3) where the index part-4j indicates the parton level coming from FOURJPHACT, and part-py (had-py) the parton (hadron) level from PYTHIA. The first ratio corrects for mass effects in the LO prediction, and the second ratio assumes that the showering and hadronization corrections do not depend strongly on the quark masses. It is found that mass effects might belarge, up to.7forx. The total theoretical uncertainty, see Table 6.22, is obtained adding quadratically the contribution of the two sources described above and results in the dominant systematic uncertainty for all parameters.
8 Measurements of the Strong Coupling Constant and the Colour Factors (M Z ) x y χ 2 =N dof x μ =.5 for the angles :2545 ± :32 2:93 ± :67 :377 ± :2 64:8=8 x μ =2. for the angles :2558 ± :3 2:48 ± :59 :36 ± :9 87:9=8 x μ =.5 for R 4 :2352 ± :3 2:265 ± :62 :266 ± :8 72:6=8 x μ =2. for R 4 :272 ± :3 2:96 ± :63 :439 ± :2 86:8=8 scale sys. =: x =:5 y =:5 quark masses :2629 ± :35 :982 ± :62 :37 ± :2 84:=8 mass sys. = :3 x = :7 y = :2 Theoretical sys. = : x = :9 y = :5 ρ y ρ xy.792 -.22 Table 6.22: Systematic uncertainties due to variations in the theoretical predictions. 6.2.4 Further Checks Hadronization and Background Corrections As a cross-check, the more extreme models presented in Sect. 5.3. were used to fit and the color factor ratios. The systematic changes in the fitted parameters, see Table 6.23, would be of about 2-3%, which is covered by the total uncertainty. Finally, also the standard PYTHIA simulation, q μq + PS + hadronization, has been used to correct the four-jet angular correlations (again Table 6.23). As expected, the χ 2 of the fit is much larger than the one of the standard fit showing that the PYTHIA simulation which uses four-parton matrix elements and a parton shower describes better the shape of the angular correlations. Two- and Three-Parton Backgrounds for the Angular Correlations The background and hadronization corrections used for the angular correlations are valid provided that the number of two- and three-parton events that are clustered into four jets after hadronization is negligible. This is verified by the following study. Using the PYTHIA ME option as described in Section 5.3., million events were generated. The hadronization parameters were the standard ones. The fraction of the
6.2 A Simultaneous Measurement of the Strong Coupling Constant and the Colour Factors 9 (M Z ) x y χ 2 =N dof ME- all :2787 ± :36 :967 ± :66 :343 ± :2 96:=8 ME- angles, PYTHIA - R 4 :2632 ± :34 :979 ± :65 :37 ± :2 98:3=8 PYTHIA - angles, ME- R 4 :272 ± :32 2:54 ± :63 :394 ± :2 76:9=8 PYTHIA,Q - R 4 :2749 ± :34 2: ± :63 :423 ± :2 9:=8 PYTHIA q μq - all :2639 ± :33 :92 ± :64 :293 ± :2 62:=8 Table 6.23: Check for the Hadronization and Background corrections. Deviations from the standard analysis are covered by the systematic uncertainties already described. number of four-jet events at HL which came from two- and three-parton events with respect to the total number offour-jet events was found to be much smaller than %, and only slightly affecting the shape of the angular correlations as observed in Fig. 6.5. In order to quantify how the two- and three-parton backgrounds could bias our measurement the four-jet angular correlations obtained at HL, from the PYTHIA ME simulation, were fitted. The hadronization corrections were calculated using the PYTHIA four-parton simulation. Then, from those distributions the background contributions were subtracted (dashed distributions in Fig. 6.5), and the resulting ones are fitted again. The difference between the fitted parameters can be taken as an estimate of the two- and three-parton background uncertainty. As seen in the Table 6.24, the systematic uncertainties for the different parameters are much smaller than the ones considered in the previous section. Therefore, no further two- and three-parton background treatment was implemented in the analysis. (M Z ) x y χ 2 =N dof Full HL distrib. :258 ± :27 2:22 ± :48 :38 ± :6 5./8 Without 2- & 3-parton bckg :2582 ± :28 2:8 ± :48 :38 ± :6.2/8 Table 6.24: Check for two- and three-parton background effects. Sensitivity Checks The sensitivity of the analysis to each of the observables is studied. The fit results when taking out one observable at a time can be seen in Table 6.25. As expected, is mainly fixed by the R 4 distribution, and the color factor ratios by the angular correlations. The
Measurements of the Strong Coupling Constant and the Colour Factors (4PL-4HL)/4HL (2or3PL - 4HL)/4HL..8.6.4 Distorsion of the true distribution.2.998.996.994.992.99..2.3.4.5.6.7.9 Cosχ BZ.2.8.6.4.2..8 2 and 3 parton background.6.4.2..2.3.4.5.6.7.9 Cosχ BZ (4PL-4HL)/4HL (2or3PL - 4HL)/4HL..8.6.4.2.998.996.994.992.99.2.8.6.4.2..8.6.4.2 Distorsion of the true distribution - - -.6 -.4 -.2.2.4.6 CosΦ KSW 2 and 3 parton background - - -.6 -.4 -.2.2.4.6 CosΦ KSW (4PL-4HL)/4HL (2or3PL - 4HL)/4HL..8.6.4 Distorsion of the true distribution.2.998.996.994.992.99..2.3.4.5.6.7.9 CosΘ NR.2.8.6.4.2..8 2 and 3 parton background.6.4.2..2.3.4.5.6.7.9 CosΘ NR (4PL-4HL)/4HL (2or3PL - 4HL)/4HL..8.6.4.2.998.996.994.992.99.2.8.6.4.2..8.6.4.2 Distorsion of the true distribution - - -.6 -.4 -.2.2.4.6 Cosα 34 2 and 3 parton background - - -.6 -.4 -.2.2.4.6 Cosα 34 Figure 6.5: Distortion in the four-jet angular correlations due to two- and three-parton backgrounds. sensitivity to the different angular correlations is quite similar. The fit with the angular correlations only resulted in a nonsense eta, for which the error could not be calculated. (M Z ) x y χ 2 =N dof No cos χ BZ :253 ± :3 2:2 ± :8 :38 ± :2 58:5=63 No cos Φ KSW :256 ± :3 2:2 ± :7 :35 ± :2 52:6=6 No cos NR :255 ± :3 2:9 ± :6 :38 ± :2 59:=6 No cos ff 34 :255 ± :3 2:6 ± :6 :37 ± :2 66:=6 No R 4 :9 ± 2:27 ± :9 :37 ± :5 44:2=74 Table 6.25: Results from the sensitivity check. The analysis is repeated taking out one of the observables at a time.
6.2 A Simultaneous Measurement of the Strong Coupling Constant and the Colour Factors Dependence on the y cut Acheck was performed in order to see if the present measurement depends on the chosen value of y cut. The analysis is repeated with the four-jet events calculated for y cut =., which represents a drop in the four-jet rate from 7.% for the standard.8 y cut value to 5.4%. The results are given in Table 6.26 and are in good agreement with the standard analysis. (M Z ) x y χ 2 =N dof y cut =. :259 ± :4 2:5 ± :8 :34 ± :2 99:5=8 Table 6.26: Results for the fit when y cut is fixed to. instead of.8 as used for the standard analysis. 6.2.5 Final Results Putting together all systematic uncertainties considered above, the final result of the combined measurement of and the colour factors ratios is: (M Z )=:255 ± :3(stat) ± :2(sys) x =2:7 ± :6(stat) ± :(sys) y =:37 ± :2(stat) ± :6(sys) (ρ y ) stat = :45 (ρ y ) sys =:769 (ρ xy ) stat =:845 (ρ xy ) sys = :28 which can also be expressed in terms of the strong coupling constant and the colour factors, ff s (M Z )=:9 ± :6(stat) ± :22(sys) C A =2:93 ± :4(stat) ± :5(sys) C F =:35 ± :7(stat) ± :22(sys) These results are in good agreement with previous measurements. The dominant source of systematic uncertainty isobtained from variations in the theoretical predictions, where both the scale and the quark mass effects result in large deviations from the standard measurement.
2 Measurements of the Strong Coupling Constant and the Colour Factors Figure 6.6 shows that the measurement of the colour factor ratios is in agreement with the expectations from QCD (x=2.25 and y=.375). The agreement with previous measurements by ALEPH [2] and lately by OPAL [55] is also observed. The total systematic errors are similar to the ones of previous analyses, but the statistical uncertainty has been strongly reduced because of the specific method adopted here. Finally, Fig.6.7 shows the fitted colour factor ratios for the systematic uncertainties considered in the analysis as well as for most of the further checks listed in Section 6.2.4. The compatibility of all variations with the standard fit is observed. 6.3 Massless Gluino Hypothesis Afinalstudy was carried out in order to test the hypothesis of the existence of a massless gluino. As in the measurement described above, DEBRECEN is used to obtain the NLO perturbative prediction. This MC program provides not only the B and C functions for pure QCD, but also for QCD+massless gluino. Only the four-jet angular correlations have been used, since there is no consistent prediction for R 4, for which gluino contributions are not available in the resummation terms. The simultaneous measurement of the strong coupling constant and the colour factors has been repeated using as perturbative predictions for the four-jet angular correlations the ones outlined in Eq. 3.22. Two cases have been considered. First, the B and C functions were taking into account only pure QCD configurations. Then the gluino contributions were also included in these functions, and the QCD beta function coefficients in Eq. 2.4 where changed to [2], fi = 7 3 x2 2 yn f + x 2 N g 2 fi = 3 x 4 3 3 where N g is the number ofgluinos, set to in this analysis. yn f + x N g 2 xyn f + x 2 N g 2 ; : (6.4) Hadronization and detector corrections were taken from the standard analysis under the assumption that they are not strongly dependent on the gluino contribution. At the moment of writing this work, there is no MC program which models the gluino contributions to hadronization. All the studies of systematic uncertainties described in
6.3 Massless Gluino Hypothesis 3 T R /C F ALEPH Preliminary This analysis, 68% CL contour.6.4.2 SU(4) ALEPH-997 OPAL-2.25.5.75 2 2.25 2.5 2.75 3 3.25 C A /C F Figure 6.6: 68% confidence level contour in the (x,y) plane, calculated from statistical plus systematic errors (shaded region). For comparison also the results from previous measurements are given, as well as predictions for simple Lie groups. Section 6.2.3 have been repeated. The results of the fit together with an estimate of the systematic uncertainties are
4 Measurements of the Strong Coupling Constant and the Colour Factors T R /C F.55.5 68% CL contour, syst. uncert. 95% CL contour,syst. uncert. 95% CL contour,total. uncert..45.4.35.3.25.2 x µ =2. R 4 HERWIG-all PYTHIA-Q R 4, HERWIG-R 4 PYTHIA-ME R 4, Range syst. θ ch cut, x µ =.5 angles θ nt, fem, z, d, p t, Sph, E ch cuts, HERWIG-angles Ch.Only, N ch, TPC cut, x µ =2 angles 4p Full PYTHIA-ME all, y cut Mass, PYTHIA-ME angles 2 and 3 bckg PYTHIA qq _ -all x µ =.5 R 4 Sources for Systematic Uncertainties Further Checks.9 2 2. 2.2 2.3 2.4 2.5 2.6 2.7 C A /C F Figure 6.7: 68% and 95% confidence level contours in the (x,y) plane when taking into account systematic uncertainties only. The 95% confidence level contour with the total uncertainty isshown for comparison. Results for the sources of systematic uncertainty are also plotted. x =2:27 ± :9(stat) ± :8(sys) y =:38 ± :5(stat) ± :7(sys) (ρ xy ) total = :45
6.4 Conclusions 5 for the pure QCD case, and x =2:26 ± :8(stat) ± :7(sys) y =:5 ± :6(stat) ± :6(sys) (ρ xy ) total = :87 for the QCD+gluino hypothesis. The reduction of the total systematic uncertainty for x with respect to the standard fit comes from the lack of one of the contributions to the theoretical uncertainty estimate, namely the scale uncertainty for the four-jet rate. Figure 6.8 shows that these results exclude the existence of a massless gluino at more than 95% confidence level, since the measured colour factor ratios do not agree with SU(3) anymore. It is worth noting that effects of a massive gluino have not been studied in this thesis, since no NLO predictions including mass terms are available. 6.4 Conclusions Results were presented for a measurement of the strong coupling constant from the four-jet rate as obtained by the Durham clustering algorithm (E-scheme). The result for ff s (M Z ) using 994-95 ALEPH data is: ff s (M Z )=:7 ± :(stat) ± :4(sys) ; with the renormalization scale x μ =. However, data shows a preference for x μ values smaller than, as can be seen when ff s and x μ are fitted simultaneously. In this case the results are ff s (M Z )=:75 ± :2(stat) ± :6(sys) x μ =:73 ± :5, where the error on x μ is statistical only. The preferred small x μ indicates that missing higher order corrections are still important. These results are in perfect agreement with previous measurements based on three-jet quantities [9][54].
6 Measurements of the Strong Coupling Constant and the Colour Factors ALEPH Preliminary T R /C F pure QCD, 68% CL contour pure QCD, 95% CL contour.6.4.2 -.2 QCD+gluino, 68% CL contour QCD+gluino, 95% CL contour.4.6.8 2 2.2 2.4 2.6 2.8 3 C A /C F Figure 6.8: 68% and 95% confidence level contours in the (x,y) plane for the QCD and QCD+gluino hypotheses, based on four-jet angular correlations. The uncertainties include statistical as well as systematic errors. Acombined measurement ofthe strong coupling constant and the colour factors from angular correlations in four-jet events and the four-jet rate has been presented. For the jet finding the Durham clustering algorithm (E-scheme) was used, with y cut =.8 for the four-jet events used to compute the angular correlations. The results are ff s (M Z )=:9 ± :6(stat) ± :22(sys) C A =2:93 ± :4(stat) ± :5(sys)
6.4 Conclusions 7 C F =:35 ± :7(stat) ± :22(sys) These results are in good agreement with previous measurements, with similar systematic uncertainties, but with an important reduction in the statistical error. Large discrepancies have been found when using either PYTHIA or HERWIG predictions for the calculations of hadronization corrections to the angular correlations. This indicates some problem either in the tuning or in the showering and hadronization implementation in these MC programs. The problems in the description of the shape for cos ff 34 are further hints together with the worsening of the shape of the j cos NR j shape when using a full simulation starting from four-parton massless MEs.