Period map of the Dwork quartic K3 family

edit · history · discussion · files · short url · period modular function special values
Numbers
$t$ 
$P(t)$
0:
1/2 + i*1/2
1/2:
0.3557494653356022809351377615008825598564274388039376440663697253911986199088299135481818901603557202 + i*0.6110992700972838758971165517928815061499794823142292540556124058901480717585141341242637599742104808
2:
i*1.319676446447160358511590485409009448980441637939271382975125053493778995506162002405834584227564825
1/3:
0.4112547398722945739768227851426013602918058825635239289698569554198675349809185769496500295825515213 + i*0.5752126032455924235676714167810084598862738873156991006720222491640155793970935489795467482535560167
2/3:
0.2894140617897159893491940923730476660532029538196813413510196828465948771933283423294903139860799748 + i*0.6451662582919060680811158329996396271267869084827529439817952164190645069066869497808374036979344266
3/2:
i*1.126854934770877718873360632024233938679648863943350934165329740268049093598191591647559534705046786
3:
i*1.581138830084189665999446772216359266859777569662608413428752426396297219319619110672124054189650148
1/4:
0.4359611368939099187985960856100973067748802370882400544832744297315075661988194862731353455817291222 + i*0.5567206544741891789179625714902372594840217939890498273196128302011356729012463237339312551594531914
3/4:
1/4 + i*0.6614378277661476476254039384098151064275647957706125450920836148002672058075709069400982216186359027
4/3:
i*1.042324463428387124228613186531130066504683152150512445391766147723270180350854944674060216548062382
4:
i*1.764831872812151059183744729872439434719258703287716425391165481090531021485475097875945097774917041
1/5:
0.4499313747407122795416993775191090947696195665635099977166350573989006926857759437373018705639067185 + i*0.5454922162817107843260650400246020495352913627906874252071664042856281246729224160544284826928979774
2/5:
0.3901040771918746033294684367245251946211162841511342072558104043127500539933233957012457074982760424 + i*0.5897616543640964293411411789753992042847856443636835364192239579484715698261926660388878912751873138
3/5:
0.3175806736473533984247704707986680265763117488732477245336680904723510682547953631250748460129720846 + i*0.6317772674967762835418668282699505827720219802286733148512738028101520439552588401975930917957285222
4/5:
0.2232207582083288142927421608678567957312577500800185100212795099824107896977457901374680778429720024 + i*0.6709489497010177184030725603874287825208656998783154407032496273908744467979575192050421536601976650
5/4:
i*0.9928869047742279829628866264068202257927105613287148505160587097072073659516994959025133163326816511
5/3:
i*1.198934488330625177340095210361209025468821239912970682425430855350152835925357901604950679704064299
5/2:
i*1.464200197523786337580603552287685753991753124927398503334859488805239956046468951465592630787553754
5:
i*1.907038886495286573464475244239025327160963913168595906692679421368751015592884832826789735454514247
1/6:
0.4589079045890776990385661257676171054835069114578186900492938008197239520781362623961491689538299096 + i*0.5379623919064063713643065455134987373196213244111385872105009503547321132074850356916060636137787270
5/6:
0.2035225140355705959356378098411210279859845946222259426201083396727020728507727246201801147009715724 + i*0.6771843074677978461787519902692211194959701505487905416048284559898705169136299212966147370395037000
6/5:
i*0.9597181528135624205412609645381539472515250485411915647903169184512651698482598145793356065811057505
6:
i*2.023162002475728672237185236353393968336835820723864815153772608903464579617735471763947994577326203
Definition
The table gives the scalar period coordinate $P(t)$ of the mirror-quartic K3 surface $X_t$, with Hartmann's integral lattice marking and the continuation lift fixed here [1].
Parameters
$t$
—   quartic coefficient ($t\in\mathbb Q$, $t\geq0$, $t\ne1$)
Formulas
(1)
Let $H_n=\sum_{j=1}^{n}1/j$, $H_0=0$, $A_n=(4n)!/(256^n(n!)^4)$, $F(z)=\sum_{n\geq0}A_nz^n$ and $L(z)=F(z)\log(z/256)+4\sum_{n\geq0}A_n(H_{4n}-H_n)z^n$. Then $P(t)=L(t^{-4})/(2\pi iF(t^{-4}))$, initially for $t>1$, with the continuation in (9) [1], Example 4.27.
(2)
Let $K(m)=\int_0^{\pi/2}(1-m\sin^2\theta)^{-1/2}d\theta$ use the elliptic parameter $m$. Under $z=16(1-\lambda)\lambda^2/(\lambda-2)^4$, the branch $\lambda\sim\sqrt z>0$ at $z=0^+$ gives $P=iK(1-\lambda)/K(\lambda)$ [2], equations (5.3), (5.15)–(5.17). Transport both periods on the specified lift.
(3)
Explicitly, $s=\sqrt{1-z}$, $u=\sqrt{(1-s)/2}$ and $\lambda=2u/(1+u)$. For $0<z<1$ both roots are positive real; for upper-bank $z>1$, take $s=-i\sqrt{z-1}$ and the root $u$ with positive real part. Substitution gives $z=4u^2(1-u^2)$ and (2).
(4)
$P(0)=(1+i)/2$ and $P(3/4)=(1+i\sqrt7)/4$. For $P(3/4)$, the exact algebraic pullback and the lambda duplication and modular transformations [4] [5] imply $2P^2-P+1=0$ on the selected lift; the attached mathematical-checks.md gives the branch and injectivity certificate.
(5)
For $\operatorname{Im}\tau>0$, put $q=\exp(2\pi i\tau)$ and $\eta(\tau)=q^{1/24}\prod_{n\geq1}(1-q^n)$. Then $\rho(\tau)=256\eta(\tau)^{24}\eta(2\tau)^{24}/(\eta(\tau)^{24}+64\eta(2\tau)^{24})^2$ satisfies $\rho(P(t))=t^{-4}$ for $t>0$ [3], equations (2.4)–(2.5). This identity alone does not distinguish integer shifts of $P$.
(6)
With the zero-constant normalization in the McKay–Thompson table, $T_{2A}(P(t))=256t^4-104$. Indeed, if $R=(\eta(\tau)/\eta(2\tau))^{24}$, then $T_{2A}=R+24+4096/R$ and $256/\rho=R+128+4096/R$.
Comments
(7)
The surface $X_t$ is the minimal resolution of $F_t/G$, where $F_t:X_0^4+X_1^4+X_2^4+X_3^4-4tX_0X_1X_2X_3=0$. The quotient group is $G=\{(a_0,a_1,a_2,a_3)\in\mu_4^4:a_0a_1a_2a_3=1\}/\mu_4$, where $\mu_4$ is the group of fourth roots of unity and the denominator acts diagonally in projective space. The marked surface is the minimal resolution of $F_t/G$, rather than an unspecified marking of $F_t$ [1]. The excluded singular fibres have $t^4=1$ or $t=\infty$; the origin is smooth.
(8)
Use the rank-three lattice $T_0=\langle4\rangle\oplus U$ in Hartmann's Theorem 4.8, where $U$ is the hyperbolic plane. In the ordered basis $(h,e,f)$, $h^2=4$, $e\cdot f=1$, and all other pairings are zero. The coordinate is fixed by $[\Omega_t]=[P(t)h-e+2P(t)^2f]$ [1], equation (10) and Theorem 4.29. Poincaré duality identifies the basis with cycles, so $P(t)=-\int_h\Omega_t/(4\int_f\Omega_t)$. The common differential scale cancels.
(9)
Put $z=t^{-4}$ for $t>0$. Start the Frobenius ratio in (1) on $0<z<1$ with the real logarithm. Continue through the upper $z$-half-plane without winding around $0$ or $1$; real $z>1$ uses the upper-bank value. This fixes a lift, with no subsequent modular reduction. The smooth origin uses its limiting value $P(0)=(1+i)/2$ [1], Theorem 1.1 and proof of Theorem 4.37.
(10)
For $t>1$ the value is purely positive imaginary. For $0<t<1$ it lies on the arc $0<\operatorname{Re}P<1/2$, $\operatorname{Im}P>0$, $|P|^2=1/2$. The singular limit $t\to1$ is $i/\sqrt2$ and is not an entry [1]. The isolated vertex $i/2$ in the statement of Theorem 4.37 conflicts with its proof, introduction and Theorem 4.28; the defining germ and these consistent limits fix the normalization.
(11)
The geometric exponential coordinate is $q_{K3}=\exp(2\pi iP)$, while the elliptic theta nome is $q_{\rm ell}=\exp(\pi iP)$, so $q_{K3}=q_{\rm ell}^2$ [2]. Recovering $P$ from either exponential requires the same logarithm lift.
(12)
The positive rational fibres form a bounded exploratory reference slice of small-coefficient quartic models. The height bound six is noncanonical; no source is claimed to use every selected coefficient. Both smooth regimes are retained without selecting for known CM relations. The attached argument-selection.md describes this choice.
Programs
(P1)
Sage
import numberdb.sage as numberdb
from sage.rings.rational_field import QQ
from sage.rings.complex_arb import ComplexBallField

def period_map(t, digits=100):
    t = QQ(t)
    if t < 0 or t == 1:
        raise ValueError("t must be nonnegative and different from 1")
    C = ComplexBallField(numberdb.bits(digits, losing=180))
    i = C(0, 1)
    if t == 0:
        return C(QQ(1)/2, QQ(1)/2)
    z = C(t**(-4))
    s = (1-z).sqrt() if t > 1 else -i*(z-1).sqrt()
    u = ((1-s)/2).sqrt()
    m = 2*u/(1+u)
    v = i*(1-m).elliptic_k()/m.elliptic_k()
    return C(QQ(1)/4, v.imag()) if t == QQ(3)/4 else v

period_map(QQ(5)/6)
# Stored selection: t=0 and every reduced a/b != 1 with 1<=a,b<=6,
# ordered by height max(a,b), then by value within each height.
References
[1]
Heinrich Hartmann, Period- and Mirror-maps for the Quartic K3 (2011), Sections 4.1–4.8, Theorem 4.8, equation (10), Example 4.27, Theorem 4.29 and proof of Theorem 4.37. (arXiv)
[2]
Wenzhe Yang, K3 mirror symmetry, Legendre family and Deligne's conjecture for Fermat quartic, Nuclear Physics B 959 (2020), 115303, equations (5.3), (5.15)–(5.17). (arXiv) (doi)
[3]
Lassina Dembélé, Alexei Panchishkin, John Voight and Wadim Zudilin, Special hypergeometric motives and their L-functions: Asai recognition (2020), equations (2.4)–(2.5). (arXiv)
Links
Similar tables
Complete elliptic integral of the first kind $K(m)$ —   gives individual real elliptic periods used in the algebraic pullback, rather than this marked ratio
Values of the elliptic nome $q(m)$ —   uses $q=\exp(\pi iP)$ at rational Legendre parameters; here the Legendre parameter is an algebraic pullback of the quartic coefficient
Values of monstrous McKay–Thompson series at rectangular CM points —   gives modular-function values related by (6); the present table evaluates their geometrically marked inverse coordinate
Pham periods of Fermat varieties —   includes the dimension-two degree-four Fermat surface with a different integral marking and differential convention; its individual periods are not identified with this ratio
Normalized holomorphic periods of hypergeometric Calabi–Yau families —   gives rank-four periods in a different family; this K3 period system has rank three
Canonical coordinates (mirror maps) of hypergeometric Calabi–Yau families —   gives exponential coordinates for the rank-four families, with different normalizations
Data properties
Entries are of type: complex number
Table is complete: no (it holds the smooth Fermat-quotient fibre $t=0$ and all 22 positive reduced rational coefficients $t=a/b\ne1$ with $1\leq a,b\leq6$, ordered by height and then by value. The noncanonical height bound gives a whole exploratory slice of simple rational quartic models on both sides of the singular fibre, without filtering by CM status. Other rational coefficients remain uncomputed)
How they were obtained:

Arb ball arithmetic encloses exact rational coefficients, the algebraic pullback and the ratio of elliptic integrals at 513 working bits, with the denominator separated from zero. The 100-significant-digit decimals enclose the resulting component balls. The exact origin, exactly zero real parts for $t>1$, and the real component $1/4$ at $t=3/4$ are preserved; the component $1/4$ has an exact lambda-transformation certificate in mathematical-checks.md.

more

Every inexact fibre was also checked by separately implemented Picard–Fuchs Taylor continuation in mpmath at 145 and 180 decimal digits, using the Frobenius germ rather than the elliptic pullback. Finite eta products with analytic tail bounds check the modular inverse at every nonzero fibre, including propagated stored-interval uncertainty and inverse conditioning. Known-answer and deliberately wrong branch, normalization, parameter and enclosure controls are retained in the attached executable checks and results. The Picard–Fuchs convergence check is numerical, not a separate rigorous enclosure proof; production bounds come from Arb.