3 <title>PDF Selection</title>
4 <link rel="stylesheet" type="text/css" href="pythia.css"/>
5 <link rel="shortcut icon" href="pythia32.gif"/>
11 This page contains five subsections. The first deals with how to
12 pick the parton distribution set for protons, including from LHAPDF,
13 to be used for all proton and antiproton beams. The second is a special
14 option that allows a separate PDF set to be used for the hard process
15 only, while the first choice would still apply to everything else.
16 The third and fourth give access to pion and Pomeron PDF's, respectively,
17 the latter being used to describe diffractive systems.
18 The fifth gives the possibility to switch off the lepton
21 <h3>Parton densities for protons</h3>
23 The selection of parton densities is made once and then is propagated
24 through the program. It is essential to make an informed choice,
25 for several reasons [<a href="Bibliography.html" target="page">Kas10</a>]:
26 <br/><b>Warning 1:</b> the choice of PDF set affects a number of
27 properties of events. A change of PDF therefore requires a complete
28 retuning e.g. of the multiple-interactions model for minimum-bias and
30 <br/><b>Warning 2:</b> People often underestimate the differences
31 between different sets on the market. The sets for the same order are
32 constructed to behave more or less similarly at large <i>x</i> and
33 <i>Q^2</i>, while the multiple interactions are dominated by the
34 behaviour in the region of small <i>x</i> and <i>Q^2</i>. A good
35 PDF parametrization ought to be sensible down to <i>x = 10^-6</i>
36 (<i>x = 10^-7</i>) and <i>Q^2 = 1</i> GeV^2 for Tevatron (LHC)
37 applications. Unfortunately there are distributions on the market that
38 completely derail in that region. The <code>main41.cc</code> and
39 <code>main42.cc</code> programs in the <code>examples</code>
40 subdirectory provide some examples of absolutely minimal sanity checks
41 before a new PDF set is put in production.
42 <br/><b>Warning 3:</b> NLO and LO sets tend to have quite different
43 behaviours, e.g. NLO ones have less gluons at small x, which then is
44 compensated by positive corrections in the NLO matrix elements.
45 Therefore do not blindly assume that an NLO tune has to be better than
46 an LO one when combined with the LO matrix elements in PYTHIA. There are
47 explicit examples where such thinking can lead you down the wrong alley,
48 especially if you study low-<i>pT</i> physics. In the list below you
49 should therefore be extra cautious when using set 6 or set 9.
52 The simplest option is to pick one
53 of the distributions available internally:
55 <p/><code>mode </code><strong> PDF:pSet </strong>
56 (<code>default = <strong>2</strong></code>; <code>minimum = 1</code>; <code>maximum = 12</code>)<br/>
57 Parton densities to be used for proton beams (and, by implication,
59 <br/><code>option </code><strong> 1</strong> : GRV 94L, LO <i>alpha_s(M_Z) = 0.128</i>
60 (this set is out of date, but retained for historical comparisons).
61 <br/><code>option </code><strong> 2</strong> : CTEQ 5L, LO <i>alpha_s(M_Z) = 0.127</i>
62 (this set is also out of date, but not badly so, and many tunes
64 <br/><code>option </code><strong> 3</strong> : MRST LO* (2007),
65 NLO <i>alpha_s(M_Z) = 0.12032</i>.
66 <br/><code>option </code><strong> 4</strong> : MRST LO** (2008),
67 NLO <i>alpha_s(M_Z) = 0.11517</i>.
68 <br/><code>option </code><strong> 5</strong> : MSTW 2008 LO (central member),
69 LO <i>alpha_s(M_Z) = 0.13939</i>.
70 <br/><code>option </code><strong> 6</strong> : MSTW 2008 NLO (central member),
71 NLO <i>alpha_s(M_Z) = 0.12018</i> (NLO, see Warning 3 above).
72 <br/><code>option </code><strong> 7</strong> : CTEQ6L, NLO <i>alpha_s(M_Z) = 0.1180</i>.
73 <br/><code>option </code><strong> 8</strong> : CTEQ6L1, LO <i>alpha_s(M_Z) = 0.1298</i>.
74 <br/><code>option </code><strong> 9</strong> : CTEQ66.00 (NLO, central member),
75 NLO <i>alpha_s(M_Z) = 0.1180</i> (NLO, see Warning 3 above).
76 <br/><code>option </code><strong> 10</strong> : CT09MC1, LO <i>alpha_s(M_Z) = 0.1300</i>.
77 <br/><code>option </code><strong> 11</strong> : CT09MC2, NLO <i>alpha_s(M_Z) = 0.1180</i>.
78 <br/><code>option </code><strong> 12</strong> : CT09MCS, NLO <i>alpha_s(M_Z) = 0.1180</i>.
80 <br/><b>Note:</b> the <i>alpha_s(M_Z)</i> values and the order of the
81 running in the description above is purely informative, and does not
82 affect any other parts of the program. Instead you have the freedom to
83 set <i>alpha_s(M_Z)</i> value and running separately for
84 <a href="CouplingsAndScales.html" target="page">hard processes</a>
85 (including resonance decays),
86 <a href="MultipleInteractions.html" target="page">multiple interactions</a>,
87 <a href="SpacelikeShowers.html" target="page">initial-state radiation</a>, and
88 <a href="TimelikeShowers.html" target="page">final-state radiation</a>.
91 This is a reasonably complete list of recent LO fits, both
92 ones within the normal LO context and ones with modifications for better
93 matching to event generators. In addition two older sets are
94 included for backwards reference (most studies to date are based on
95 CTEQ 5L). If you link to the
96 <a href="http://projects.hepforge.org/lhapdf/" target="page">LHAPDF
97 library</a> [<a href="Bibliography.html" target="page">Wha05</a>] you get access to a much wider selection.
98 <br/><b>Warning 1:</b> owing to previous problems with the behaviour
99 of PDF's beyond the <i>x</i> and <i>Q^2</i> boundaries of a set,
100 you should only use LHAPDF <b>version 5.3.0 or later</b>.
101 <br/><b>Warning 2:</b> the behaviour of the LHAPDF sets need not be
102 identical with the implementation found in PYTHIA. Specifically we
103 are aware of the following points that may influence a comparison.
104 <br/>(a) CTEQ 5L in PYTHIA is the parametrization, in LHAPDF the grid
106 <br/>(b) MRST LO* and LO** in PYTHIA is based on an updated edition,
107 where one makes use of the expanded MSTW grid format, while LHAPDF
108 is based on the original smaller grid.
109 <br/>(c) The CTEQ 6 and CT09MC sets in PYTHIA are frozen at the
110 boundaries of the grid, by recommendation of the authors, while
111 LHAPDF also offers an option with a smooth extrapolation outside
114 <p/><code>flag </code><strong> PDF:useLHAPDF </strong>
115 (<code>default = <strong>off</strong></code>)<br/>
116 If off then the choice of proton PDF is based on <code>PDF:pSet</code>
117 above. If on then it is instead based on the choice of
118 <code>PDF:LHAPDFset</code> and <code>PDF:LHAPDFmember</code> below.
119 <br/><b>Note:</b> in order for this option to work you must have
120 compiled PYTHIA appropriately and have set the <code>LHAPATH</code>
121 environment variable to provide the data-files directory of your local
122 LHAPDF installation. See the README file in the <code>examples</code>
123 directory for further instructions.
126 <p/><code>word </code><strong> PDF:LHAPDFset </strong>
127 (<code>default = <strong>MRST2004FF4lo.LHgrid</strong></code>)<br/>
128 Name of proton PDF set from LHAPDF to be used. You have to choose
130 <a href="http://projects.hepforge.org/lhapdf/pdfsets" target="page">
131 list of available sets</a>. Examples of some fairly recent ones
132 (but still less recent than found above) would be
133 cteq61.LHpdf, cteq61.LHgrid, cteq6l.LHpdf, cteq6ll.LHpdf,
134 MRST2004nlo.LHpdf, MRST2004nlo.LHgrid, MRST2004nnlo.LHgrid and
135 MRST2004FF3lo.LHgrid. If you pick a LHpdf set it will require some
136 calculation the first time it is called.
137 <br/><b>Technical note:</b> if you provide a name beginning with a
138 slash (/) it is assumed you want to provide the full file path and then
139 <code>initPDFsetM(name)</code> is called, else the correct path is assumed
140 already set and <code>initPDFsetByNameM(name)</code> is called.
143 <p/><code>mode </code><strong> PDF:LHAPDFmember </strong>
144 (<code>default = <strong>0</strong></code>; <code>minimum = 0</code>)<br/>
145 Further choice of a specific member from the set picked above. Member 0
146 should normally correspond to the central value, with higher values
147 corresponding to different error PDF's somewhat off in different
148 directions. You have to check from set to set which options are open.
149 <br/><b>Note:</b> you can only use one member in a run, so if you
150 want to sweep over many members you either have to do many separate
151 runs or, as a simplification, save the
152 <a href="EventInformation.html" target="page">pdf weights</a> at the hard scattering
153 and do an offline reweighting of events.
156 <p/><code>flag </code><strong> PDF:extrapolateLHAPDF </strong>
157 (<code>default = <strong>off</strong></code>)<br/>
158 Parton densities have a guaranteed range of validity in <i>x</i>
159 and <i>Q^2</i>, and what should be done beyond that range usually is
160 not explained by the authors of PDF sets. Nevertheless these boundaries
161 very often are exceeded, e.g. minimum-bias studies at LHC may sample
162 <i>x</i> values down to <i>10^-8</i>, while many PDF sets stop
163 already at <i>10^-5</i>. The default behaviour is then that the
164 PDF's are frozen at the boundary, i.e. <i>xf(x,Q^2)</i> is fixed at
165 its value at <i>x_min</i> for all values <i>x < x_min</i>,
166 and so on. This is a conservative approach. Alternatively, if you
167 switch on extrapolation, then parametrizations will be extended beyond
168 the boundaries, by some prescription. In some cases this will provide a
169 more realistic answer, in others complete rubbish. Another problem is
170 that some of the PDF-set codes will write a warning message anytime the
171 limits are exceeded, thus swamping your output file. Therefore you should
172 study a set seriously before you run it with this switch on.
176 If you want to use PDF's not found in LHAPDF, or you want to interface
177 LHAPDF another way, you have full freedom to use the more generic
178 <a href="PartonDistributions.html" target="page">interface options</a>.
180 <h3>Parton densities for protons in the hard process</h3>
182 The above options provides a PDF set that will be used everywhere:
183 for the hard process, the parton showers and the multiple interactions
184 alike. As already mentioned, therefore a change of PDF should be
185 accompanied by a <b>complete</b> retuning of the whole MI framework,
186 and maybe more. There are cases where one may want to explore
187 different PDF options for the hard process, but would not want to touch
188 the rest. If several different sets are to be compared, a simple
189 reweighting based on the <a href="EventInformation.html" target="page">originally
190 used</a> flavour, <i>x</i>, <i>Q^2</i> and PDF values may offer the
191 best route. The options in this section allow a choice of the PDF set
192 for the hard process alone, while the choice made in the previous section
193 would still be used for everything else. The hardest interaction
194 of the minimum-bias process is part of the multiple-interactions
195 framework and so does not count as a hard process here.
198 Of course it is inconsistent to use different PDF's in different parts
199 of an event, but if the <i>x</i> and <i>Q^2</i> ranges mainly accessed
200 by the components are rather different then the contradiction would not be
201 too glaring. Furthermore, since standard PDF's are one-particle-inclusive
202 we anyway have to 'invent' our own PDF modifications to handle configurations
203 where more than one parton is kicked out of the proton [<a href="Bibliography.html" target="page">Sjo04</a>].
206 The PDF choices that can be made are the same as above, so we do not
207 repeat the detailed discussion.
209 <p/><code>flag </code><strong> PDF:useHard </strong>
210 (<code>default = <strong>off</strong></code>)<br/>
211 If on then select a separate PDF set for the hard process, using the
212 variables below. If off then use the same PDF set for everything,
213 as already chosen above.
216 <p/><code>mode </code><strong> PDF:pHardSet </strong>
217 (<code>default = <strong>2</strong></code>; <code>minimum = 1</code>; <code>maximum = 12</code>)<br/>
218 Parton densities to be used for proton beams (and, by implication,
220 <br/><code>option </code><strong> 1</strong> : GRV 94L, LO <i>alpha_s(M_Z) = 0.128</i>
222 <br/><code>option </code><strong> 2</strong> : CTEQ 5L, LO <i>alpha_s(M_Z) = 0.127</i>
223 (slightly out of date; many tunes are based on it).
224 <br/><code>option </code><strong> 3</strong> : MRST LO* (2007),
225 NLO <i>alpha_s(M_Z) = 0.12032</i>.
226 <br/><code>option </code><strong> 4</strong> : MRST LO** (2008),
227 NLO <i>alpha_s(M_Z) = 0.11517</i>.
228 <br/><code>option </code><strong> 5</strong> : MSTW 2008 LO (central member),
229 LO <i>alpha_s(M_Z) = 0.13939</i>.
230 <br/><code>option </code><strong> 6</strong> : MSTW 2008 NLO (central member),
231 LO <i>alpha_s(M_Z) = 0.12018</i>.
232 <br/><code>option </code><strong> 7</strong> : CTEQ6L, NLO <i>alpha_s(M_Z) = 0.1180</i>.
233 <br/><code>option </code><strong> 8</strong> : CTEQ6L1, LO <i>alpha_s(M_Z) = 0.1298</i>.
234 <br/><code>option </code><strong> 9</strong> : CTEQ66.00 (NLO, central member),
235 NLO <i>alpha_s(M_Z) = 0.1180</i>.
236 <br/><code>option </code><strong> 10</strong> : CT09MC1, LO <i>alpha_s(M_Z) = 0.1300</i>.
237 <br/><code>option </code><strong> 11</strong> : CT09MC2, NLO <i>alpha_s(M_Z) = 0.1180</i>.
238 <br/><code>option </code><strong> 12</strong> : CT09MCS, NLO <i>alpha_s(M_Z) = 0.1180</i>.
241 <p/><code>flag </code><strong> PDF:useHardLHAPDF </strong>
242 (<code>default = <strong>off</strong></code>)<br/>
243 If off then the choice of proton PDF is based on <code>hardpPDFset</code>
244 above. If on then it is instead based on the choice of
245 <code>hardLHAPDFset</code> and <code>hardLHAPDFmember</code> below.
248 <p/><code>word </code><strong> PDF:hardLHAPDFset </strong>
249 (<code>default = <strong>MRST2004FF4lo.LHgrid</strong></code>)<br/>
250 Name of proton PDF set from LHAPDF to be used.
253 <p/><code>mode </code><strong> PDF:hardLHAPDFmember </strong>
254 (<code>default = <strong>0</strong></code>; <code>minimum = 0</code>)<br/>
255 Further choice of a specific member from the set picked above.
259 Note that there is no separate equivalent of the
260 <code>PDF:extrapolateLHAPDF</code> flag specifically for the hard
261 PDF. Since LHAPDF only has one global flag for extrapolation or not,
262 the choice for the normal PDF's also applies to the hard ones.
264 <h3>Parton densities for pions</h3>
266 The parton densities of the pion are considerably less well known than
267 those of the proton. There are only rather few sets on the market,
268 and none particularly recent. Only one comes built-in, but others can
269 be accessed from LHAPDF. Input parametrizations are for the <i>pi+</i>.
270 >From this the <i>pi-</i> is obtained by charge conjugation and the
271 <i>pi0</i> from averaging (half the pions have <i>d dbar</i>
272 valence quark content, half <i>u ubar</i>.
275 Much of the switches are taken over from the proton case, with obvious
276 modifications; therefore the description is briefer. Currently we have
277 not seen the need to allow separate parton densities for hard processes.
278 When using LHAPDF the <code>PDF:extrapolateLHAPDF</code> switch of the
279 proton also applies to pions.
281 <p/><code>mode </code><strong> PDF:piSet </strong>
282 (<code>default = <strong>1</strong></code>; <code>minimum = 1</code>; <code>maximum = 1</code>)<br/>
283 Internal parton densities that can be used for pion beams, currently with
285 <br/><code>option </code><strong> 1</strong> : GRV 92 L.
288 <p/><code>flag </code><strong> PDF:piUseLHAPDF </strong>
289 (<code>default = <strong>off</strong></code>)<br/>
290 If off then the choice of proton PDF is based on <code>PDF:piSet</code>
291 above. If on then it is instead based on the choice of
292 <code>PDF:piLHAPDFset</code> and <code>PDF:piLHAPDFmember</code> below.
295 <p/><code>word </code><strong> PDF:piLHAPDFset </strong>
296 (<code>default = <strong>OWPI.LHgrid</strong></code>)<br/>
297 Name of pion PDF set from LHAPDF to be used. You have to choose from the
298 <a href="http://projects.hepforge.org/lhapdf/pdfsets" target="page">
299 list of available sets</a>.
302 <p/><code>mode </code><strong> PDF:piLHAPDFmember </strong>
303 (<code>default = <strong>0</strong></code>; <code>minimum = 0</code>)<br/>
304 Further choice of a specific member from the set picked above.
307 <h3>Parton densities for Pomerons</h3>
309 The Pomeron is introduced in the description of diffractive events,
310 i.e. a diffractive system is viewed as a Pomeron-proton collision at a
311 reduced CM energy. Here the PDF's are even less well known.
312 Most experimental parametrizations are NLO, which makes them less
313 well suited for Monte Carlo applications. Furthemore note that
314 the momentum sum is arbitrarily normalized to a non-unity value.
316 <p/><code>mode </code><strong> PDF:PomSet </strong>
317 (<code>default = <strong>6</strong></code>; <code>minimum = 1</code>; <code>maximum = 6</code>)<br/>
318 Parton densities that can be used for Pomeron beams.
319 <br/><code>option </code><strong> 1</strong> : <i>Q^2</i>-independent parametrizations
320 <i>xf(x) = N_ab x^a (1 - x)^b</i>, where <i>N_ab</i> ensures
321 unit momentum sum. The <i>a</i> and <i>b</i> parameters can be
322 set separately for the gluon and the quark distributions. The
323 momentum fraction of gluons and quarks can be freely mixed, and
324 production of <i>s</i> quarks can be suppressed relative to
325 that of <i>d</i> and <i>u</i> ones, with antiquarks as likely
326 as quarks. See further below how to set the six parameters of this
329 <br/><code>option </code><strong> 2</strong> : <i>pi0</i> distributions, as specified in the
332 <br/><code>option </code><strong> 3</strong> : the H1 2006 Fit A NLO <i>Q^2</i>-dependent
333 parametrization, based on a tune to their data [<a href="Bibliography.html" target="page">H1P06</a>],
334 rescaled by the factor <code>PomRescale</code> below.
336 <br/><code>option </code><strong> 4</strong> : the H1 2006 Fit B NLO <i>Q^2</i>-dependent
337 parametrization, based on a tune to their data [<a href="Bibliography.html" target="page">H1P06</a>],
338 rescaled by the factor <code>PomRescale</code> below.
340 <br/><code>option </code><strong> 5</strong> : the H1 2007 Jets NLO <i>Q^2</i>-dependent
341 parametrization, based on a tune to their data [<a href="Bibliography.html" target="page">H1P07</a>],
342 rescaled by the factor <code>PomRescale</code> below.
344 <br/><code>option </code><strong> 6</strong> : the H1 2006 Fit B LO <i>Q^2</i>-dependent
345 parametrization, based on a tune to their data [<a href="Bibliography.html" target="page">H1P06</a>],
346 rescaled by the factor <code>PomRescale</code> below.
350 <p/><code>parm </code><strong> PDF:PomGluonA </strong>
351 (<code>default = <strong>0.</strong></code>; <code>minimum = -0.5</code>; <code>maximum = 2.</code>)<br/>
352 the parameter <i>a</i> in the ansatz <i>xg(x) = N_ab x^a (1 - x)^b</i>
356 <p/><code>parm </code><strong> PDF:PomGluonB </strong>
357 (<code>default = <strong>3.</strong></code>; <code>minimum = 0.</code>; <code>maximum = 10.</code>)<br/>
358 the parameter <i>b</i> in the ansatz <i>xg(x) = N_ab x^a (1 - x)^b</i>
362 <p/><code>parm </code><strong> PDF:PomQuarkA </strong>
363 (<code>default = <strong>0.</strong></code>; <code>minimum = -0.5</code>; <code>maximum = 2.</code>)<br/>
364 the parameter <i>a</i> in the ansatz <i>xq(x) = N_ab x^a (1 - x)^b</i>
368 <p/><code>parm </code><strong> PDF:PomQuarkB </strong>
369 (<code>default = <strong>3.</strong></code>; <code>minimum = 0.</code>; <code>maximum = 10.</code>)<br/>
370 the parameter <i>b</i> in the ansatz <i>xq(x) = N_ab x^a (1 - x)^b</i>
374 <p/><code>parm </code><strong> PDF:PomQuarkFrac </strong>
375 (<code>default = <strong>0.2</strong></code>; <code>minimum = 0.</code>; <code>maximum = 1.</code>)<br/>
376 the fraction of the Pomeron momentum carried by quarks
377 for option 1 above, with the rest carried by gluons.
380 <p/><code>parm </code><strong> PDF:PomStrangeSupp </strong>
381 (<code>default = <strong>0.5</strong></code>; <code>minimum = 0.</code>; <code>maximum = 1.</code>)<br/>
382 the suppression of the <i>s</i> quark density relative to that of the
383 <i>d</i> and <i>u</i> ones for option 1 above.
386 <p/><code>parm </code><strong> PDF:PomRescale </strong>
387 (<code>default = <strong>1.0</strong></code>; <code>minimum = 0.5</code>; <code>maximum = 5.0</code>)<br/>
388 Rescale the four H1 fits above by this uniform factor, e.g. to bring
389 up their momentum sum to around unity. By default all three have
390 a momentum sum of order 0.5, suggesting that a factor around 2.0
391 should be used. You can use <code>examples/main41.cc</code> to get
392 a more precise value. Note that also other parameters in the
393 <a href="Diffraction.html" target="page">diffraction</a> framework may need to
394 be retuned when this parameter is changed.
397 <h3>Parton densities for leptons</h3>
399 For electrons/leptons there is no need to choose between different
400 parametrizations, since only one implementation is available, and
401 should be rather uncontroversial (apart from some technical details).
402 However, insofar as e.g. <i>e^+ e^-</i> data often are corrected
403 back to a world without any initial-state photon radiation, it is
404 useful to have a corresponding option available here.
406 <p/><code>flag </code><strong> PDF:lepton </strong>
407 (<code>default = <strong>on</strong></code>)<br/>
408 Use parton densities for lepton beams or not. If off the colliding
409 leptons carry the full beam energy, if on part of the energy is
410 radiated away by initial-state photons. In the latter case the
411 initial-state showers will generate the angles and energies of the
412 set of photons that go with the collision. In addition one collinear
413 photon per beam carries any leftover amount of energy not described
414 by shower emissions. If the initial-state showers are switched off
415 these collinear photons will carry the full radiated energy.
418 <h3>Incoming parton selection</h3>
420 There is one useful degree of freedom to restrict the set of incoming
421 quark flavours for hard processes. It does not change the PDF's as such,
422 only which quarks are allowed to contribute to the hard-process cross
423 sections. Note that separate but similarly named modes are available
424 for multiple interactions and spacelike showers.
426 <p/><code>mode </code><strong> PDFinProcess:nQuarkIn </strong>
427 (<code>default = <strong>5</strong></code>; <code>minimum = 0</code>; <code>maximum = 5</code>)<br/>
428 Number of allowed incoming quark flavours in the beams; a change
429 to 4 would thus exclude <i>b</i> and <i>bbar</i> as incoming
436 <!-- Copyright (C) 2010 Torbjorn Sjostrand -->