1: %
2: %
3: %
4: %
5: \section{Introduction}
6:
7: The W boson mass and width results presented in this chapter are obtained from
8: data recorded over a range of centre-of-mass energies,
9: $\roots=161-209$~\GeV, during the 1996-2000 operation of the LEP
10: collider. The results reported by the ALEPH, DELPHI and L3
11: collaborations include an analysis of the year 2000 data, and have an
12: integrated luminosity per experiment of about $700$~\ipb. The OPAL
13: collaboration has analysed the data up to and including 1999 and has
14: an integrated luminosity of approximately $450$~\ipb. The ALEPH result does not include an analysis of the small amount of data (about $10$~\ipb) collected in 1996 at a centre-of-mass energy of 172 GeV.
15:
16: The results on the W mass and width quoted below correspond to a
17: definition based on a Breit-Wigner denominator with an $s$-dependent width,
18: $|(s-\Mw^2) + i s \Gw/\Mw|$.
19:
20: \section{W Mass Measurements}
21:
22: Since 1996 the LEP $\epem$ collider has been operating above the
23: threshold for $\WW$ pair production. Initially, 10~\ipb\ of data
24: were recorded close to the $\WW$ pair production threshold. At this
25: energy the $\WW$ cross section is sensitive to the W boson mass, $\Mw$.
26: Table \ref{mw:tab:wmass_threshold} summarises the W mass results from the four
27: LEP collaborations based
28: on these data~\cite{common_bib:adloww161}.
29: \begin{table}[htbp]
30: \begin{center}
31: \begin{tabular}{|r|c|}\hline
32: \multicolumn{2}{|c|}{THRESHOLD ANALYSIS~\cite{common_bib:adloww161}} \\
33: Experiment & \Mw(threshold)/\GeVm \\ \hline
34: ALEPH & $80.14\pm0.35$ \\
35: DELPHI & $80.40\pm0.45$ \\
36: L3 & $80.80^{+0.48}_{-0.42}$ \\
37: OPAL & $80.40^{+0.46}_{-0.43}$ \\ \hline
38: \end{tabular}
39: \caption{W mass measurements from the $\WW$ threshold cross section
40: at $\roots=161$~\GeV. The errors
41: include statistical and systematic contributions.}
42: \label{mw:tab:wmass_threshold}
43: \end{center}
44: \end{table}
45:
46: Subsequently LEP has operated at energies significantly above the
47: $\WW$ threshold, where the $\epem\rightarrow\WW$ cross section has
48: little sensitivity to $\Mw$. For these higher energy data $\Mw$ is
49: measured through the direct reconstruction of the W boson's invariant
50: mass from the observed jets and leptons. Table
51: \ref{mw:tab:wmass_experiments} summarises the W mass results presented
52: individually by the four LEP experiments using the direct
53: reconstruction method. The combined values of $\Mw$ from each
54: collaboration take into account the correlated systematic
55: uncertainties between the decay channels and between the different
56: years of data taking. In addition to the combined numbers, each
57: experiment presents mass measurements from $\WWqqln$ and $\WWqqqq$
58: channels separately. The DELPHI and OPAL collaborations provide results from independent fits to the data in the $\qqln$ and $\qqqq$
59: decay channels separately and hence account for correlations between
60: years but do not need to include correlations between the two channels. The
61: $\qqln$ and $\qqqq$ results quoted by the ALEPH and L3 collaborations
62: are obtained from a simultaneous fit to all data which, in addition to
63: other correlations, takes into account the correlated systematic
64: uncertainties between the two channels. The L3 result is unchanged
65: when determined through separate fits. The
66: systematic uncertainties in the $\WWqqqq$ channel show
67: a large variation between experiments; this is caused by
68: differing estimates of the possible effects of Colour Reconnection
69: (CR) and Bose-Einstein Correlations (BEC), discussed below.
70: The systematic errors in the $\WWqqln$ channel are dominated by
71: uncertainties from hadronisation, with estimates ranging from
72: 15 to 30~$\MeVm$.
73:
74: The results presented in this note differ from those in the previous
75: combination \cite{bib-EWEP-02} due to revised measurements from the
76: ALEPH Collaboration~\cite{mw:bib:ALEPHRevised}; otherwise the results
77: are identical. The ALEPH measurements have been revised due to a
78: change in their event reconstruction algorithm. This change makes the
79: analysis less sensitive to detector simulation inaccuracies which were
80: not taken into account in their previous preliminary result.
81:
82: %
83: %
84: %
85: %
86: %
87: %
88: %
89: %
90: %
91: %
92: %
93: %
94: %
95: %
96: %
97: %
98: %
99: %
100: %
101: %
102: %
103: %
104: %
105: %
106: %
107: %
108: %
109: %
110: %
111:
112: \begin{table}[htbp]
113: \begin{center}
114: \begin{tabular}{|r|c|c||c|}\hline
115: \multicolumn{1}{|c|}{ } & \multicolumn{3}{c|}{DIRECT RECONSTRUCTION } \\
116: & \WWqqln & \WWqqqq & Combined \\
117: Experiment & \Mw/\GeVm & \Mw/\GeVm & \Mw/\GeVm \\ \hline
118: ALEPH \cite{mw:bib:ALEPHRevised}
119: & $80.375 \pm 0.062$ & $80.431 \pm 0.117$ & $80.385 \pm 0.058$ \\
120: DELPHI \cite{common_bib:delww172,mw:bib:D-mw183,mw:bib:D-mw189,mw:bib:D-mw20X}
121: & $80.414 \pm 0.089$ & $80.374 \pm 0.119$ & $80.402 \pm 0.075$ \\
122: L3 \cite{common_bib:ltrww172,mw:bib:L-mw183,mw:bib:L-mw189,mw:bib:L-mw19X,mw:bib:L-mw20X}
123: & $80.314 \pm 0.087$ & $80.485 \pm 0.127$ & $80.367 \pm 0.078$ \\
124: OPAL\cite{common_bib:opaww172,mw:bib:O-mw183,mw:bib:O-mw189,mw:bib:O-mw19X,mw:bib:O-mwlvlv}
125: & $80.516 \pm 0.073$ & $80.407 \pm 0.120$ & $80.495 \pm 0.067$ \\ \hline
126: \end{tabular}
127: \caption{Preliminary W mass measurements from direct reconstruction
128: ($\roots=172-209$~\GeV). Results are given for the
129: semi-leptonic, fully-hadronic channels and the combined value.
130: The $\WWqqln$ results from the OPAL collaboration include mass information from
131: the $\WWlnln$ channel. The results given here differ from those in the publications of the individual experiments as they have been recalculated imposing common FSI uncertainties.}
132: \label{mw:tab:wmass_experiments}
133: \end{center}
134: \end{table}
135:
136: \section{Combination Procedure}
137:
138: A combined LEP W mass measurement is obtained from the results
139: of the four experiments. In order to perform a reliable combination of
140: the measurements, a more detailed input than that given in
141: Table~\ref{mw:tab:wmass_experiments} is required. Each experiment
142: provided a W mass measurement for both the $\WWqqln$ and $\WWqqqq$
143: channels for each of the data taking years (1996-2000) that it had
144: analysed. In addition to the four threshold measurements a total of 36
145: direct reconstruction measurements are supplied: DELPHI
146: provided 10 measurements (1996-2000), L3 gave 8 measurements
147: (1996-2000) having already combined the 1996 and 1997 results, ALEPH provided 8 measurements (1997-2000) and OPAL also gave 8 measurements (1996-1999). The $\WWlnln$ channel is also
148: analysed by the OPAL(1997-1999)
149: collaboration; the lower precision results obtained from this channel
150: are combined with the $\WWqqln$ channel mass
151: determinations.
152:
153: Subdividing the results by data-taking years enables a proper
154: treatment of the correlated systematic uncertainty from the LEP beam
155: energy and other dependences on the centre-of-mass energy or
156: data-taking period. A detailed breakdown of the sources of systematic
157: uncertainty are provided for each result and the correlations
158: specified. The inter-year, inter-channel and inter-experiment
159: correlations are included in the combination. The main sources of
160: correlated systematic errors are: colour reconnection, Bose-Einstein
161: correlations, hadronisation, the LEP beam energy, and uncertainties
162: from initial and final state radiation. The full correlation matrix
163: for the LEP beam energy is employed\cite{mw:bib:energy}.
164: %
165: %
166: %
167: The combination is performed and the evaluation of the components of
168: the total error assessed using the Best Linear Unbiased Estimate
169: (BLUE) technique, see Reference~\citen{common_bib:BLUE}.
170:
171: \begin{figure}[tbp]
172: \begin{center}
173: \mbox{\epsfxsize=9.5cm\epsffile{sk1.eps}}
174: \caption{W mass bias obtained in the SK-I model of colour reconnection
175: relative to a simulation without colour reconnection as a function
176: of the fraction of events reconnected for the fully-hadronic decay channel
177: at a centre of mass energy of 189 \GeV . The
178: analyses of the four LEP experiments show similar sensitivity to
179: this effect. The points connected by the lines have correlated
180: uncertainties increasing to the right in the range indicated.}
181: \label{mw:fig:sk1}
182: \end{center}
183: \end{figure}
184:
185: A preliminary study of colour reconnection has been made by the LEP
186: experiments using the particle flow method \cite{mw:bib:CRcomb} on a
187: sample of fully-hadronic WW events, see chapter~\ref{sec-CR}. These
188: results are interpreted in terms of the reconnection parameter $k_i$
189: of the SK-I model \cite{mw:bib:ski} and yield a $68\%$ confidence
190: level range of:
191: \begin{eqnarray}
192: 0.39 < k_i < 2.13 \,.
193: \end{eqnarray}
194: The method was found to be insensitive to the \HERWIG\ and \ARIADNE-II
195: models of colour reconnection.
196:
197:
198: \begin{figure}[tbp]
199: \begin{center}
200: \mbox{\epsfxsize=9.5cm\epsffile{massnote_s02_dm_vs_energy.eps}}
201: \caption{The values used in the W Mass combination for the uncertainty due to colour reconnection are shown as a function of the centre of mass energy. These values were obtained from a linear fit to simulation results obtained with the SK1 model of colour reconnection at $k_i = 2.13$. }
202: \label{mw:fig:sk1cm}
203: \end{center}
204: \end{figure}
205:
206: Studies of simulation samples have demonstrated that the four
207: experiments are equally sensitive to colour reconnection effects, {\em
208: i.e.} when looking at the same CR model similar biases are seen by
209: all experiments. This is shown in Figure \ref{mw:fig:sk1} for the SKI
210: model as a function of the fraction of reconnected events. For this
211: reason a common value for all experiments of the CR systematic
212: uncertainty is used in the combination.
213:
214: For this combination, no offset has been applied to the central value
215: of \Mw\ due to colour reconnection effects and a symmetric systematic
216: error has been imposed. The $\Mw$ error is set from a linear
217: extrapolation of simulation results obtained at $k_i = 2.13$, the
218: values used in the combination ere: $74~\MeVm$ shift for the 1996
219: data at a centre-of-mass energy of $172~\GeV$, $84~\MeVm$ for 1997 at
220: $183~\GeV$, $90~\MeV$ for 1998 at $189~\GeV$, $95~\MeV$ for 1999 at
221: $195~\GeV$ and $105~\MeVm$ for 2000 at $207~\GeV$, they are shown in
222: Figure \ref{mw:fig:sk1cm}. Previous \Mw\ combinations have relied upon
223: theoretical expectations of colour reconnection effects, in which
224: there is considerable uncertainty. This new data driven approach
225: achieves a more robust uncertainty estimate at the expense of a
226: significantly increased colour reconnection uncertainty. The
227: \ARIADNE-II and \HERWIG\ models of colour reconnection have also been
228: studied and the W Mass shift was found to be lower than that from SK1
229: with $k_i = 2.13$ used for the combination.
230:
231: \label{mw:sec:cr}
232:
233: For Bose-Einstein correlations, a similar test has been made of the
234: respective experimental sensitivities with the LUBOEI
235: \cite{mw:bib:LUBOEI} model: the experiments observed compatible mass
236: shifts. A common value of the systematic uncertainty from BEC of
237: 35~\MeVm\ is assumed from studies of the LUBOEI model. This value may
238: be compared with recent direct measurements from LEP of this effect,
239: Chapter~\ref{sec-BE}, where the observed
240: Bose-Einstein effect was of smaller magnitude than in the LUBOEI
241: model, see chapter~\ref{sec-BE}. Hence, the currently assigned
242: 35~\MeVm\ uncertainty is considered a conservative estimate.
243:
244: \section{LEP Combined W Boson Mass }
245:
246: The combined W mass from direct reconstruction is
247: \begin{eqnarray}
248: \Mw(\mathrm{direct}) = 80.412\pm0.029(\mathrm{stat.})\pm0.031(\mathrm{syst.})~\GeVm,
249: \end{eqnarray}
250: with a $\chi^2$/d.o.f. of 28.2/33, corresponding to a $\chi^2$ probability
251: of 70\%.
252: The weight of the fully-hadronic channel in the combined fit is
253: 0.10. This reduced weight is a consequence of the relatively large
254: size of the current estimates of the systematic errors from
255: CR and BEC. Table \ref{mw:tab:errors} gives a breakdown of the contribution
256: to the total error of the various sources of systematic errors. The largest
257: contribution to the systematic error comes from hadronisation uncertainties,
258: which are conservatively treated as correlated between the two channels, between experiments and between years. In the absence of systematic effects the current LEP statistical precision on $\Mw$ would be $21$~$\MeVm$: the statistical error contribution in the LEP combination is larger than this (29~$\MeVm$) due to the significantly reduced weight of the fully-hadronic channel.
259: \begin{table}[tbp]
260: \begin{center}
261: \begin{tabular}{|l|r|r||r|}\hline
262: Source & \multicolumn{3}{|c|}{Systematic Error on \Mw\ ($\MeVm$)} \\
263: & \qqln & \qqqq & Combined \\ \hline
264: ISR/FSR & 8 & 8 & 8 \\
265: Hadronisation & 19 & 18 & 18 \\
266: Detector Systematics & 14 & 10 & 14 \\
267: LEP Beam Energy & 17 & 17 & 17 \\
268: Colour Reconnection & $-$& 90 & 9 \\
269: Bose-Einstein Correlations & $-$& 35 & 3 \\
270: Other & 4 & 5 & 4 \\ \hline
271: Total Systematic & 31 & 101 & 31 \\ \hline
272: Statistical & 32 & 35 & 29 \\ \hline\hline
273: Total & 44 & 107 & 43 \\ \hline
274: & & & \\
275: Statistical in absence of Systematics & 32 & 28 & 21 \\ \hline
276:
277: \end{tabular}
278: \caption{Error decomposition for the combined LEP W mass results.
279: Detector systematics include uncertainties
280: in the jet and lepton energy scales and resolution. The `Other'
281: category refers to errors, all of which are uncorrelated
282: between experiments, arising from: simulation statistics,
283: background estimation, four-fermion treatment, fitting method
284: and event selection. The error decomposition
285: in the $\qqln$ and $\qqqq$
286: channels refers to the independent fits to the results from
287: the two channels separately.}
288: \label{mw:tab:errors}
289: \end{center}
290: \end{table}
291:
292: In addition to the above results, the W boson mass is measured at
293: LEP from the 10~\ipb\ per experiment of
294: data recorded at threshold for W pair production:
295: \begin{eqnarray}
296: {\Mw(\mathrm{threshold}) =
297: 80.40\pm0.20(\mathrm{stat.})\pm
298: 0.07(\mathrm{syst.})\pm0.03(\mathrm{E_{beam}})~\GeVm}.
299: \end{eqnarray}
300: When the threshold measurements are combined with the much more precise results obtained from direct reconstruction one achieves a W mass measurement of
301: \begin{eqnarray}
302: \Mw = 80.412\pm0.029(\mathrm{stat.})\pm0.031(\mathrm{syst.}) \GeVm.
303: \end{eqnarray}
304: The LEP beam energy uncertainty is the only correlated systematic error source
305: between the threshold and direct reconstruction measurements.
306: The threshold measurements have a weight of only $0.03$ in the combined fit.
307: This LEP combined result is compared with the results (threshold and direct reconstruction combined) of the four LEP experiments in Figure \ref{mw:fig:mwgw}.
308:
309:
310: \section{Consistency Checks}
311:
312: The difference between the combined W boson mass measurements
313: obtained from the fully-hadronic and semi-leptonic channels,
314: $\Delta\Mw(\qqqq-\qqln)$, is determined:
315: \begin{eqnarray*}
316: \Delta\Mw(\qqqq-\qqln) = +22\pm43~\MeVm.
317: \end{eqnarray*}
318:
319: A significant non-zero value for $\Delta\Mw$ could indicate that CR and BEC
320: effects are biasing the value of \Mw\ determined from \WWqqqq\ events.
321: Since $\Delta\Mw$ is primarily of interest as a check of the possible
322: effects of final state interactions, the errors from CR and BEC are
323: set to zero in its determination. The result is obtained from a fit
324: where the imposed correlations are the same as those for the results
325: given in the previous sections. This result is almost unchanged if the
326: systematic part of the error on $\Mw$ from hadronisation effects is
327: considered as uncorrelated between channels, although the uncertainty
328: increases by 16\%: $\Delta\Mw=19\pm50~\MeVm$.
329: %
330: %
331: %
332: %
333: %
334: %
335: %
336: %
337:
338: The masses from the two channels obtained from this fit with the BEC and CR errors now included are:
339: \begin{eqnarray*}
340: \Mw(\WWqqln) = 80.411\pm0.032(\mathrm{stat.})\pm0.030(\mathrm{syst.})~\GeVm,\\
341: \Mw(\WWqqqq) = 80.420\pm0.035(\mathrm{stat.})\pm0.101(\mathrm{syst.})~\GeVm.
342: \end{eqnarray*}
343: These two results are correlated and have a correlation coefficient of
344: 0.18. The value of $\chi^2$/d.o.f is 28.2/32, corresponding to a
345: $\chi^2$ probability of 66$\%$.
346: These results and the correlation between them
347: can be used to combine the two measurements or to form the mass
348: difference. The LEP combined results from the two channels
349: are compared with those quoted by the individual experiments in
350: Figure \ref{mw:fig-qqlnqqqq}, where the common CR and BEC errors have been imposed.
351:
352: Experimentally, separate $\Mw$ measurements are obtained from the
353: $\WWqqln$ and $\WWqqqq$ channels for each of the years of data.
354: The combination using only the $\qqlv$ measurements yields:
355: \begin{eqnarray*}
356: \Mwindep(\WWqqln) = 80.413\pm0.032(\mathrm{stat.})\pm0.031(\mathrm{syst.})~\GeVm.
357: \end{eqnarray*}
358: The systematic error is dominated by
359: hadronisation uncertainties ($\pm19$~$\MeVm$) and the
360: uncertainty in the LEP beam energy ($\pm17$~$\MeVm$).
361: The combination using only the $\qqqq$ measurements gives:
362: \begin{eqnarray*}
363: \Mwindep(\WWqqqq) = 80.411\pm0.035(\mathrm{stat.})\pm0.107(\mathrm{syst.})~\GeVm.
364: \end{eqnarray*}
365: where the dominant contributions to the systematic error are from
366: CR ($\pm90$~$\MeVm$) and BEC ($\pm35$~$\MeVm$).
367:
368: \section{LEP Combined W Boson Width}
369:
370: The method of direct reconstruction is also well suited to the
371: direct measurement of the width of the W boson. The results of the four
372: LEP experiments are shown in Table \ref{mw:tab:wwidth_experiments}
373: and in Figure \ref{mw:fig:mwgw}.
374: \begin{table}[htbp]
375: \begin{center}
376: \begin{tabular}{|c|c|}\hline
377: Experiment & \Gw\ (\GeVm) \\ \hline
378: ALEPH & $2.13\pm0.11\pm0.09$ \\
379: DELPHI & $2.11\pm0.10\pm0.07$ \\
380: L3 & $2.24\pm0.11\pm0.15$ \\
381: OPAL & $2.04\pm0.16\pm0.09$ \\ \hline
382: \end{tabular}
383: \caption{Preliminary W width measurements ($\roots=172-209$~\GeV)
384: from the individual experiments. The first error is statistical
385: and the second systematic.}
386: \label{mw:tab:wwidth_experiments}
387: \end{center}
388: \end{table}
389:
390: Each experiment provided a W width measurement for both $\WWqqln$ and
391: $\WWqqqq$ channels for each of the data taking years (1996-2000) that
392: it has analysed. A total of 25 measurements are supplied: ALEPH
393: provided 3 $\WWqqqq$ results (1998-2000) and two $\WWqqln$ results
394: (1998-1999), DELPHI 8 measurements (1997-2000), L3 8 measurements
395: (1996-2000) having already combined the 1996 and 1997 results and OPAL
396: provided 4 measurements (1996-1998) where for the first two years the
397: $\WWqqln$ and $\WWqqqq$ results are already combined.
398:
399: A common colour reconnection error of 65 $\MeVm$ and a common
400: Bose-Einstein correlation error of 35 $\MeVm$ are used in the
401: combination. These common errors were determined such that the same
402: error was obtained on \Gw\ as when using the BEC/CR errors supplied by the experiments. The change in the value of the width is only 2 \MeVm. The BEC and CR values supplied by the experiments were based on studies of phenomenological models of these effects, the uncertainty has not yet been determined from the particle flow measurements of colour reconnection.
403:
404: A simultaneous fit to the results of the four LEP collaborations is
405: performed in the same way as for the $\Mw$ measurement. Correlated
406: systematic uncertainties are taken into account and the combination gives:
407: \begin{eqnarray}
408: \Gw = 2.150\pm0.068(\mathrm{stat.})\pm0.060
409: (\mathrm{syst.})~\GeVm,
410: \end{eqnarray}
411: with a $\chi^2$/d.o.f. of 19.7/24, corresponding to a
412: $\chi^2$ probability of 71$\%$.
413:
414:
415: %
416: %
417: %
418:
419: \section{Summary}
420:
421: The results of the four LEP experiments on the mass and width of the W
422: boson are combined taking into account correlated systematic
423: uncertainties, giving:
424: \begin{eqnarray*}
425: \Mw & = & 80.412\pm0.042~\GeVm, \\
426: \Gw & = & 2.150\pm0.091~\GeVm.
427: \end{eqnarray*}
428: The statistical correlation between mass and width is small and
429: neglected. Their correlation due to common systematic effects is
430: under study.
431:
432:
433:
434: \begin{figure}[hbt]
435: \begin{center}
436: \mbox{\epsfxsize=8.25cm\epsffile{mw_ffff.eps}
437: \epsfxsize=8.25cm\epsffile{gw_ffff.eps}}
438: \vskip-0.5cm
439: \caption{\label{mw:fig:mwgw}
440: The combined results for the measurements of the
441: W mass (left) and W width (right) compared to the results
442: obtained by the four LEP collaborations. The combined
443: values take into account correlations between experiments
444: and years and hence, in general, do not give the same central
445: value as a simple average. In the LEP combination of
446: the $\qqqq$ results common values (see text) for the CR and BEC
447: errors are used. The individual and combined $\Mw$ results
448: include the measurements from
449: the threshold cross section. The $\Mw$ values from the experiments
450: have been recalculated for this plot including the common LEP
451: CR and BEC errors.}
452: \end{center}
453: \end{figure}
454:
455: %
456:
457: \begin{figure}[hbt]
458: \begin{center}
459: \mbox{\epsfxsize=8.25cm\epsffile{mw_qqln.eps} \newline
460: \epsfxsize=8.25cm\epsffile{mw_qqqq.eps}}
461: \vspace*{-0.5cm}
462: \caption{\label{mw:fig-qqlnqqqq}
463: The W mass measurements
464: from the $\WWqqln$ (left) and $\WWqqqq$ (right) channels
465: obtained by the four LEP collaborations compared to the
466: combined value. The combined values take into account
467: correlations between experiments, years and the two channels.
468: In the LEP combination of
469: the $\qqqq$ results common values (see text) for the CR and BEC
470: errors are used.
471: The ALEPH and L3 $\qqln$ and $\qqqq$ results are
472: correlated since they are obtained from a fit to both channels
473: taking into account inter-channel correlations. The $\Mw$ values from the experiments have been recalculated
474: for this plot including the common LEP CR and BEC errors.}
475: \end{center}
476: \end{figure}
477:
478:
479: %
480: %
481: %
482:
483:
484:
485:
486:
487:
488:
489:
490:
491: