This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Integer group determinants for C24{\rm C}_{2}^{4}

Yuka Yamaguchi and Naoya Yamaguchi
Abstract.

We determine all possible values of the integer group determinant of C24{\rm C}_{2}^{4}, where C2{\rm C}_{2} is the cyclic group of order 22.

Key words and phrases:
Group determinant, Integer group determinant, Cyclic group
1991 Mathematics Subject Classification:
11C20, 11E76, 20C15

1. Introduction

For a finite group GG, assigning a variable xgx_{g} for each gGg\in G, the group determinant of GG is defined as det(xgh1)g,hG\det{\left(x_{gh^{-1}}\right)}_{g,h\in G}. When the variables xgx_{g} are all integers, the group determinant is called an integer group determinant of GG. For a broad context of group determinants, see their use in the identification of a group [3], [7], [11]; the related abelian question of circulants [6], [9]; the Lind-Lehmer problem [1], [2], [9]; and the representation theoretic background [5].

Let S(G)S(G) denote the set of all possible values of the integer group determinant of GG:

S(G):={det(xgh1)g,hGxg}.S(G):=\left\{\det{\left(x_{gh^{-1}}\right)}_{g,h\in G}\mid x_{g}\in\mathbb{Z}\right\}.

Let C2{\rm C}_{2} be the cyclic group of order 22. The complete descriptions of S(C22)S\left({\rm C}_{2}^{2}\right) and S(C23)S\left({\rm C}_{2}^{3}\right) are obtained in [1, Theorem 5.3] and [10, Theorem 3.1], respectively:

  • S(C22)={4m+1, 24(2m+1), 26mm}S\left({\rm C}_{2}^{2}\right)=\left\{4m+1,\>2^{4}(2m+1),\>2^{6}m\mid m\in\mathbb{Z}\right\};

  • S(C23)={8m+1, 28(4m+1), 212mm}S\left({\rm C}_{2}^{3}\right)=\left\{8m+1,\>2^{8}(4m+1),\>2^{12}m\mid m\in\mathbb{Z}\right\}.

In this paper, we determine S(C24)S\left({\rm C}_{2}^{4}\right).

Theorem 1.1.

Let A:={(8k3)(8l+3)k,l}{8m1m}A:=\left\{(8k-3)(8l+3)\mid k,l\in\mathbb{Z}\right\}\subsetneq\left\{8m-1\mid m\in\mathbb{Z}\right\}. Then we have

S(C24)\displaystyle S\left({\rm C}_{2}^{4}\right) ={16m+1, 216(4m+1), 224(4m+1), 224(8m+3), 224m, 226mm,mA}.\displaystyle=\left\{16m+1,\>2^{16}(4m+1),\>2^{24}(4m+1),\>2^{24}(8m+3),\>2^{24}m^{\prime},\>2^{26}m\mid m\in\mathbb{Z},\>m^{\prime}\in A\right\}.

Let odd\mathbb{Z}_{\rm odd} be the set of all odd numbers. We remark that all of the odd values in S(C24)S\left({\rm C}_{2}^{4}\right) are already known: S(C2n)odd={2nm+1m}S\left({\rm C}_{2}^{n}\right)\cap\mathbb{Z}_{\rm odd}=\left\{2^{n}m+1\mid m\in\mathbb{Z}\right\} holds for any nn. For example, see [2, Lemmas 2.1 and 2.2]. Although S(C24)odd={16m+1m}S\left({\rm C}_{2}^{4}\right)\cap\mathbb{Z}_{\rm odd}=\left\{16m+1\mid m\in\mathbb{Z}\right\} is not new result, we give a proof in our method.

For every group GG of order at most 1515, S(G)S(G) was determined (see [8], [10]). For the groups of order 1616, the complete descriptions of S(G)S(G) were obtained for the dihedral group [1, Theorem 5.3] and the cyclic group [13]. There are fourteen groups of order 1616 up to isomorphism [4], [14]. Theorem 1.1 determines S(G)S(G) for one of the unsolved twelve groups.

2. Relations with group determinants of subgroups

For any g=(ε0¯,ε1¯,,εn1¯)C2ng=(\overline{\varepsilon_{0}},\overline{\varepsilon_{1}},\ldots,\overline{\varepsilon_{n-1}})\in{\rm C}_{2}^{n} with εi{0,1}\varepsilon_{i}\in\{0,1\}, we denote the variable xgx_{g} by xjx_{j}, where j:=εn12n1+εn22n2++ε020j:=\varepsilon_{n-1}\cdot 2^{n-1}+\varepsilon_{n-2}\cdot 2^{n-2}+\cdots+\varepsilon_{0}\cdot 2^{0}, and let

Dn(x0,x1,,x2n1):=det(xgh1)g,hC2n.D_{n}(x_{0},x_{1},\ldots,x_{2^{n}-1}):=\det{\left(x_{gh^{-1}}\right)}_{g,h\in{\rm C}_{2}^{n}}.

From the H=C2n1H={\rm C}_{2}^{n-1} and K=C2K={\rm C}_{2} case of [12, Theorem 1.1], we have the following corollary.

Corollary 2.1.

For any positive integer nn,

Dn+1(x0,,x2n+11)=Dn(x0+x2n,,x2n1+x2n+11)Dn(x0x2n,,x2n1x2n+11).D_{n+1}(x_{0},\ldots,x_{2^{n+1}-1})=D_{n}(x_{0}+x_{2^{n}},\ldots,x_{2^{n}-1}+x_{2^{n+1}-1})D_{n}(x_{0}-x_{2^{n}},\ldots,x_{2^{n}-1}-x_{2^{n+1}-1}).

Throughout this paper, we assume that a0,a1,,a15a_{0},a_{1},\ldots,a_{15}\in\mathbb{Z} and, for any 0i30\leq i\leq 3, put

bi\displaystyle b_{i} :=(ai+ai+8)+(ai+4+ai+12),\displaystyle:=(a_{i}+a_{i+8})+(a_{i+4}+a_{i+12}), ci\displaystyle c_{i} :=(ai+ai+8)(ai+4+ai+12),\displaystyle:=(a_{i}+a_{i+8})-(a_{i+4}+a_{i+12}),
di\displaystyle d_{i} :=(aiai+8)+(ai+4ai+12),\displaystyle:=(a_{i}-a_{i+8})+(a_{i+4}-a_{i+12}), ei\displaystyle e_{i} :=(aiai+8)(ai+4ai+12).\displaystyle:=(a_{i}-a_{i+8})-(a_{i+4}-a_{i+12}).

The following relations will be frequently used in this paper:

D4(a0,a1,,a15)\displaystyle D_{4}(a_{0},a_{1},\ldots,a_{15}) =D3(a0+a8,,a7+a15)D3(a0a8,,a7a15),\displaystyle=D_{3}(a_{0}+a_{8},\ldots,a_{7}+a_{15})D_{3}(a_{0}-a_{8},\ldots,a_{7}-a_{15}),
=D2(b0,b1,b2,b3)D2(c0,c1,c2,c3)D2(d0,d1,d2,d3)D2(e0,e1,e2,e3).\displaystyle=D_{2}(b_{0},b_{1},b_{2},b_{3})D_{2}(c_{0},c_{1},c_{2},c_{3})D_{2}(d_{0},d_{1},d_{2},d_{3})D_{2}(e_{0},e_{1},e_{2},e_{3}).
Remark 2.2.

The following hold:

  1. (1)(1)

    bicidiei(mod2)for  0i3b_{i}\equiv c_{i}\equiv d_{i}\equiv e_{i}\pmod{2}\>\>\text{for}\>\>0\leq i\leq 3;

  2. (2)(2)

    bi+ci+di+ei0(mod4)for  0i3b_{i}+c_{i}+d_{i}+e_{i}\equiv 0\pmod{4}\>\>\text{for}\>\>0\leq i\leq 3.

Noting that det(αst)s,tdet(βst)s,t(mod2)\det{(\alpha_{st})_{s,t}}\equiv\det{(\beta_{st})_{s,t}}\pmod{2} holds if αstβst(mod2)\alpha_{st}\equiv\beta_{st}\pmod{2} for any ss and tt, we have the following lemma.

Lemma 2.3.

The following hold:

  1. (1)(1)

    D4(a0,a1,,a15)D3(a0+a8,,a7+a15)D3(a0a8,,a7a15)(mod2)D_{4}(a_{0},a_{1},\ldots,a_{15})\equiv D_{3}(a_{0}+a_{8},\ldots,a_{7}+a_{15})\equiv D_{3}(a_{0}-a_{8},\ldots,a_{7}-a_{15})\pmod{2};

  2. (2)(2)

    D4(a0,a1,,a15)D2(b0,b1,b2,b3)D2(c0,c1,c2,c3)D_{4}(a_{0},a_{1},\ldots,a_{15})\equiv D_{2}(b_{0},b_{1},b_{2},b_{3})\equiv D_{2}(c_{0},c_{1},c_{2},c_{3})
                                 D2(d0,d1,d2,d3)D2(e0,e1,e2,e3)(mod2)\equiv D_{2}(d_{0},d_{1},d_{2},d_{3})\equiv D_{2}(e_{0},e_{1},e_{2},e_{3})\pmod{2}.

Lemma 2.4.

We have D4(a0,a1,,a15)2b0+b2b1+b3(mod2)D_{4}(a_{0},a_{1},\ldots,a_{15})\in 2\mathbb{Z}\iff b_{0}+b_{2}\equiv b_{1}+b_{3}\pmod{2}.

Proof.

From Lemma 2.3 (2) and

D2(b0,b1,b2,b3)\displaystyle D_{2}(b_{0},b_{1},b_{2},b_{3}) =D1(b0+b2,b1+b3)D1(b0b2,b1b3)\displaystyle=D_{1}(b_{0}+b_{2},b_{1}+b_{3})D_{1}(b_{0}-b_{2},b_{1}-b_{3})
={(b0+b2)2(b1+b3)2}{(b0b2)2(b1b3)2}\displaystyle=\left\{(b_{0}+b_{2})^{2}-(b_{1}+b_{3})^{2}\right\}\left\{(b_{0}-b_{2})^{2}-(b_{1}-b_{3})^{2}\right\}
{2,b0+b2b1+b3(mod2),odd,b0+b2b1+b3(mod2),\displaystyle\in\begin{cases}2\mathbb{Z},&b_{0}+b_{2}\equiv b_{1}+b_{3}\pmod{2},\\ \mathbb{Z}_{\rm odd},&b_{0}+b_{2}\not\equiv b_{1}+b_{3}\pmod{2},\end{cases}

the lemma is proved. ∎

3. Group determinant of C22{\rm C}_{2}^{2}

In this section, we prove five lemmas which give properties of the group determinant of C22{\rm C}_{2}^{2}. These lemmas are used in the next section.

By direct calculation, we have the following lemma.

Lemma 3.1.

The identity D2(x0,x1,x2,x3)=i=03xi420i<j3xi2xj2+8x0x1x2x3D_{2}(x_{0},x_{1},x_{2},x_{3})=\sum_{i=0}^{3}x_{i}^{4}-2\sum_{0\leq i<j\leq 3}x_{i}^{2}x_{j}^{2}+8x_{0}x_{1}x_{2}x_{3} holds. Therefore, D2(x0,x1,x2,x3)D_{2}(x_{0},x_{1},x_{2},x_{3}) is a symmetric polynomial in x0,x1,x2,x3x_{0},x_{1},x_{2},x_{3}.

Lemma 3.2.

For any k,l,m,nk,l,m,n\in\mathbb{Z}, the following hold:

  1. (1)(1)

    D2(2k,2l,2m,2n+1)8(k+l+m)+1(mod16)D_{2}(2k,2l,2m,2n+1)\equiv 8(k+l+m)+1\pmod{16};

  2. (2)(2)

    D2(2k,2l+1,2m+1,2n+1)8k3(mod16)D_{2}(2k,2l+1,2m+1,2n+1)\equiv 8k-3\pmod{16}.

Proof.

We have

D2(2k,2l,2m,2n+1)\displaystyle D_{2}(2k,2l,2m,2n+1)
=D1(2k+2m,2l+2n+1)D1(2k2m,2l2n1)\displaystyle\hskip 14.22636pt=D_{1}(2k+2m,2l+2n+1)D_{1}(2k-2m,2l-2n-1)
={4(k+m)24(l+n)24(l+n)1}{4(km)24(ln)2+4(ln)1}\displaystyle\hskip 14.22636pt=\left\{4(k+m)^{2}-4(l+n)^{2}-4(l+n)-1\right\}\left\{4(k-m)^{2}-4(l-n)^{2}+4(l-n)-1\right\}
4(k+m)2+4(l+n)2+4(l+n)4(km)2+4(ln)24(ln)+1\displaystyle\hskip 14.22636pt\equiv-4(k+m)^{2}+4(l+n)^{2}+4(l+n)-4(k-m)^{2}+4(l-n)^{2}-4(l-n)+1
8k28m2+8l2+8n(n+1)+1\displaystyle\hskip 14.22636pt\equiv-8k^{2}-8m^{2}+8l^{2}+8n(n+1)+1
8(k+l+m)+1(mod16),\displaystyle\hskip 14.22636pt\equiv 8(k+l+m)+1\pmod{16},
D2(2k,2l+1,2m+1,2n+1)\displaystyle D_{2}(2k,2l+1,2m+1,2n+1)
=D1(2k+2m+1,2l+2n+2)D1(2k2m1,2l2n)\displaystyle\hskip 14.22636pt=D_{1}(2k+2m+1,2l+2n+2)D_{1}(2k-2m-1,2l-2n)
={4(k+m)2+4(k+m)+14(l+n+1)2}{4(km)24(km)+14(ln)2}\displaystyle\hskip 14.22636pt=\left\{4(k+m)^{2}+4(k+m)+1-4(l+n+1)^{2}\right\}\left\{4(k-m)^{2}-4(k-m)+1-4(l-n)^{2}\right\}
4(k+m)2+4(k+m)4(l+n+1)2+4(km)24(km)4(ln)2+1\displaystyle\hskip 14.22636pt\equiv 4(k+m)^{2}+4(k+m)-4(l+n+1)^{2}+4(k-m)^{2}-4(k-m)-4(l-n)^{2}+1
8k2+8m(m+1)8l(l+1)8n(n+1)3\displaystyle\hskip 14.22636pt\equiv 8k^{2}+8m(m+1)-8l(l+1)-8n(n+1)-3
8k3(mod16).\displaystyle\hskip 14.22636pt\equiv 8k-3\pmod{16}.

Lemma 3.3.

For any k,l,m,nk,l,m,n\in\mathbb{Z}, let α:=D2(2k,2l,2m,2n)\alpha:=D_{2}(2k,2l,2m,2n). The following hold:

  1. (1)(1)

    If k+ml+n,kmln(mod2)k+m\not\equiv l+n,\>km\equiv ln\pmod{2}, then α{24(8a+1)a}\alpha\in\left\{2^{4}(8a+1)\mid a\in\mathbb{Z}\right\};

  2. (2)(2)

    If k+ml+n,kmln(mod2)k+m\not\equiv l+n,\>km\not\equiv ln\pmod{2}, then α{24(8a3)a}\alpha\in\left\{2^{4}(8a-3)\mid a\in\mathbb{Z}\right\};

  3. (3)(3)

    If k,l,m,nk,l,m,n are even and k+ml+n(mod4)k+m\not\equiv l+n\pmod{4}, then α{28(4a+1)a}\alpha\in\left\{2^{8}(4a+1)\mid a\in\mathbb{Z}\right\};

  4. (4)(4)

    If k,l,m,nk,l,m,n are odd and k+ml+n(mod4)k+m\not\equiv l+n\pmod{4}, then α{28(4a1)a}\alpha\in\left\{2^{8}(4a-1)\mid a\in\mathbb{Z}\right\};

  5. (5)(5)

    If klmn(mod2)k\equiv l\equiv m\equiv n\pmod{2} and k+ml+n(mod4)k+m\equiv l+n\pmod{4}, then α212\alpha\in 2^{12}\mathbb{Z};

  6. (6)(6)

    If exactly two of k,l,m,nk,l,m,n are even numbers, then α210\alpha\in 2^{10}\mathbb{Z}.

Proof.

First, we prove (1) and (2). We have (k+m)2(l+n)2=(km)2(ln)2+4(kmln)(k+m)^{2}-(l+n)^{2}=(k-m)^{2}-(l-n)^{2}+4(km-ln). Therefore, if k+ml+n(mod2)k+m\not\equiv l+n\pmod{2}, then

24α\displaystyle 2^{-4}\alpha ={(k+m)2(l+n)2}{(km)2(ln)2}\displaystyle=\left\{(k+m)^{2}-(l+n)^{2}\right\}\left\{(k-m)^{2}-(l-n)^{2}\right\}
{1(mod8),kmln(mod2),3(mod8),kmln(mod2).\displaystyle\equiv\begin{cases}1\pmod{8},&km\equiv ln\pmod{2},\\ -3\pmod{8},&km\not\equiv ln\pmod{2}.\end{cases}

Second, we prove (3). If k,l,m,nk,l,m,n are even and k+ml+n(mod4)k+m\not\equiv l+n\pmod{4}, then there exist k,l,m,nk^{\prime},l^{\prime},m^{\prime},n^{\prime}\in\mathbb{Z} satisfying k=2kk=2k^{\prime}, l=2ll=2l^{\prime}, m=2mm=2m^{\prime}, n=2nn=2n^{\prime} and k+ml+n(mod2)k^{\prime}+m^{\prime}\not\equiv l^{\prime}+n^{\prime}\pmod{2}. Therefore,

28α={(k+m)2(l+n)2}{(km)2(ln)2}1(mod4).2^{-8}\alpha=\left\{(k^{\prime}+m^{\prime})^{2}-(l^{\prime}+n^{\prime})^{2}\right\}\left\{(k^{\prime}-m^{\prime})^{2}-(l^{\prime}-n^{\prime})^{2}\right\}\equiv 1\pmod{4}.

Third, we prove (4). If k,l,m,nk,l,m,n are odd and k+ml+n(mod4)k+m\not\equiv l+n\pmod{4}, then there exist k,l,m,nk^{\prime},l^{\prime},m^{\prime},n^{\prime}\in\mathbb{Z} satisfying k=2k+1k=2k^{\prime}+1, l=2l+1l=2l^{\prime}+1, m=2m+1m=2m^{\prime}+1, n=2n+1n=2n^{\prime}+1 and k+ml+n(mod2)k^{\prime}+m^{\prime}\not\equiv l^{\prime}+n^{\prime}\pmod{2}. Therefore,

28α={(k+m+1)2(l+n+1)2}{(km)2(ln)2}1(mod4).\displaystyle 2^{-8}\alpha=\left\{(k^{\prime}+m^{\prime}+1)^{2}-(l^{\prime}+n^{\prime}+1)^{2}\right\}\left\{(k^{\prime}-m^{\prime})^{2}-(l^{\prime}-n^{\prime})^{2}\right\}\equiv-1\pmod{4}.

Fourth, we prove (5). If klmn(mod2)k\equiv l\equiv m\equiv n\pmod{2} and k+ml+n(mod4)k+m\equiv l+n\pmod{4}, then (k+m)2(l+n)2(km)2(ln)20(mod16)(k+m)^{2}-(l+n)^{2}\equiv(k-m)^{2}-(l-n)^{2}\equiv 0\pmod{16}. Therefore,

α=24{(k+m)2(l+n)2}{(km)2(ln)2}212.\displaystyle\alpha=2^{4}\left\{(k+m)^{2}-(l+n)^{2}\right\}\left\{(k-m)^{2}-(l-n)^{2}\right\}\in 2^{12}\mathbb{Z}.

Finally, we prove (6). If (k,l,m,n)(0,0,1,1)(mod2)(k,l,m,n)\equiv(0,0,1,1)\pmod{2}, then (k+m)2(l+n)2(km)2(ln)20(mod8)(k+m)^{2}-(l+n)^{2}\equiv(k-m)^{2}-(l-n)^{2}\equiv 0\pmod{8}. Therefore,

α=24{(k+m)2(l+n)2}{(km)2(ln)2}210.\displaystyle\alpha=2^{4}\left\{(k+m)^{2}-(l+n)^{2}\right\}\left\{(k-m)^{2}-(l-n)^{2}\right\}\in 2^{10}\mathbb{Z}.

Lemma 3.1 completes the proof of (6). ∎

Lemma 3.4.

For any k,l,m,nk,l,m,n\in\mathbb{Z}, let α:=D2(2k,2l+1,2m,2n+1)\alpha:=D_{2}(2k,2l+1,2m,2n+1). The following hold:

  1. (1)(1)

    If km0(mod2)k\equiv m\equiv 0\pmod{2} and (I) hold, then α{26(8a1)(4b1)a,b}\alpha\in\left\{2^{6}(8a-1)(4b-1)\mid a,b\in\mathbb{Z}\right\};

  2. (2)(2)

    If km0(mod2)k\equiv m\equiv 0\pmod{2} and (II) hold, then α{26(8a+3)(4b+1)a,b}\alpha\in\left\{2^{6}(8a+3)(4b+1)\mid a,b\in\mathbb{Z}\right\};

  3. (3)(3)

    If km1(mod2)k\equiv m\equiv 1\pmod{2} and (I) hold, then α{26(8a+3)(4b1)a,b}\alpha\in\left\{2^{6}(8a+3)(4b-1)\mid a,b\in\mathbb{Z}\right\};

  4. (4)(4)

    If km1(mod2)k\equiv m\equiv 1\pmod{2} and (II) hold, then α{26(8a1)(4b+1)a,b}\alpha\in\left\{2^{6}(8a-1)(4b+1)\mid a,b\in\mathbb{Z}\right\},

where

  1. (I)

    k+m1ln0(mod4)k+m\equiv 1-l-n\equiv 0\pmod{4} or km2l+n0(mod4)k-m\equiv 2-l+n\equiv 0\pmod{4};

  2. (II)

    k+m1ln2(mod4)k+m\equiv 1-l-n\equiv 2\pmod{4} or km2l+n2(mod4)k-m\equiv 2-l+n\equiv 2\pmod{4}.

Proof.

Note that

α\displaystyle\alpha =D1(2k+2m,2l+2n+2)D1(2k2m,2l2n)\displaystyle=D_{1}(2k+2m,2l+2n+2)D_{1}(2k-2m,2l-2n)
=24{(k+m)2(l+n+1)2}{(km)2(ln)2}.\displaystyle=2^{4}\left\{(k+m)^{2}-(l+n+1)^{2}\right\}\left\{(k-m)^{2}-(l-n)^{2}\right\}.

If km0(mod2)k\equiv m\equiv 0\pmod{2}, then

(k+m)2(l+n+1)2{4(mod16),ifk+m1ln0(mod4),4(mod16),ifk+m1ln2(mod4),1(mod8),ifkm2l+n0(mod4),3(mod8),ifkm2l+n2(mod4),\displaystyle(k+m)^{2}-(l+n+1)^{2}\equiv\begin{cases}-4\pmod{16},&\text{if}\>\>k+m\equiv 1-l-n\equiv 0\pmod{4},\\ 4\pmod{16},&\text{if}\>\>k+m\equiv 1-l-n\equiv 2\pmod{4},\\ -1\pmod{8},&\text{if}\>\>k-m\equiv 2-l+n\equiv 0\pmod{4},\\ 3\pmod{8},&\text{if}\>\>k-m\equiv 2-l+n\equiv 2\pmod{4},\end{cases}
(km)2(ln)2{1(mod8),ifk+m1ln0(mod4),3(mod8),ifk+m1ln2(mod4),4(mod16),ifkm2l+n0(mod4),4(mod16),ifkm2l+n2(mod4).\displaystyle(k-m)^{2}-(l-n)^{2}\equiv\begin{cases}-1\pmod{8},&\text{if}\>\>k+m\equiv 1-l-n\equiv 0\pmod{4},\\ 3\pmod{8},&\text{if}\>\>k+m\equiv 1-l-n\equiv 2\pmod{4},\\ -4\pmod{16},&\text{if}\>\>k-m\equiv 2-l+n\equiv 0\pmod{4},\\ 4\pmod{16},&\text{if}\>\>k-m\equiv 2-l+n\equiv 2\pmod{4}.\end{cases}

If km1(mod2)k\equiv m\equiv 1\pmod{2}, then

(k+m)2(l+n+1)2{4(mod16),ifk+m1ln0(mod4),4(mod16),ifk+m1ln2(mod4),3(mod8),ifkm2l+n0(mod4),1(mod8),ifkm2l+n2(mod4),\displaystyle(k+m)^{2}-(l+n+1)^{2}\equiv\begin{cases}-4\pmod{16},&\text{if}\>\>k+m\equiv 1-l-n\equiv 0\pmod{4},\\ 4\pmod{16},&\text{if}\>\>k+m\equiv 1-l-n\equiv 2\pmod{4},\\ 3\pmod{8},&\text{if}\>\>k-m\equiv 2-l+n\equiv 0\pmod{4},\\ -1\pmod{8},&\text{if}\>\>k-m\equiv 2-l+n\equiv 2\pmod{4},\end{cases}
(km)2(ln)2{3(mod8),ifk+m1ln0(mod4),1(mod8),ifk+m1ln2(mod4),4(mod16),ifkm2l+n0(mod4),4(mod16),ifkm2l+n2(mod4).\displaystyle(k-m)^{2}-(l-n)^{2}\equiv\begin{cases}3\pmod{8},&\text{if}\>\>k+m\equiv 1-l-n\equiv 0\pmod{4},\\ -1\pmod{8},&\text{if}\>\>k+m\equiv 1-l-n\equiv 2\pmod{4},\\ -4\pmod{16},&\text{if}\>\>k-m\equiv 2-l+n\equiv 0\pmod{4},\\ 4\pmod{16},&\text{if}\>\>k-m\equiv 2-l+n\equiv 2\pmod{4}.\end{cases}

From the above, the lemma is proved. ∎

Note that for any integers a,ba,b and cc, where bb is odd, it holds that

abc(mod8)abc(mod8)ab\equiv c\pmod{8}\iff a\equiv bc\pmod{8}

since b21(mod8)b^{2}\equiv 1\pmod{8}. From this, we have the following.

Remark 3.5.

For any k,l,m,nk,l,m,n\in\mathbb{Z}, the following hold:

  1. (1)(1)

    (2k+2l+1)(2m+2n+1)1(mod8)kml+n(mod4)(2k+2l+1)(2m+2n+1)\equiv 1\pmod{8}\iff k-m\equiv-l+n\pmod{4};

  2. (2)(2)

    (2k+2l+1)(2m+2n+1)1(mod8)k+mln1(mod4)(2k+2l+1)(2m+2n+1)\equiv-1\pmod{8}\iff k+m\equiv-l-n-1\pmod{4};

  3. (3)(3)

    (2k+2l+1)(2m+2n+1)3(mod8)k+m1ln(mod4)(2k+2l+1)(2m+2n+1)\equiv 3\pmod{8}\iff k+m\equiv 1-l-n\pmod{4};

  4. (4)(4)

    (2k+2l+1)(2m+2n+1)3(mod8)km2l+n(mod4)(2k+2l+1)(2m+2n+1)\equiv-3\pmod{8}\iff k-m\equiv 2-l+n\pmod{4}.

Lemma 3.6.

For any k,l,m,nk,l,m,n\in\mathbb{Z}, the following hold:

  1. (1)(1)

    D2(2k,2l,2m,2n){24odd,k+ml+n(mod2),28odd210,k+ml+n(mod2);D_{2}(2k,2l,2m,2n)\in\begin{cases}2^{4}\mathbb{Z}_{\rm odd},&k+m\not\equiv l+n\pmod{2},\\ 2^{8}\mathbb{Z}_{\rm odd}\cup 2^{10}\mathbb{Z},&k+m\equiv l+n\pmod{2};\end{cases}

  2. (2)(2)

    D2(2k+1,2l+1,2m+1,2n+1){24odd,k+ml+n(mod2),29,k+ml+n(mod2);D_{2}(2k+1,2l+1,2m+1,2n+1)\in\begin{cases}2^{4}\mathbb{Z}_{\rm odd},&k+m\not\equiv l+n\pmod{2},\\ 2^{9}\mathbb{Z},&k+m\equiv l+n\pmod{2};\end{cases}

  3. (3)(3)

    D2(2k,2l+1,2m,2n+1)D_{2}(2k,2l+1,2m,2n+1)
    {27,km(mod2),28,km(mod2)and(2k+2l+1)(2m+2n+1)±1(mod8),26odd,km(mod2)and(2k+2l+1)(2m+2n+1)±3(mod8).\in\begin{cases}2^{7}\mathbb{Z},&k\not\equiv m\pmod{2},\\ 2^{8}\mathbb{Z},&k\equiv m\pmod{2}\>\>\text{and}\>\>(2k+2l+1)(2m+2n+1)\equiv\pm 1\pmod{8},\\ 2^{6}\mathbb{Z}_{\rm odd},&k\equiv m\pmod{2}\>\>\text{and}\>\>(2k+2l+1)(2m+2n+1)\equiv\pm 3\pmod{8}.\end{cases}

Proof.

From Lemma 3.3, it follows that (1) holds. Also, we obtain (2) from

D2(2k+1,2l+1,2m+1,2n+1)\displaystyle D_{2}(2k+1,2l+1,2m+1,2n+1) =D1(2k+2m+2,2l+2n+2)D1(2k2m,2l2n)\displaystyle=D_{1}(2k+2m+2,2l+2n+2)D_{1}(2k-2m,2l-2n)
=24{(k+m+1)2(l+n+1)2}{(km)2(ln)2}.\displaystyle=2^{4}\left\{(k+m+1)^{2}-(l+n+1)^{2}\right\}\left\{(k-m)^{2}-(l-n)^{2}\right\}.

Below, we prove (3). Let α:=D2(2k,2l+1,2m,2n+1)\alpha:=D_{2}(2k,2l+1,2m,2n+1). Note that

α\displaystyle\alpha =D1(2k+2m,2l+2n+2)D1(2k2m,2l2n)\displaystyle=D_{1}(2k+2m,2l+2n+2)D_{1}(2k-2m,2l-2n)
=24{(k+m)2(l+n+1)2}{(km)2(ln)2}.\displaystyle=2^{4}\left\{(k+m)^{2}-(l+n+1)^{2}\right\}\left\{(k-m)^{2}-(l-n)^{2}\right\}.

If km(mod2)k\not\equiv m\pmod{2}, then α27\alpha\in 2^{7}\mathbb{Z}. If km(mod2)k\equiv m\pmod{2} and (2k+2l+1)(2m+2n+1)1(mod8)(2k+2l+1)(2m+2n+1)\equiv 1\pmod{8}, then (km)2(ln)224(k-m)^{2}-(l-n)^{2}\in 2^{4}\mathbb{Z} from Remark 3.5 (1)(1). Thus, α28\alpha\in 2^{8}\mathbb{Z}. If km(mod2)k\equiv m\pmod{2} and (2k+2l+1)(2m+2n+1)1(mod8)(2k+2l+1)(2m+2n+1)\equiv-1\pmod{8}, then (k+m)2(l+n+1)224(k+m)^{2}-(l+n+1)^{2}\in 2^{4}\mathbb{Z} from Remark 3.5 (2)(2). Thus, α28\alpha\in 2^{8}\mathbb{Z}. The cases of km(mod2)k\equiv m\pmod{2} and (2k+2l+1)(2m+2n+1)±3(mod8)(2k+2l+1)(2m+2n+1)\equiv\pm 3\pmod{8} are proved from Lemma 3.4 and Remark 3.5 (3)(3) and (4)(4). ∎

4. Impossible values

In this section, we consider impossible values.

Lemma 4.1.

We have S(C24)odd{16m+1m}S\left({\rm C}_{2}^{4}\right)\cap\mathbb{Z}_{\rm odd}\subset\left\{16m+1\mid m\in\mathbb{Z}\right\}.

Proof.

Let D4(a0,,a15)=D2(b0,b1,b2,b3)D2(c0,c1,c2,c3)D2(d0,d1,d2,d3)D2(e0,e1,e2,e3)D_{4}(a_{0},\ldots,a_{15})=D_{2}(b_{0},b_{1},b_{2},b_{3})D_{2}(c_{0},c_{1},c_{2},c_{3})D_{2}(d_{0},d_{1},d_{2},d_{3})D_{2}(e_{0},e_{1},e_{2},e_{3}) be an odd number. Then, b0+b2b1+b3(mod2)b_{0}+b_{2}\not\equiv b_{1}+b_{3}\pmod{2} holds from Lemma 2.4. We divide the proof into the following cases:

  1. (i)

    Exactly three of b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} are even;

  2. (ii)

    Exactly one of b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} is even.

First, we consider the case of (i). If (b0,b1,b2,b3)(0,0,0,1)(mod2)(b_{0},b_{1},b_{2},b_{3})\equiv(0,0,0,1)\pmod{2}, then there exist ki,li,mik_{i},l_{i},m_{i}\in\mathbb{Z} satisfying (b0,b1,b2)=(2k0,2l0,2m0)(b_{0},b_{1},b_{2})=(2k_{0},2l_{0},2m_{0}), (c0,c1,c2)=(2k1,2l1,2m1)(c_{0},c_{1},c_{2})=(2k_{1},2l_{1},2m_{1}), (d0,d1,d2)=(2k2,2l2,2m2)(d_{0},d_{1},d_{2})=(2k_{2},2l_{2},2m_{2}), (e0,e1,e2)=(2k3,2l3,2m3)(e_{0},e_{1},e_{2})=(2k_{3},2l_{3},2m_{3}) and i=03kii=03lii=03mi0(mod2)\sum_{i=0}^{3}k_{i}\equiv\sum_{i=0}^{3}l_{i}\equiv\sum_{i=0}^{3}m_{i}\equiv 0\pmod{2} from Remark 2.2. Therefore, from Lemma 3.2 (1)(1), we have

D4(a0,a1,,a15)\displaystyle D_{4}(a_{0},a_{1},\ldots,a_{15}) i=03{8(ki+li+mi)+1}8i=03(ki+li+mi)+11(mod16).\displaystyle\equiv\prod_{i=0}^{3}\left\{8(k_{i}+l_{i}+m_{i})+1\right\}\equiv 8\sum_{i=0}^{3}(k_{i}+l_{i}+m_{i})+1\equiv 1\pmod{16}.

Lemma 3.1 completes the proof for the case (i). Next, we consider the case of (ii). If (b0,b1,b2,b3)(0,1,1,1)(mod2)(b_{0},b_{1},b_{2},b_{3})\equiv(0,1,1,1)\pmod{2}, then there exist kik^{\prime}_{i}\in\mathbb{Z} satisfying b0=2k0b_{0}=2k^{\prime}_{0}, c0=2k1c_{0}=2k^{\prime}_{1}, d0=2k2d_{0}=2k^{\prime}_{2}, e0=2k3e_{0}=2k^{\prime}_{3} and i=03ki0(mod2)\sum_{i=0}^{3}k^{\prime}_{i}\equiv 0\pmod{2} from Remark 2.2. Therefore, from Lemma 3.2 (2)(2), we have

D4(a0,a1,,a15)i=03(8ki3)8i=03ki+11(mod16).\displaystyle D_{4}(a_{0},a_{1},\ldots,a_{15})\equiv\prod_{i=0}^{3}(8k^{\prime}_{i}-3)\equiv 8\sum_{i=0}^{3}k^{\prime}_{i}+1\equiv 1\pmod{16}.

Lemma 3.1 completes the proof for the case (ii). ∎

Lemma 4.2.

We have S(C24)2216odd224odd226S\left({\rm C}_{2}^{4}\right)\cap 2\mathbb{Z}\subset 2^{16}\mathbb{Z}_{\rm odd}\cup 2^{24}\mathbb{Z}_{\rm odd}\cup 2^{26}\mathbb{Z}.

Proof.

Let D4(a0,,a15)=D2(b0,b1,b2,b3)D2(c0,c1,c2,c3)D2(d0,d1,d2,d3)D2(e0,e1,e2,e3)D_{4}(a_{0},\ldots,a_{15})=D_{2}(b_{0},b_{1},b_{2},b_{3})D_{2}(c_{0},c_{1},c_{2},c_{3})D_{2}(d_{0},d_{1},d_{2},d_{3})D_{2}(e_{0},e_{1},e_{2},e_{3}) be an even number. Then, b0+b2b1+b3(mod2)b_{0}+b_{2}\equiv b_{1}+b_{3}\pmod{2} holds from Lemma 2.4. We prove the following:

  1. (i)

    If b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} are even, then D4(a0,,a15)216odd224odd226D_{4}(a_{0},\ldots,a_{15})\in 2^{16}\mathbb{Z}_{\rm odd}\cup 2^{24}\mathbb{Z}_{\rm odd}\cup 2^{26}\mathbb{Z};

  2. (ii)

    If b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} are odd, then D4(a0,,a15)216odd226D_{4}(a_{0},\ldots,a_{15})\in 2^{16}\mathbb{Z}_{\rm odd}\cup 2^{26}\mathbb{Z};

  3. (iii)

    If exactly two of b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} are even, then D4(a0,,a15)224odd226D_{4}(a_{0},\ldots,a_{15})\in 2^{24}\mathbb{Z}_{\rm odd}\cup 2^{26}\mathbb{Z}.

First, we prove (i). If b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} are even, then there exist ki,li,mi,nik_{i},l_{i},m_{i},n_{i}\in\mathbb{Z} satisfying

(b0,b1,b2,b3)\displaystyle(b_{0},b_{1},b_{2},b_{3}) =(2k0,2l0,2m0,2n0),\displaystyle=(2k_{0},2l_{0},2m_{0},2n_{0}), (c0,c1,c2,c3)=(2k1,2l1,2m1,2n1),\displaystyle(c_{0},c_{1},c_{2},c_{3})=(2k_{1},2l_{1},2m_{1},2n_{1}),
(d0,d1,d2,d3)\displaystyle(d_{0},d_{1},d_{2},d_{3}) =(2k2,2l2,2m2,2n2),\displaystyle=(2k_{2},2l_{2},2m_{2},2n_{2}), (e0,e1,e2,e3)=(2k3,2l3,2m3,2n3)\displaystyle(e_{0},e_{1},e_{2},e_{3})=(2k_{3},2l_{3},2m_{3},2n_{3})

and i=03kii=03lii=03mii=03ni0(mod2)\sum_{i=0}^{3}k_{i}\equiv\sum_{i=0}^{3}l_{i}\equiv\sum_{i=0}^{3}m_{i}\equiv\sum_{i=0}^{3}n_{i}\equiv 0\pmod{2} from Remark 2.2. Let NN denote the cardinal number of the set {i0i3,ki+mili+ni(mod2)}\left\{i\mid 0\leq i\leq 3,\>\>k_{i}+m_{i}\equiv l_{i}+n_{i}\pmod{2}\right\}. Then, N{0,2,4}N\in\{0,2,4\} holds from i=03(ki+milini)0(mod2)\sum_{i=0}^{3}(k_{i}+m_{i}-l_{i}-n_{i})\equiv 0\pmod{2}. Therefore, from Lemma 3.6 (1)(1), we obtain (i). Also, in the same way, we can prove (ii) by using Lemma 3.6 (2). Finally, we prove (iii). If (b0,b1,b2,b3)(0,1,0,1)(mod2)(b_{0},b_{1},b_{2},b_{3})\equiv(0,1,0,1)\pmod{2}, then there exist ki,mik^{\prime}_{i},m^{\prime}_{i}\in\mathbb{Z} satisfying (b0,b2)=(2k0,2m0)(b_{0},b_{2})=(2k^{\prime}_{0},2m^{\prime}_{0}), (c0,c2)=(2k1,2m1)(c_{0},c_{2})=(2k^{\prime}_{1},2m^{\prime}_{1}), (d0,d2)=(2k2,2m2)(d_{0},d_{2})=(2k^{\prime}_{2},2m^{\prime}_{2}), (e0,e2)=(2k3,2m3)(e_{0},e_{2})=(2k^{\prime}_{3},2m^{\prime}_{3}) and i=03kii=03mi0(mod2)\sum_{i=0}^{3}k^{\prime}_{i}\equiv\sum_{i=0}^{3}m^{\prime}_{i}\equiv 0\pmod{2} from Remark 2.2. Let NN^{\prime} denote the cardinal number of the set {i0i3,kimi(mod2)}\left\{i\mid 0\leq i\leq 3,\>\>k^{\prime}_{i}\equiv m^{\prime}_{i}\pmod{2}\right\}. Then, N{0,2,4}N^{\prime}\in\{0,2,4\} holds from i=03(kimi)0(mod2)\sum_{i=0}^{3}(k^{\prime}_{i}-m^{\prime}_{i})\equiv 0\pmod{2}. Therefore, from Lemma 3.6 (3)(3), we have D4(a0,,a15)224odd226D_{4}(a_{0},\ldots,a_{15})\in 2^{24}\mathbb{Z}_{\rm odd}\cup 2^{26}\mathbb{Z}. Lemma 3.1 completes the proof of (iii). ∎

Lemma 4.3.

We have S(C24)216odd{216(4m+1)m}S\left({\rm C}_{2}^{4}\right)\cap 2^{16}\mathbb{Z}_{\rm odd}\subset\left\{2^{16}(4m+1)\mid m\in\mathbb{Z}\right\}.

Proof.

Note that we have S(C23)2={28(4m+1),212mm}S\left({\rm C}_{2}^{3}\right)\cap 2\mathbb{Z}=\left\{2^{8}(4m+1),2^{12}m\mid m\in\mathbb{Z}\right\} from the description in the introduction. Let D4(a0,,a15)=D3(a0+a8,,a7+a15)D3(a0a8,,a7a15)216oddD_{4}(a_{0},\ldots,a_{15})=D_{3}(a_{0}+a_{8},\ldots,a_{7}+a_{15})D_{3}(a_{0}-a_{8},\ldots,a_{7}-a_{15})\in 2^{16}\mathbb{Z}_{\rm odd}. Then, from Lemma 2.3 (1)(1), there exist k,lk,l\in\mathbb{Z} satisfying D3(a0+a8,,a7+a15)=28(4k+1)D_{3}(a_{0}+a_{8},\ldots,a_{7}+a_{15})=2^{8}(4k+1) and D3(a0a8,,a7a15)=28(4l+1)D_{3}(a_{0}-a_{8},\ldots,a_{7}-a_{15})=2^{8}(4l+1). Therefore, we have D4(a0,,a15){216(4m+1)m}D_{4}(a_{0},\ldots,a_{15})\in\left\{2^{16}(4m+1)\mid m\in\mathbb{Z}\right\}. ∎

Let A:={(8k3)(8l+3)k,l}A:=\left\{(8k-3)(8l+3)\mid k,l\in\mathbb{Z}\right\}.

Lemma 4.4.

If m1(mod8)m\equiv-1\pmod{8} and mAm\not\in A, then 224mS(C24)2^{24}m\not\in S\left({\rm C}_{2}^{4}\right).

To prove Lemma 4.4, we use the following two lemmas.

Lemma 4.5.

If m1(mod8)m\equiv-1\pmod{8} and mAm\not\in A, then D4(a0,a1,,a15)224mD_{4}(a_{0},a_{1},\ldots,a_{15})\neq 2^{24}m for any a0,a1,,a15a_{0},a_{1},\ldots,a_{15}\in\mathbb{Z} with b0b1b2b30(mod2)b_{0}\equiv b_{1}\equiv b_{2}\equiv b_{3}\equiv 0\pmod{2}.

Proof.

We prove by contradiction. Assume that there exist a0,a1,,a15a_{0},a_{1},\ldots,a_{15}\in\mathbb{Z} with b0b1b2b30(mod2)b_{0}\equiv b_{1}\equiv b_{2}\equiv b_{3}\equiv 0\pmod{2} satisfying D4(a0,a1,,a15)=224mD_{4}(a_{0},a_{1},\ldots,a_{15})=2^{24}m. Then there exist ki,li,mi,nik_{i},l_{i},m_{i},n_{i}\in\mathbb{Z} satisfying

D2(0)\displaystyle D_{2}^{(0)} :=D2(2k0,2l0,2m0,2n0)=D2(b0,b1,b2,b3),\displaystyle:=D_{2}(2k_{0},2l_{0},2m_{0},2n_{0})=D_{2}(b_{0},b_{1},b_{2},b_{3}),
D2(1)\displaystyle D_{2}^{(1)} :=D2(2k1,2l1,2m1,2n1)=D2(c0,c1,c2,c3),\displaystyle:=D_{2}(2k_{1},2l_{1},2m_{1},2n_{1})=D_{2}(c_{0},c_{1},c_{2},c_{3}),
D2(2)\displaystyle D_{2}^{(2)} :=D2(2k2,2l2,2m2,2n2)=D2(d0,d1,d2,d3),\displaystyle:=D_{2}(2k_{2},2l_{2},2m_{2},2n_{2})=D_{2}(d_{0},d_{1},d_{2},d_{3}),
D2(3)\displaystyle D_{2}^{(3)} :=D2(2k3,2l3,2m3,2n3)=D2(e0,e1,e2,e3)\displaystyle:=D_{2}(2k_{3},2l_{3},2m_{3},2n_{3})=D_{2}(e_{0},e_{1},e_{2},e_{3})

and i=03kii=03lii=03mii=03ni0(mod2)\sum_{i=0}^{3}k_{i}\equiv\sum_{i=0}^{3}l_{i}\equiv\sum_{i=0}^{3}m_{i}\equiv\sum_{i=0}^{3}n_{i}\equiv 0\pmod{2} from Remark 2.2. Note that i=03(ki+milini)0(mod2)\sum_{i=0}^{3}(k_{i}+m_{i}-l_{i}-n_{i})\equiv 0\pmod{2}. Since D2(0)D2(1)D2(2)D2(3)224oddD_{2}^{(0)}D_{2}^{(1)}D_{2}^{(2)}D_{2}^{(3)}\in 2^{24}\mathbb{Z}_{\rm odd}, one of the following holds: (i) Three of D2(i)D_{2}^{(i)} are of type (1) or (2) in Lemma 3.3 and the other one is of type (5) in Lemma 3.3; (ii) Two of D2(i)D_{2}^{(i)} are of type (1) or (2) in Lemma 3.3 and the others are of type (3) or (4) in Lemma 3.3. In the case of (i), we have i=03(ki+milini)0(mod2)\sum_{i=0}^{3}(k_{i}+m_{i}-l_{i}-n_{i})\not\equiv 0\pmod{2}. This is a contradiction. In the case of (ii), since m1(mod8)m\equiv-1\pmod{8} and mAm\not\in A, we have

D2(α),D2(β){24(8a+1)a},D2(γ){28(4a+1)a},D2(δ){28(4a1)a},\displaystyle D_{2}^{(\alpha)},D_{2}^{(\beta)}\in\left\{2^{4}(8a+1)\mid a\in\mathbb{Z}\right\},D_{2}^{(\gamma)}\in\left\{2^{8}(4a+1)\mid a\in\mathbb{Z}\right\},D_{2}^{(\delta)}\in\left\{2^{8}(4a-1)\mid a\in\mathbb{Z}\right\},

where {α,β,γ,δ}={0,1,2,3}\left\{\alpha,\beta,\gamma,\delta\right\}=\left\{0,1,2,3\right\}. This implies that

(kα,lα,mα,nα)\displaystyle(k_{\alpha},l_{\alpha},m_{\alpha},n_{\alpha}) (0,0,0,1),(0,0,1,0),(0,1,0,0)or(1,0,0,0)(mod2),\displaystyle\equiv(0,0,0,1),\>(0,0,1,0),\>(0,1,0,0)\>\text{or}\>(1,0,0,0)\pmod{2},
(kβ,lβ,mβ,nβ)\displaystyle(k_{\beta},l_{\beta},m_{\beta},n_{\beta}) (0,0,0,1),(0,0,1,0),(0,1,0,0)or(1,0,0,0)(mod2),\displaystyle\equiv(0,0,0,1),\>(0,0,1,0),\>(0,1,0,0)\>\text{or}\>(1,0,0,0)\pmod{2},
(kγ,lγ,mγ,nγ)\displaystyle(k_{\gamma},l_{\gamma},m_{\gamma},n_{\gamma}) (0,0,0,0)(mod2),\displaystyle\equiv(0,0,0,0)\pmod{2},
(kδ,lδ,mδ,nδ)\displaystyle(k_{\delta},l_{\delta},m_{\delta},n_{\delta}) (1,1,1,1)(mod2).\displaystyle\equiv(1,1,1,1)\pmod{2}.

Therefore, at least two of i=03ki\sum_{i=0}^{3}k_{i}, i=03li\sum_{i=0}^{3}l_{i}, i=03mi\sum_{i=0}^{3}m_{i}, i=03ni\sum_{i=0}^{3}n_{i} are odd. This is a contradiction. ∎

Lemma 4.6.

If m1(mod8)m\equiv-1\pmod{8} and mAm\not\in A, then D4(a0,a1,,a15)224mD_{4}(a_{0},a_{1},\ldots,a_{15})\neq 2^{24}m for any a0,a1,,a15a_{0},a_{1},\ldots,a_{15}\in\mathbb{Z}, where exactly two of b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} are even.

Proof.

We prove by contradiction. Assume that there exist a0,a1,,a15a_{0},a_{1},\ldots,a_{15}\in\mathbb{Z}, where exactly two of b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} are even, satisfying D4(a0,a1,,a15)=224mD_{4}(a_{0},a_{1},\ldots,a_{15})=2^{24}m. If (b0,b1,b2,b3)(0,1,0,1)(mod2)(b_{0},b_{1},b_{2},b_{3})\equiv(0,1,0,1)\pmod{2}, then there exist ki,li,mi,nik_{i},l_{i},m_{i},n_{i}\in\mathbb{Z} satisfying

D2(0)\displaystyle D_{2}^{(0)} :=D2(2k0,2l0+1,2m0,2n0+1)=D2(b0,b1,b2,b3),\displaystyle:=D_{2}(2k_{0},2l_{0}+1,2m_{0},2n_{0}+1)=D_{2}(b_{0},b_{1},b_{2},b_{3}),
D2(1)\displaystyle D_{2}^{(1)} :=D2(2k1,2l1+1,2m1,2n1+1)=D2(c0,c1,c2,c3),\displaystyle:=D_{2}(2k_{1},2l_{1}+1,2m_{1},2n_{1}+1)=D_{2}(c_{0},c_{1},c_{2},c_{3}),
D2(2)\displaystyle D_{2}^{(2)} :=D2(2k2,2l2+1,2m2,2n2+1)=D2(d0,d1,d2,d3),\displaystyle:=D_{2}(2k_{2},2l_{2}+1,2m_{2},2n_{2}+1)=D_{2}(d_{0},d_{1},d_{2},d_{3}),
D2(3)\displaystyle D_{2}^{(3)} :=D2(2k3,2l3+1,2m3,2n3+1)=D2(e0,e1,e2,e3)\displaystyle:=D_{2}(2k_{3},2l_{3}+1,2m_{3},2n_{3}+1)=D_{2}(e_{0},e_{1},e_{2},e_{3})

and i=03kii=03lii=03mii=03ni0(mod2)\sum_{i=0}^{3}k_{i}\equiv\sum_{i=0}^{3}l_{i}\equiv\sum_{i=0}^{3}m_{i}\equiv\sum_{i=0}^{3}n_{i}\equiv 0\pmod{2} from Remark 2.2. Since D2(0)D2(1)D2(2)D2(3)224oddD_{2}^{(0)}D_{2}^{(1)}D_{2}^{(2)}D_{2}^{(3)}\in 2^{24}\mathbb{Z}_{\rm odd}, it follows from Lemma 3.6 (3)(3) that kimi(mod2)k_{i}\equiv m_{i}\pmod{2} and (2ki+2li+1)(2mi+2ni+1)±3(mod8)(2k_{i}+2l_{i}+1)(2m_{i}+2n_{i}+1)\equiv\pm 3\pmod{8} for any 0i30\leq i\leq 3. Then, from Remark 3.5 (3)(3) and (4)(4), we find that every D2(i)D_{2}^{(i)} is of type (1)(1) or (4)(4) in Lemma 3.4 since mAm\not\in A. In addition, since m1(mod8)m\equiv-1\pmod{8}, exactly one or three of D2(i)D_{2}^{(i)} are of type (1)(1) in Lemma 3.4. Then we have i=03ki1(mod2)\sum_{i=0}^{3}k_{i}\equiv 1\pmod{2}. This is a contradiction. Lemma 3.1 completes the proof. ∎

Proof of Lemma 4.4.

Let m1(mod8)m\equiv-1\pmod{8} and mAm\not\in A. From Lemma 2.4, it is sufficient to prove that D4(a0,,a15)224mD_{4}(a_{0},\ldots,a_{15})\neq 2^{24}m for any a0,,a15a_{0},\ldots,a_{15}\in\mathbb{Z} with b0+b2b1+b3(mod2)b_{0}+b_{2}\equiv b_{1}+b_{3}\pmod{2}. If b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} are odd, then as mentioned in the proof of Lemma 4.2, it holds that D4(a0,,a15)216odd226D_{4}(a_{0},\ldots,a_{15})\in 2^{16}\mathbb{Z}_{\rm odd}\cup 2^{26}\mathbb{Z}. Thus, D4(a0,,a15)224mD_{4}(a_{0},\ldots,a_{15})\neq 2^{24}m. If all or exactly two of b0,b1,b2,b3b_{0},b_{1},b_{2},b_{3} are even, then we have D4(a0,,a15)224mD_{4}(a_{0},\ldots,a_{15})\neq 2^{24}m from Lemmas 4.5 and 4.6. ∎

5. Possible values

In this section, we determine all possible values. Lemmas 4.14.4 imply that S(C24)S\left({\rm C}_{2}^{4}\right) does not include every integer that is not mentioned in the following lemma.

Lemma 5.1.

For any m,nm,n\in\mathbb{Z}, the following hold:

  1. (1)(1)

    16m+1S(C24)16m+1\in S\left({\rm C}_{2}^{4}\right);

  2. (2)(2)

    216(4m+1)S(C24)2^{16}(4m+1)\in S\left({\rm C}_{2}^{4}\right);

  3. (3)(3)

    224(4m+1)S(C24)2^{24}(4m+1)\in S\left({\rm C}_{2}^{4}\right);

  4. (4)(4)

    224(4m+1)(8n+3)S(C24)2^{24}(4m+1)(8n+3)\in S\left({\rm C}_{2}^{4}\right);

  5. (5)(5)

    226mS(C24)2^{26}m\in S\left({\rm C}_{2}^{4}\right).

Proof.

We obtain (1) from

D4(m+1,m,m,,m)\displaystyle D_{4}(m+1,m,m,\ldots,m) =D3(2m+1,2m,2m,,2m)D3(1,0,0,,0)\displaystyle=D_{3}(2m+1,2m,2m,\ldots,2m)D_{3}(1,0,0,\ldots,0)
=D2(4m+1,4m,4m,4m)D2(1,0,0,0)3\displaystyle=D_{2}(4m+1,4m,4m,4m)D_{2}(1,0,0,0)^{3}
=D1(8m+1,8m)D1(1,0)7\displaystyle=D_{1}(8m+1,8m)D_{1}(1,0)^{7}
=(8m+1)2(8m)2\displaystyle=(8m+1)^{2}-(8m)^{2}
=16m+1.\displaystyle=16m+1.

We obtain (2) from

D4(k+2,k,k,,k)\displaystyle D_{4}(k+2,k,k,\ldots,k) =D3(2k+2,2k,2k,,2k)D3(2,0,0,,0)\displaystyle=D_{3}(2k+2,2k,2k,\ldots,2k)D_{3}(2,0,0,\ldots,0)
=D2(4k+2,4k,4k,4k)D2(2,0,0,0)3\displaystyle=D_{2}(4k+2,4k,4k,4k)D_{2}(2,0,0,0)^{3}
=D1(8k+2,8k)D1(2,0)7\displaystyle=D_{1}(8k+2,8k)D_{1}(2,0)^{7}
={(8k+2)2(8k)2}47\displaystyle=\left\{(8k+2)^{2}-(8k)^{2}\right\}\cdot 4^{7}
=216(8k+1)\displaystyle=2^{16}(8k+1)

and

D4(1k,k,1k,k,k,k,1k,k1,1k,k,k,k1,k,k,k,k)\displaystyle D_{4}(1-k,k,1-k,k,-k,k,1-k,k-1,1-k,k,-k,k-1,-k,k,-k,k)
=D3(22k,2k,12k,2k1,2k,2k,12k,2k1)D3(0,0,1,1,0,0,1,1)\displaystyle\quad=D_{3}(2-2k,2k,1-2k,2k-1,-2k,2k,1-2k,2k-1)D_{3}(0,0,1,1,0,0,1,-1)
=D2(24k,4k,24k,4k2)D2(2,0,0,0)D2(0,0,2,0)D2(0,0,0,2)\displaystyle\quad=D_{2}(2-4k,4k,2-4k,4k-2)D_{2}(2,0,0,0)D_{2}(0,0,2,0)D_{2}(0,0,0,2)
=D1(48k,8k2)D1(0,2)2D1(2,0)3D1(2,0)D1(0,2)\displaystyle\quad=D_{1}(4-8k,8k-2)D_{1}(0,2)^{2}D_{1}(2,0)^{3}D_{1}(-2,0)D_{1}(0,-2)
={(48k)2(8k2)2}(4)344\displaystyle\quad=\left\{(4-8k)^{2}-(8k-2)^{2}\right\}(-4)^{3}\cdot 4^{4}
=216(8k3).\displaystyle\quad=2^{16}(8k-3).

We obtain (3) from

D4(m+3,m+1,m,m,,m)\displaystyle D_{4}(m+3,m+1,m,m,\ldots,m) =D3(2m+3,2m+1,2m,2m,,2m)D3(3,1,0,0,,0)\displaystyle=D_{3}(2m+3,2m+1,2m,2m,\ldots,2m)D_{3}(3,1,0,0,\ldots,0)
=D2(4m+3,4m+1,4m,4m)D2(3,1,0,0)3\displaystyle=D_{2}(4m+3,4m+1,4m,4m)D_{2}(3,1,0,0)^{3}
=D1(8m+3,8m+1)D1(3,1)7\displaystyle=D_{1}(8m+3,8m+1)D_{1}(3,1)^{7}
={(8m+3)2(8m+1)2}87\displaystyle=\left\{(8m+3)^{2}-(8m+1)^{2}\right\}\cdot 8^{7}
=224(4m+1).\displaystyle=2^{24}(4m+1).

We obtain (4) by substituting

a0=mn+2,a1=(m+n+1),a2=m+n+1,a3=nm,\displaystyle a_{0}=m-n+2,\>\>a_{1}=-(m+n+1),\>\>a_{2}=m+n+1,\>\>a_{3}=n-m,
a4=mn1,a5=(m+n),a6=m+n+1,a7=nm,\displaystyle a_{4}=m-n-1,\>\>a_{5}=-(m+n),\>\>a_{6}=m+n+1,\>\>a_{7}=n-m,
a8=mn1,a9=(m+n),a10=m+n+1,a11=nm,\displaystyle a_{8}=m-n-1,\>\>a_{9}=-(m+n),\>\>a_{10}=m+n+1,\>\>a_{11}=n-m,
a12=mn1,a13=(m+n),a14=m+n+1,a15=nm.\displaystyle a_{12}=m-n-1,\>\>a_{13}=-(m+n),\>\>a_{14}=m+n+1,\>\>a_{15}=n-m.

In fact,

D4(a0,,a15)\displaystyle D_{4}(a_{0},\ldots,a_{15}) =D2(b0,b1,b2,b3)D2(c0,c1,c2,c3)D2(d0,d1,d2,d3)D2(e0,e1,e2,e3)\displaystyle=D_{2}(b_{0},b_{1},b_{2},b_{3})D_{2}(c_{0},c_{1},c_{2},c_{3})D_{2}(d_{0},d_{1},d_{2},d_{3})D_{2}(e_{0},e_{1},e_{2},e_{3})
=D2(4m4n1,4m4n1,4m+4n+4,4n4m)D2(3,1,0,0)3\displaystyle=D_{2}(4m-4n-1,-4m-4n-1,4m+4n+4,4n-4m)D_{2}(3,-1,0,0)^{3}
=D1(8m+3,8m1)D1(8n5,8n1)D1(3,1)6\displaystyle=D_{1}(8m+3,-8m-1)D_{1}(-8n-5,-8n-1)D_{1}(3,-1)^{6}
={(8m+3)2(8m+1)2}{(8n+5)2(8n+1)2}86\displaystyle=\left\{(8m+3)^{2}-(8m+1)^{2}\right\}\left\{(8n+5)^{2}-(8n+1)^{2}\right\}\cdot 8^{6}
=224(4m+1)(8n+3).\displaystyle=2^{24}(4m+1)(8n+3).

We obtain (5) from

D4(k+1,1k,k+1,k,k1,k,k+1,k,k,k,k+1,k,k2,1k,k+1,k)\displaystyle D_{4}(k+1,1-k,k+1,-k,k-1,-k,k+1,-k,k,-k,k+1,-k,k-2,1-k,k+1,-k)
=D3(2k+1,12k,2k+2,2k,2k3,12k,2k+2,2k)D3(1,1,0,0,1,1,0,0)\displaystyle\quad=D_{3}(2k+1,1-2k,2k+2,-2k,2k-3,1-2k,2k+2,-2k)D_{3}(1,1,0,0,1,-1,0,0)
=D2(4k2,24k,4k+4,4k)D2(4,0,0,0)D2(2,0,0,0)D2(0,2,0,0)\displaystyle\quad=D_{2}(4k-2,2-4k,4k+4,-4k)D_{2}(4,0,0,0)D_{2}(2,0,0,0)D_{2}(0,2,0,0)
=D1(8k+2,28k)D1(6,2)D1(4,0)2D1(2,0)2D1(0,2)2\displaystyle\quad=D_{1}(8k+2,2-8k)D_{1}(-6,2)D_{1}(4,0)^{2}D_{1}(2,0)^{2}D_{1}(0,2)^{2}
={(8k+2)2(28k)2}3216242(4)2\displaystyle\quad=\left\{(8k+2)^{2}-(2-8k)^{2}\right\}\cdot 32\cdot 16^{2}\cdot 4^{2}\cdot(-4)^{2}
=226(2k)\displaystyle\quad=2^{26}(2k)

and

D4(k1,k+1,k+1,k+2,k+1,k+2,k+2,k,k,k,k,k,k,k,k,k)\displaystyle D_{4}(k-1,k+1,k+1,k+2,k+1,k+2,k+2,k,k,k,k,k,k,k,k,k)
=D3(2k1,2k+1,2k+1,2k+2,2k+1,2k+2,2k+2,2k)D3(1,1,1,2,1,2,2,0)\displaystyle\quad=D_{3}(2k-1,2k+1,2k+1,2k+2,2k+1,2k+2,2k+2,2k)D_{3}(-1,1,1,2,1,2,2,0)
=D2(4k,4k+3,4k+3,4k+2)D2(2,1,1,2)2D2(0,3,3,2)\displaystyle\quad=D_{2}(4k,4k+3,4k+3,4k+2)D_{2}(-2,-1,-1,2)^{2}D_{2}(0,3,3,2)
=D1(8k+3,8k+5)D1(3,1)4D1(1,3)2D1(3,5)\displaystyle\quad=D_{1}(8k+3,8k+5)D_{1}(-3,1)^{4}D_{1}(-1,-3)^{2}D_{1}(3,5)
={(8k+3)2(8k+5)2}84(8)2(16)\displaystyle\quad=\left\{(8k+3)^{2}-(8k+5)^{2}\right\}\cdot 8^{4}\cdot(-8)^{2}\cdot(-16)
=226(2k+1).\displaystyle\quad=2^{26}(2k+1).

Remark that Lemma 5.1 (4)(4) implies that 224(8m+3),224mS(C24)2^{24}(8m+3),2^{24}m^{\prime}\in S\left({\rm C}_{2}^{4}\right) for any mm\in\mathbb{Z} and mAm^{\prime}\in A. From Lemmas 4.14.4 and 5.1, Theorem 1.1 is proved.

References

  • [1] Ton Boerkoel and Christopher Pinner. Minimal group determinants and the Lind-Lehmer problem for dihedral groups. Acta Arith., 186(4):377–395, 2018.
  • [2] Dilum DeSilva and Christopher Pinner. The Lind Lehmer constant for pn\mathbb{Z}_{p}^{n}. Proc. Amer. Math. Soc., 142(6):1935–1941, 2014.
  • [3] Edward Formanek and David Sibley. The group determinant determines the group. Proc. Amer. Math. Soc., 112(3):649–656, 1991.
  • [4] Otto Hölder. Die Gruppen der Ordnungen p3p^{3}, pq2pq^{2}, pqrpqr, p4p^{4}. Math. Ann., 43(2-3):301–412, 1893.
  • [5] K.W. Johnson. Group Matrices, Group Determinants and Representation Theory: The Mathematical Legacy of Frobenius. Lecture Notes in Mathematics. Springer International Publishing, 2019.
  • [6] Norbert Kaiblinger. Progress on Olga Taussky-Todd’s circulant problem. Ramanujan J., 28(1):45–60, 2012.
  • [7] Richard Mansfield. A group determinant determines its group. Proc. Amer. Math. Soc., 116(4):939–941, 1992.
  • [8] Bishnu Paudel and Chris Pinner. Integer circulant determinants of order 15. Integers, 22:Paper No. A4, 21, 2022.
  • [9] Vincent Pigno and Christopher Pinner. The Lind-Lehmer constant for cyclic groups of order less than 892,371,480. Ramanujan J., 33(2):295–300, 2014.
  • [10] Christopher Pinner and Christopher Smyth. Integer group determinants for small groups. Ramanujan J., 51(2):421–453, 2020.
  • [11] Naoya Yamaguchi and Yuka Yamaguchi. Generalized group determinant gives a necessary and sufficient condition for a subset of a finite group to be a subgroup. Comm. Algebra, 49(4):1805–1811, 2021.
  • [12] Naoya Yamaguchi and Yuka Yamaguchi. Remark on Laquer’s theorem for circulant determinants. International Journal of Group Theory, 12(4):265–269, 2023.
  • [13] Yuka Yamaguchi and Naoya Yamaguchi. Integer circulant determinants of order 16. The Ramanujan Journal, 2022.
  • [14] J. W. A. Young. On the Determination of Groups Whose Order is a Power of a Prime. Amer. J. Math., 15(2):124–178, 1893.

Yuka Yamaguchi

Faculty of Education

University of Miyazaki

1-1 Gakuen Kibanadai-nishi

Miyazaki 889-2192

JAPAN

[email protected]

Naoya Yamaguchi

Faculty of Education

University of Miyazaki

1-1 Gakuen Kibanadai-nishi

Miyazaki 889-2192

JAPAN

[email protected]