]>
Commit | Line | Data |
---|---|---|
9419eeef | 1 | <chapter name="PDF Selection"> |
2 | ||
3 | <h2>PDF Selection</h2> | |
4 | ||
5 | This page contains five subsections. The first deals with how to | |
6 | pick the parton distribution set for protons, including from LHAPDF, | |
7 | to be used for all proton and antiproton beams. The second is a special | |
8 | option that allows a separate PDF set to be used for the hard process | |
9 | only, while the first choice would still apply to everything else. | |
10 | The third and fourth give access to pion and Pomeron PDF's, respectively, | |
11 | the latter being used to describe diffractive systems. | |
12 | The fifth gives the possibility to switch off the lepton | |
13 | "parton density". | |
14 | ||
15 | <h3>Parton densities for protons</h3> | |
16 | ||
17 | The selection of parton densities is made once and then is propagated | |
18 | through the program. It is essential to make an informed choice, | |
19 | for several reasons <ref>Kas10</ref>: | |
20 | <note>Warning 1:</note> the choice of PDF set affects a number of | |
21 | properties of events. A change of PDF therefore requires a complete | |
22 | retuning e.g. of the multiple-interactions model for minimum-bias and | |
23 | underlying events. | |
24 | <note>Warning 2:</note> People often underestimate the differences | |
25 | between different sets on the market. The sets for the same order are | |
26 | constructed to behave more or less similarly at large <ei>x</ei> and | |
27 | <ei>Q^2</ei>, while the multiple interactions are dominated by the | |
28 | behaviour in the region of small <ei>x</ei> and <ei>Q^2</ei>. A good | |
29 | PDF parametrization ought to be sensible down to <ei>x = 10^-6</ei> | |
30 | (<ei>x = 10^-7</ei>) and <ei>Q^2 = 1</ei> GeV^2 for Tevatron (LHC) | |
31 | applications. Unfortunately there are distributions on the market that | |
32 | completely derail in that region. The <code>main41.cc</code> and | |
33 | <code>main42.cc</code> programs in the <code>examples</code> | |
34 | subdirectory provide some examples of absolutely minimal sanity checks | |
35 | before a new PDF set is put in production. | |
36 | <note>Warning 3:</note> NLO and LO sets tend to have quite different | |
37 | behaviours, e.g. NLO ones have less gluons at small x, which then is | |
38 | compensated by positive corrections in the NLO matrix elements. | |
39 | Therefore do not blindly assume that an NLO tune has to be better than | |
40 | an LO one when combined with the LO matrix elements in PYTHIA. There are | |
41 | explicit examples where such thinking can lead you down the wrong alley, | |
42 | especially if you study low-<ei>pT</ei> physics. In the list below you | |
43 | should therefore be extra cautious when using set 6 or set 9. | |
44 | ||
45 | <p/> | |
46 | The simplest option is to pick one | |
47 | of the distributions available internally: | |
48 | ||
49 | <modepick name="PDF:pSet" default="2" min="1" max="12"> | |
50 | Parton densities to be used for proton beams (and, by implication, | |
51 | antiproton ones): | |
52 | <option value="1">GRV 94L, LO <ei>alpha_s(M_Z) = 0.128</ei> | |
53 | (this set is out of date, but retained for historical comparisons).</option> | |
54 | <option value="2">CTEQ 5L, LO <ei>alpha_s(M_Z) = 0.127</ei> | |
55 | (this set is also out of date, but not badly so, and many tunes | |
56 | are based on it).</option> | |
57 | <option value="3">MRST LO* (2007), | |
58 | NLO <ei>alpha_s(M_Z) = 0.12032</ei>.</option> | |
59 | <option value="4">MRST LO** (2008), | |
60 | NLO <ei>alpha_s(M_Z) = 0.11517</ei>.</option> | |
61 | <option value="5">MSTW 2008 LO (central member), | |
62 | LO <ei>alpha_s(M_Z) = 0.13939</ei>.</option> | |
63 | <option value="6">MSTW 2008 NLO (central member), | |
64 | NLO <ei>alpha_s(M_Z) = 0.12018</ei> (NLO, see Warning 3 above).</option> | |
65 | <option value="7">CTEQ6L, NLO <ei>alpha_s(M_Z) = 0.1180</ei>.</option> | |
66 | <option value="8">CTEQ6L1, LO <ei>alpha_s(M_Z) = 0.1298</ei>.</option> | |
67 | <option value="9">CTEQ66.00 (NLO, central member), | |
68 | NLO <ei>alpha_s(M_Z) = 0.1180</ei> (NLO, see Warning 3 above).</option> | |
69 | <option value="10">CT09MC1, LO <ei>alpha_s(M_Z) = 0.1300</ei>.</option> | |
70 | <option value="11">CT09MC2, NLO <ei>alpha_s(M_Z) = 0.1180</ei>.</option> | |
71 | <option value="12">CT09MCS, NLO <ei>alpha_s(M_Z) = 0.1180</ei>.</option> | |
72 | </modepick> | |
73 | <note>Note:</note> the <ei>alpha_s(M_Z)</ei> values and the order of the | |
74 | running in the description above is purely informative, and does not | |
75 | affect any other parts of the program. Instead you have the freedom to | |
76 | set <ei>alpha_s(M_Z)</ei> value and running separately for | |
77 | <aloc href="CouplingsAndScales">hard processes</aloc> | |
78 | (including resonance decays), | |
79 | <aloc href="MultipleInteractions">multiple interactions</aloc>, | |
80 | <aloc href="SpacelikeShowers">initial-state radiation</aloc>, and | |
81 | <aloc href="TimelikeShowers">final-state radiation</aloc>. | |
82 | ||
83 | <p/> | |
84 | This is a reasonably complete list of recent LO fits, both | |
85 | ones within the normal LO context and ones with modifications for better | |
86 | matching to event generators. In addition two older sets are | |
87 | included for backwards reference (most studies to date are based on | |
88 | CTEQ 5L). If you link to the | |
89 | <a href="http://projects.hepforge.org/lhapdf/" target="page">LHAPDF | |
90 | library</a> <ref>Wha05</ref> you get access to a much wider selection. | |
91 | <note>Warning 1:</note> owing to previous problems with the behaviour | |
92 | of PDF's beyond the <ei>x</ei> and <ei>Q^2</ei> boundaries of a set, | |
93 | you should only use LHAPDF <b>version 5.3.0 or later</b>. | |
94 | <note>Warning 2:</note> the behaviour of the LHAPDF sets need not be | |
95 | identical with the implementation found in PYTHIA. Specifically we | |
96 | are aware of the following points that may influence a comparison. | |
97 | <br/>(a) CTEQ 5L in PYTHIA is the parametrization, in LHAPDF the grid | |
98 | interpolation. | |
99 | <br/>(b) MRST LO* and LO** in PYTHIA is based on an updated edition, | |
100 | where one makes use of the expanded MSTW grid format, while LHAPDF | |
101 | is based on the original smaller grid. | |
102 | <br/>(c) The CTEQ 6 and CT09MC sets in PYTHIA are frozen at the | |
103 | boundaries of the grid, by recommendation of the authors, while | |
104 | LHAPDF also offers an option with a smooth extrapolation outside | |
105 | the grid boundaries. | |
106 | ||
107 | <flag name="PDF:useLHAPDF" default="off"> | |
108 | If off then the choice of proton PDF is based on <code>PDF:pSet</code> | |
109 | above. If on then it is instead based on the choice of | |
110 | <code>PDF:LHAPDFset</code> and <code>PDF:LHAPDFmember</code> below. | |
111 | <note>Note:</note> in order for this option to work you must have | |
112 | compiled PYTHIA appropriately and have set the <code>LHAPATH</code> | |
113 | environment variable to provide the data-files directory of your local | |
114 | LHAPDF installation. See the README file in the <code>examples</code> | |
115 | directory for further instructions. | |
116 | </flag> | |
117 | ||
118 | <word name="PDF:LHAPDFset" default="MRST2004FF4lo.LHgrid"> | |
119 | Name of proton PDF set from LHAPDF to be used. You have to choose | |
120 | from the | |
121 | <a href="http://projects.hepforge.org/lhapdf/pdfsets" target="page"> | |
122 | list of available sets</a>. Examples of some fairly recent ones | |
123 | (but still less recent than found above) would be | |
124 | cteq61.LHpdf, cteq61.LHgrid, cteq6l.LHpdf, cteq6ll.LHpdf, | |
125 | MRST2004nlo.LHpdf, MRST2004nlo.LHgrid, MRST2004nnlo.LHgrid and | |
126 | MRST2004FF3lo.LHgrid. If you pick a LHpdf set it will require some | |
127 | calculation the first time it is called. | |
128 | <note>Technical note:</note> if you provide a name beginning with a | |
129 | slash (/) it is assumed you want to provide the full file path and then | |
130 | <code>initPDFsetM(name)</code> is called, else the correct path is assumed | |
131 | already set and <code>initPDFsetByNameM(name)</code> is called. | |
132 | </word> | |
133 | ||
134 | <modeopen name="PDF:LHAPDFmember" default="0" min="0"> | |
135 | Further choice of a specific member from the set picked above. Member 0 | |
136 | should normally correspond to the central value, with higher values | |
137 | corresponding to different error PDF's somewhat off in different | |
138 | directions. You have to check from set to set which options are open. | |
139 | <note>Note:</note> you can only use one member in a run, so if you | |
140 | want to sweep over many members you either have to do many separate | |
141 | runs or, as a simplification, save the | |
142 | <aloc href="EventInformation">pdf weights</aloc> at the hard scattering | |
143 | and do an offline reweighting of events. | |
144 | </modeopen> | |
145 | ||
146 | <flag name="PDF:extrapolateLHAPDF" default="off"> | |
147 | Parton densities have a guaranteed range of validity in <ei>x</ei> | |
148 | and <ei>Q^2</ei>, and what should be done beyond that range usually is | |
149 | not explained by the authors of PDF sets. Nevertheless these boundaries | |
150 | very often are exceeded, e.g. minimum-bias studies at LHC may sample | |
151 | <ei>x</ei> values down to <ei>10^-8</ei>, while many PDF sets stop | |
152 | already at <ei>10^-5</ei>. The default behaviour is then that the | |
153 | PDF's are frozen at the boundary, i.e. <ei>xf(x,Q^2)</ei> is fixed at | |
154 | its value at <ei>x_min</ei> for all values <ei>x < x_min</ei>, | |
155 | and so on. This is a conservative approach. Alternatively, if you | |
156 | switch on extrapolation, then parametrizations will be extended beyond | |
157 | the boundaries, by some prescription. In some cases this will provide a | |
158 | more realistic answer, in others complete rubbish. Another problem is | |
159 | that some of the PDF-set codes will write a warning message anytime the | |
160 | limits are exceeded, thus swamping your output file. Therefore you should | |
161 | study a set seriously before you run it with this switch on. | |
162 | </flag> | |
163 | ||
164 | <p/> | |
165 | If you want to use PDF's not found in LHAPDF, or you want to interface | |
166 | LHAPDF another way, you have full freedom to use the more generic | |
167 | <aloc href="PartonDistributions">interface options</aloc>. | |
168 | ||
169 | <h3>Parton densities for protons in the hard process</h3> | |
170 | ||
171 | The above options provides a PDF set that will be used everywhere: | |
172 | for the hard process, the parton showers and the multiple interactions | |
173 | alike. As already mentioned, therefore a change of PDF should be | |
174 | accompanied by a <b>complete</b> retuning of the whole MI framework, | |
175 | and maybe more. There are cases where one may want to explore | |
176 | different PDF options for the hard process, but would not want to touch | |
177 | the rest. If several different sets are to be compared, a simple | |
178 | reweighting based on the <aloc href="EventInformation">originally | |
179 | used</aloc> flavour, <ei>x</ei>, <ei>Q^2</ei> and PDF values may offer the | |
180 | best route. The options in this section allow a choice of the PDF set | |
181 | for the hard process alone, while the choice made in the previous section | |
182 | would still be used for everything else. The hardest interaction | |
183 | of the minimum-bias process is part of the multiple-interactions | |
184 | framework and so does not count as a hard process here. | |
185 | ||
186 | <p/> | |
187 | Of course it is inconsistent to use different PDF's in different parts | |
188 | of an event, but if the <ei>x</ei> and <ei>Q^2</ei> ranges mainly accessed | |
189 | by the components are rather different then the contradiction would not be | |
190 | too glaring. Furthermore, since standard PDF's are one-particle-inclusive | |
191 | we anyway have to 'invent' our own PDF modifications to handle configurations | |
192 | where more than one parton is kicked out of the proton <ref>Sjo04</ref>. | |
193 | ||
194 | <p/> | |
195 | The PDF choices that can be made are the same as above, so we do not | |
196 | repeat the detailed discussion. | |
197 | ||
198 | <flag name="PDF:useHard" default="off"> | |
199 | If on then select a separate PDF set for the hard process, using the | |
200 | variables below. If off then use the same PDF set for everything, | |
201 | as already chosen above. | |
202 | </flag> | |
203 | ||
204 | <modepick name="PDF:pHardSet" default="2" min="1" max="12"> | |
205 | Parton densities to be used for proton beams (and, by implication, | |
206 | antiproton ones): | |
207 | <option value="1">GRV 94L, LO <ei>alpha_s(M_Z) = 0.128</ei> | |
208 | (out of date).</option> | |
209 | <option value="2">CTEQ 5L, LO <ei>alpha_s(M_Z) = 0.127</ei> | |
210 | (slightly out of date; many tunes are based on it).</option> | |
211 | <option value="3">MRST LO* (2007), | |
212 | NLO <ei>alpha_s(M_Z) = 0.12032</ei>.</option> | |
213 | <option value="4">MRST LO** (2008), | |
214 | NLO <ei>alpha_s(M_Z) = 0.11517</ei>.</option> | |
215 | <option value="5">MSTW 2008 LO (central member), | |
216 | LO <ei>alpha_s(M_Z) = 0.13939</ei>.</option> | |
217 | <option value="6">MSTW 2008 NLO (central member), | |
218 | LO <ei>alpha_s(M_Z) = 0.12018</ei>.</option> | |
219 | <option value="7">CTEQ6L, NLO <ei>alpha_s(M_Z) = 0.1180</ei>.</option> | |
220 | <option value="8">CTEQ6L1, LO <ei>alpha_s(M_Z) = 0.1298</ei>.</option> | |
221 | <option value="9">CTEQ66.00 (NLO, central member), | |
222 | NLO <ei>alpha_s(M_Z) = 0.1180</ei>.</option> | |
223 | <option value="10">CT09MC1, LO <ei>alpha_s(M_Z) = 0.1300</ei>.</option> | |
224 | <option value="11">CT09MC2, NLO <ei>alpha_s(M_Z) = 0.1180</ei>.</option> | |
225 | <option value="12">CT09MCS, NLO <ei>alpha_s(M_Z) = 0.1180</ei>.</option> | |
226 | </modepick> | |
227 | ||
228 | <flag name="PDF:useHardLHAPDF" default="off"> | |
229 | If off then the choice of proton PDF is based on <code>hardpPDFset</code> | |
230 | above. If on then it is instead based on the choice of | |
231 | <code>hardLHAPDFset</code> and <code>hardLHAPDFmember</code> below. | |
232 | </flag> | |
233 | ||
234 | <word name="PDF:hardLHAPDFset" default="MRST2004FF4lo.LHgrid"> | |
235 | Name of proton PDF set from LHAPDF to be used. | |
236 | </word> | |
237 | ||
238 | <modeopen name="PDF:hardLHAPDFmember" default="0" min="0"> | |
239 | Further choice of a specific member from the set picked above. | |
240 | </modeopen> | |
241 | ||
242 | <p/> | |
243 | Note that there is no separate equivalent of the | |
244 | <code>PDF:extrapolateLHAPDF</code> flag specifically for the hard | |
245 | PDF. Since LHAPDF only has one global flag for extrapolation or not, | |
246 | the choice for the normal PDF's also applies to the hard ones. | |
247 | ||
248 | <h3>Parton densities for pions</h3> | |
249 | ||
250 | The parton densities of the pion are considerably less well known than | |
251 | those of the proton. There are only rather few sets on the market, | |
252 | and none particularly recent. Only one comes built-in, but others can | |
253 | be accessed from LHAPDF. Input parametrizations are for the <ei>pi+</ei>. | |
254 | >From this the <ei>pi-</ei> is obtained by charge conjugation and the | |
255 | <ei>pi0</ei> from averaging (half the pions have <ei>d dbar</ei> | |
256 | valence quark content, half <ei>u ubar</ei>. | |
257 | ||
258 | <p/> | |
259 | Much of the switches are taken over from the proton case, with obvious | |
260 | modifications; therefore the description is briefer. Currently we have | |
261 | not seen the need to allow separate parton densities for hard processes. | |
262 | When using LHAPDF the <code>PDF:extrapolateLHAPDF</code> switch of the | |
263 | proton also applies to pions. | |
264 | ||
265 | <modepick name="PDF:piSet" default="1" min="1" max="1"> | |
266 | Internal parton densities that can be used for pion beams, currently with | |
267 | only one choice. | |
268 | <option value="1">GRV 92 L.</option> | |
269 | </modepick> | |
270 | ||
271 | <flag name="PDF:piUseLHAPDF" default="off"> | |
272 | If off then the choice of proton PDF is based on <code>PDF:piSet</code> | |
273 | above. If on then it is instead based on the choice of | |
274 | <code>PDF:piLHAPDFset</code> and <code>PDF:piLHAPDFmember</code> below. | |
275 | </flag> | |
276 | ||
277 | <word name="PDF:piLHAPDFset" default="OWPI.LHgrid"> | |
278 | Name of pion PDF set from LHAPDF to be used. You have to choose from the | |
279 | <a href="http://projects.hepforge.org/lhapdf/pdfsets" target="page"> | |
280 | list of available sets</a>. | |
281 | </word> | |
282 | ||
283 | <modeopen name="PDF:piLHAPDFmember" default="0" min="0"> | |
284 | Further choice of a specific member from the set picked above. | |
285 | </modeopen> | |
286 | ||
287 | <h3>Parton densities for Pomerons</h3> | |
288 | ||
289 | The Pomeron is introduced in the description of diffractive events, | |
290 | i.e. a diffractive system is viewed as a Pomeron-proton collision at a | |
291 | reduced CM energy. Here the PDF's are even less well known. | |
292 | Most experimental parametrizations are NLO, which makes them less | |
293 | well suited for Monte Carlo applications. Furthemore note that | |
294 | the momentum sum is arbitrarily normalized to a non-unity value. | |
295 | ||
296 | <modepick name="PDF:PomSet" default="6" min="1" max="6"> | |
297 | Parton densities that can be used for Pomeron beams. | |
298 | <option value="1"><ei>Q^2</ei>-independent parametrizations | |
299 | <ei>xf(x) = N_ab x^a (1 - x)^b</ei>, where <ei>N_ab</ei> ensures | |
300 | unit momentum sum. The <ei>a</ei> and <ei>b</ei> parameters can be | |
301 | set separately for the gluon and the quark distributions. The | |
302 | momentum fraction of gluons and quarks can be freely mixed, and | |
303 | production of <ei>s</ei> quarks can be suppressed relative to | |
304 | that of <ei>d</ei> and <ei>u</ei> ones, with antiquarks as likely | |
305 | as quarks. See further below how to set the six parameters of this | |
306 | approach. | |
307 | </option> | |
308 | <option value="2"><ei>pi0</ei> distributions, as specified in the | |
309 | section above. | |
310 | </option> | |
311 | <option value="3">the H1 2006 Fit A NLO <ei>Q^2</ei>-dependent | |
312 | parametrization, based on a tune to their data <ref>H1P06</ref>, | |
313 | rescaled by the factor <code>PomRescale</code> below. | |
314 | </option> | |
315 | <option value="4">the H1 2006 Fit B NLO <ei>Q^2</ei>-dependent | |
316 | parametrization, based on a tune to their data <ref>H1P06</ref>, | |
317 | rescaled by the factor <code>PomRescale</code> below. | |
318 | </option> | |
319 | <option value="5">the H1 2007 Jets NLO <ei>Q^2</ei>-dependent | |
320 | parametrization, based on a tune to their data <ref>H1P07</ref>, | |
321 | rescaled by the factor <code>PomRescale</code> below. | |
322 | </option> | |
323 | <option value="6">the H1 2006 Fit B LO <ei>Q^2</ei>-dependent | |
324 | parametrization, based on a tune to their data <ref>H1P06</ref>, | |
325 | rescaled by the factor <code>PomRescale</code> below. | |
326 | </option> | |
327 | </modepick> | |
328 | ||
329 | <parm name="PDF:PomGluonA" default="0." min="-0.5" max="2."> | |
330 | the parameter <ei>a</ei> in the ansatz <ei>xg(x) = N_ab x^a (1 - x)^b</ei> | |
331 | for option 1 above. | |
332 | </parm> | |
333 | ||
334 | <parm name="PDF:PomGluonB" default="3." min="0." max="10."> | |
335 | the parameter <ei>b</ei> in the ansatz <ei>xg(x) = N_ab x^a (1 - x)^b</ei> | |
336 | for option 1 above. | |
337 | </parm> | |
338 | ||
339 | <parm name="PDF:PomQuarkA" default="0." min="-0.5" max="2."> | |
340 | the parameter <ei>a</ei> in the ansatz <ei>xq(x) = N_ab x^a (1 - x)^b</ei> | |
341 | for option 1 above. | |
342 | </parm> | |
343 | ||
344 | <parm name="PDF:PomQuarkB" default="3." min="0." max="10."> | |
345 | the parameter <ei>b</ei> in the ansatz <ei>xq(x) = N_ab x^a (1 - x)^b</ei> | |
346 | for option 1 above. | |
347 | </parm> | |
348 | ||
349 | <parm name="PDF:PomQuarkFrac" default="0.2" min="0." max="1."> | |
350 | the fraction of the Pomeron momentum carried by quarks | |
351 | for option 1 above, with the rest carried by gluons. | |
352 | </parm> | |
353 | ||
354 | <parm name="PDF:PomStrangeSupp" default="0.5" min="0." max="1."> | |
355 | the suppression of the <ei>s</ei> quark density relative to that of the | |
356 | <ei>d</ei> and <ei>u</ei> ones for option 1 above. | |
357 | </parm> | |
358 | ||
359 | <parm name="PDF:PomRescale" default="1.0" min="0.5" max="5.0"> | |
360 | Rescale the four H1 fits above by this uniform factor, e.g. to bring | |
361 | up their momentum sum to around unity. By default all three have | |
362 | a momentum sum of order 0.5, suggesting that a factor around 2.0 | |
363 | should be used. You can use <code>examples/main41.cc</code> to get | |
364 | a more precise value. Note that also other parameters in the | |
365 | <aloc href="Diffraction">diffraction</aloc> framework may need to | |
366 | be retuned when this parameter is changed. | |
367 | </parm> | |
368 | ||
369 | <h3>Parton densities for leptons</h3> | |
370 | ||
371 | For electrons/leptons there is no need to choose between different | |
372 | parametrizations, since only one implementation is available, and | |
373 | should be rather uncontroversial (apart from some technical details). | |
374 | However, insofar as e.g. <ei>e^+ e^-</ei> data often are corrected | |
375 | back to a world without any initial-state photon radiation, it is | |
376 | useful to have a corresponding option available here. | |
377 | ||
378 | <flag name="PDF:lepton" default="on"> | |
379 | Use parton densities for lepton beams or not. If off the colliding | |
380 | leptons carry the full beam energy, if on part of the energy is | |
381 | radiated away by initial-state photons. In the latter case the | |
382 | initial-state showers will generate the angles and energies of the | |
383 | set of photons that go with the collision. In addition one collinear | |
384 | photon per beam carries any leftover amount of energy not described | |
385 | by shower emissions. If the initial-state showers are switched off | |
386 | these collinear photons will carry the full radiated energy. | |
387 | </flag> | |
388 | ||
389 | <h3>Incoming parton selection</h3> | |
390 | ||
391 | There is one useful degree of freedom to restrict the set of incoming | |
392 | quark flavours for hard processes. It does not change the PDF's as such, | |
393 | only which quarks are allowed to contribute to the hard-process cross | |
394 | sections. Note that separate but similarly named modes are available | |
395 | for multiple interactions and spacelike showers. | |
396 | ||
397 | <modeopen name="PDFinProcess:nQuarkIn" default="5" min="0" max="5"> | |
398 | Number of allowed incoming quark flavours in the beams; a change | |
399 | to 4 would thus exclude <ei>b</ei> and <ei>bbar</ei> as incoming | |
400 | partons, etc. | |
401 | </modeopen> | |
402 | ||
403 | </chapter> | |
404 | ||
405 | <!-- Copyright (C) 2010 Torbjorn Sjostrand --> |