On the existence of linear rank-metric intersecting codes Martino Borello 1,2, Olga Polverino 3, Ferdinando Zullo 3 Abstract . Intersecting codes are a classical object in coding theory whose rank-metric analogue has recently been introduced. Although the definition formally parallels the Hamming-metric case, the structure and parameter constraints of rank-metric intersecting codes exhibit substantially different behavior. It was previously shown that a nondegenerate [ n , k , d ] q m / q [n,k,d]_{q^{m}/q} rank-metric intersecting code must satisfy 2 ​ k − 1 ≤ n ≤ 2 ​ m − 3 2k-1\leq n\leq 2m-3 , and the tightness of the upper bound was left open. Using the geometric interpretation of rank-metric codes via q q -systems, we prove that the dual subspace associated with a rank-metric intersecting code must satisfy strong evasiveness properties. This connection allows us to derive new restrictions on the parameters of such codes and to show that the bound n = 2 ​ m − 3 n=2m-3 can be attained only when k = 3 k=3 and m ≥ 6 m\geq 6 . More generally, we show that n ≤ 2 ​ m − ⌊ ( k + 4 ) / 2 ⌋ n\leq 2m-\lfloor(k+4)/2\rfloor . Moreover, we obtain a geometric characterization of these extremal codes in terms of scattered 𝔽 q \mathbb{F}_{q} -subspaces of 𝔽 q m 3 \mathbb{F}_{q^{m}}^{3} . As a consequence, the existence problem for [ 2 ​ m − 3 , 3 , d ] q m / q [2m-3,3,d]_{q^{m}/q} rank-metric intersecting codes is reduced to the existence of scattered subspaces of dimension m + 3 m+3 . Using known constructions of maximum scattered subspaces, we derive existence results when m m is even. Finally, we prove that [ 6 , 3 , 3 ] q 5 / q [6,3,3]_{q^{5}/q} rank-metric intersecting codes do not exist for any prime power q q , thus resolving an open problem posed in [ BBMS25 ] . Keywords: rank-metric code; intersecting code; scattered subspace. Mathematics Subject Classification . 51E20; 94B27; 05B35. Introduction Intersecting codes are a classical object in coding theory. In the Hamming metric, a linear code is said to be intersecting if the supports of any two nonzero codewords share at least one coordinate; [ KS83 , Mik84 ] . Since their introduction, intersecting codes have been studied extensively due to their rich combinatorial structure and their connections to several areas such as secret sharing schemes, oblivious transfer, separating systems, frameproof codes, and additive combinatorics; see e.g. [ Mas93 , BCS02 , Bla03 , Ran13 , BSS25 , PS11 , Sco24 ] . In the binary case, intersecting codes coincide with minimal codes, a class that has attracted significant attention over the past decades. In parallel, the theory of rank-metric codes has emerged as a central topic in modern coding theory, driven both by applications, most notably in network coding [ SKK08 ] and post-quantum cryptography [ BHL+22 ] , and by deep connections with finite geometry; see [ BHL+22 , PZ20 , She19a ] . In this paper, for rank-metric codes we will mean only \mathbb{F}_{q^{m}} -subspaces of \mathbb{F}_{q^{m}}^{n} equipped with the rank distance over \mathbb{F}_{q} . Rank-metric codes admit a natural geometric interpretation via q -systems, as shown in [ Ran20 ] (see also [ She19b , ABNR22 ] ). This viewpoint has proven particularly fruitful in the study of extremal families such as MRD codes, minimal rank-metric codes, and related structures (see [ ABNR22 , ZZ21 ] ). Motivated by these developments, [ BBMS25 ] has introduced and investigated the notion of intersecting codes in the rank metric . In this setting, a rank-metric code is said to be rank-metric intersecting if the rank supports of any two nonzero codewords intersect nontrivially. While this definition formally parallels the Hamming-metric case, the resulting class of codes exhibits substantially different behavior. In particular, rank-metric intersecting codes do not coincide with minimal rank-metric codes, although the two families intersect, and their existence and structure depend in a delicate way on the parameters of the code. A key insight in the rank-metric setting is the geometric characterization of intersecting codes in terms of q -systems proved in [ BBMS25 ] . More precisely, a rank-metric code is rank-metric intersecting if and only if its associated q -system is not generated by the sum of its intersections with two hyperplanes. This geometric viewpoint leads to nontrivial constraints on the parameters of rank-metric intersecting codes, provides new constructions beyond those arising from distance considerations alone, and reveals the presence of parameter ranges where the existence problem remains open. More precisely, the authors in [ BBMS25 ] showed that for a rank-metric intersecting code in \mathbb{F}_{q^{m}}^{n} of dimension k we have that 2k-1\leq n\leq 2m-3. The tightness of the lower bound was proved in the same paper, whereas the tightness of the upper bound was left as an open problem. The main goal of this paper is to investigate the tightness of the upper bound n\leq 2m-3 for rank-metric intersecting codes and to determine when this bound can actually be attained. Our approach relies on the geometric interpretation of rank-metric codes via q -systems. By analyzing the structure of the \mathbb{F}_{q} -subspace associated with a rank-metric intersecting code and its dual, we obtain new structural constraints on such codes. In particular, we show that the dual subspace associated with a rank-metric intersecting code must satisfy strong evasiveness properties. This connection with evasive subspaces allows us to derive new bounds on the parameters of these codes. As a consequence, we prove that the maximum possible length n=2m-3 can occur only in the case k=3 and m\geq 6 . More generally, we show that n\leq 2m-\lfloor(k+4)/2\rfloor, see Theorem 3.4 for a more precise statement. Moreover, we obtain a geometric characterization of these extremal codes in terms of scattered subspaces. More precisely, a nondegenerate [2m-3,k,d]_{q^{m}/q} rank-metric intersecting code exists if and only if k=3 , m\geq 6 , and the dual system associated with the code is a scattered \mathbb{F}_{q} -subspace of \mathbb{F}_{q^{m}}^{3} of dimension m+3 ; this forces the minimum distance to be m-1 . This characterization reduces the existence problem for such codes to the existence of scattered subspaces with prescribed dimension. Using known results on maximum scattered subspaces, we derive existence results when m is even. Finally, we complete the analysis by showing that [6,3,3]_{q^{5}/q} rank-metric intersecting codes do not exist for any prime power q , thus resolving an open problem posed in [ BBMS25 ] . The paper is organized as follows. In Section 1 we recall the necessary background on rank-metric codes, q -systems, and \mathbb{F}_{q} -subspaces of \mathbb{F}_{q^{m}}^{k} . In Section 2 we review the notion of rank-metric intersecting codes and the known parameter bounds. Section 3 contains our main structural results, where we connect intersecting codes with evasive subspaces and derive new bounds. In Section 4 we study the existence problem for codes of length 2m-3 , relating it to scattered subspaces. Finally, in Section 5 we analyze the case [6,3,3]_{q^{5}/q} and prove its non-existence. 1. Preliminaries 1.1. Rank-metric codes and their geometry Rank-metric codes were introduced by Delsarte [ Del78 ] in 1978 as subsets of matrices and they have been intensively investigated in recent years because of their applications; we refer to [ She19a , PZ20 ] . In this section we will be interested in rank-metric codes in \mathbb{F}_{q^{m}}^{n} . An essential concept is the rank support of an element of \mathbb{F}_{q^{m}}^{n} . Let \Gamma=(\gamma_{1},\ldots,\gamma_{m}) be an ordered \mathbb{F}_{q} -basis of \mathbb{F}_{q^{m}} . For any vector x=(x_{1},\ldots,x_{n})\in\mathbb{F}_{q^{m}}^{n} define the matrix \Gamma(x)\in\mathbb{F}_{q}^{m\times n} , where x_{j}=\sum_{i=1}^{m}\Gamma(x)_{i,j}\gamma_{i},\qquad\mbox{ for all }j\in\{1,\ldots,n\}, that is \Gamma(x) is the matrix expansion of the vector x with respect to the \mathbb{F}_{q} -basis \Gamma of \mathbb{F}_{q^{m}} . The rank support of x is defined as the row span of \Gamma(x) : \mathrm{supp}(x)=\mathrm{rowsp}(\Gamma(x))\subseteq\mathbb{F}_{q}^{n}. As shown in [ ABNR22 , Proposition 2.1] , the support of a vector is independent of \Gamma , allowing us to talk about the support of a word without reference to \Gamma . The rank (weight) w(v) of a vector v=(v_{1},\ldots,v_{n})\in\mathbb{F}_{q^{m}}^{n} is defined as w(v)=\dim(\mathrm{supp}(v)) . A (linear vector) rank-metric code \mathcal{C} is an \mathbb{F}_{q^{m}} -subspace of \mathbb{F}_{q^{m}}^{n} endowed with the rank distance, where such a distance is defined as d(x,y)=w(x-y) , where x,y\in\mathbb{F}_{q^{m}}^{n} . For details and applications we refer to [ BHL+22 , GR18 ] . Let \mathcal{C}\subseteq\mathbb{F}_{q^{m}}^{n} be a rank-metric code. We will write that \mathcal{C} is an [n,k,d]_{q^{m}/q} code (or [n,k]_{q^{m}/q} code) if k=\dim_{\mathbb{F}_{q^{m}}}(\mathcal{C}) and d is its minimum distance, that is d=\min\{d(x,y)\colon x,y\in\mathcal{C},x\neq y\} . By the classification of \mathbb{F}_{q^{m}} -linear isometries of \mathbb{F}_{q^{m}}^{n} (see [ Ber03 ] ), we say that two rank-metric codes \mathcal{C},\mathcal{C}^{\prime}\subseteq\mathbb{F}_{q^{m}}^{n} are (linearly) equivalent if and only if there exists a matrix A\in\mathrm{GL}(n,q) such that \mathcal{C}^{\prime}=\mathcal{C}A=\{vA:v\in\mathcal{C}\} . The codes we will consider are nondegenerate , i.e. those for which the columns of any generator matrix of \mathcal{C} are \mathbb{F}_{q} -linearly independent. The geometric counterpart of a rank-metric code is the so-called q -system. An [n,k,d]_{q^{m}/q} system U is an \mathbb{F}_{q} -subspace of \mathbb{F}_{q^{m}}^{k} of dimension n , such that \langle U\rangle_{\mathbb{F}_{q^{m}}}=\mathbb{F}_{q^{m}}^{k} and d=n-\max\left\{\dim_{\mathbb{F}_{q}}(U\cap H)\mid H\textnormal{ is an }\mathbb{F}_{q^{m}}\textnormal{-hyperplane of }\mathbb{F}_{q^{m}}^{k}\right\}. Moreover, two [n,k,d]_{q^{m}/q} systems U and U^{\prime} are equivalent if there exists an \mathbb{F}_{q^{m}} -isomorphism \varphi\in\mathrm{GL}(k,q^{m}) such that \varphi(U)=U^{\prime}. Thanks to the following result, we can construct a system starting from a nondegenerate rank-metric code and conversely. Theorem 1.1 ( [ Ran20 ] ) . Let \mathcal{C} be a nondegenerate [n,k,d]_{q^{m}/q} rank-metric code and let G be a generator matrix. Let U\subseteq\mathbb{F}_{q^{m}}^{k} be the \mathbb{F}_{q} -span of the columns of G . The rank weight of an element xG\in\mathcal{C} , with x\in\mathbb{F}_{q^{m}}^{k} is (1) w(xG)=n-\dim_{\mathbb{F}_{q}}(U\cap x^{\perp}), where x^{\perp}=\{y\in\mathbb{F}_{q^{m}}^{k}\colon\langle x,y\rangle=0\}. In particular, (2) d=n-\max\left\{\dim_{\mathbb{F}_{q}}(U\cap H)\colon H\mbox{ is an }\mathbb{F}_{q^{m}}\mbox{-hyperplane of }\mathbb{F}_{q^{m}}^{k}\right\}. The above result allows us to give a one-to-one correspondence between equivalence classes of nondegenerate [n,k,d]_{q^{m}/q} codes and equivalence classes of [n,k,d]_{q^{m}/q} systems, see [ Ran20 ] and also [ ABNR22 ] . The system U and the code \mathcal{C} as in Theorem 1.1 are said to be associated and a system associated with a code \mathcal{C} will usually be written as U_{\mathcal{C}} . 1.2. Properties of \mathbb{F}_{q} -subspaces in \mathbb{F}_{q^{m}}^{k} It is well-known that \mathbb{F}_{q^{m}}^{k} can be seen as an \mathbb{F}_{q^{m}} -vector space of dimension k and as an \mathbb{F}_{q} -vector space of dimension mk . The theory of linear sets is essentially based on the idea of studying \mathbb{F}_{q} -subspaces of \mathbb{F}_{q^{m}}^{k} and how they behave with respect to \mathbb{F}_{q^{m}} -subspaces of \mathbb{F}_{q^{m}}^{k} . For instance, given an \mathbb{F}_{q} -subspace U of \mathbb{F}_{q^{m}}^{k} and W an \mathbb{F}_{q^{m}} -subspace of \mathbb{F}_{q^{m}}^{k} , we define the weight of W in U as \textrm{wt}_{U}(W)=\dim_{\mathbb{F}_{q}}(W\cap U). A classical problem is the following: given the subspace U , what are the possible weights of the subspaces of a fixed dimension with respect to U ? We refer to [ Pol10 , LVdV15 ] for more details. In this section, we will rephrase some definitions and properties of linear sets in vector terms in order to fulfill the scope of the paper. We start with the definition of scattered subspaces, originally introduced by Blokhuis and Lavrauw in [ BL00 ] . An \mathbb{F}_{q} -subspace of \mathbb{F}_{q^{m}}^{k} is said to be scattered if \textrm{wt}_{U}(\langle v\rangle_{\mathbb{F}_{q^{m}}})=\dim_{\mathbb{F}_{q}}(U\cap\langle v\rangle_{\mathbb{F}_{q^{m}}})\leq 1, for any v\in\mathbb{F}_{q^{m}}^{k} . The dimension of a scattered subspace is bounded as follows. Theorem 1.2 ( [ BL00 ] ) . Let U be a scattered \mathbb{F}_{q} -subspace in \mathbb{F}_{q^{m}}^{k} , then \dim_{\mathbb{F}_{q}}(U)\leq\frac{mk}{2}. More recently, some attention has been paid to \mathbb{F}_{q} -subspaces in \mathbb{F}_{q^{m}}^{k} which are evasive subspaces . This definition extends the notion of scattered subspaces, later extended to h -scattered subspaces in [ Lun17 , SVdV20 , CMPZ21 ] ; see also [ BCMT21 , GRSZ24 ] . Definition 1.3 . Let \mathbb{V}(k,q^{m}) be an \mathbb{F}_{q^{m}} -vector space of dimension k and let U be an \mathbb{F}_{q} -subspace of \mathbb{V}(k,q^{m}) with \dim_{\mathbb{F}_{q}}(U)=n . Let h and r be positive integers such that h\in\{1,\ldots,k\} and h\leq r\leq km . We say that U is an (h,r) -evasive subspace if \dim_{\mathbb{F}_{q}}(U\cap T)\leq r for each h -dimensional \mathbb{F}_{q^{m}} -subspace T . If r=h , an (h,h) -evasive subspace is called h -scattered subspace. Furthermore, if h=1 , then a 1 -scattered subspace will be simply called a scattered subspace. By [ CMPZ21 , Theorem 2.3] , if U is an h -scattered \mathbb{F}_{q} -subspace of \mathbb{V}(k,q^{m}) with \dim_{\mathbb{F}_{q}}(U)>k , then \dim_{\mathbb{F}_{q}}(U)\leq\frac{km}{h+1}. Subspaces that reach the equality in the previous bound are called maximum h -scattered subspaces . In particular, the following bound has been proved. Theorem 1.4 (Corollary 4.9 in [ BCMT21 ] ) . Let U be an \mathbb{F}_{q} -subspace of an \mathbb{F}_{q^{m}} -vector space of dimension k . If k\textnormal{wt}_{U^{\perp^{\prime}}}(H_{1}^{\perp})+\textnormal{wt}_{U^{\perp^{\prime}}}(H_{2}^{\perp}), i.e. the intersection of U^{\perp^{\prime}} with any 2 -dimensional \mathbb{F}_{q^{m}} -subspace of \mathbb{F}_{q^{m}}^{k} is not complementary weight. ∎ Using the characterization of non- 2 -spannable subspaces obtained in Proposition 3.1 , we derive a strong restriction on the dual subspace associated with a rank-metric intersecting code. In particular, it must have an evasive property with respect to the one-dimensional \mathbb{F}_{q^{m}} -subspaces of \mathbb{F}_{q^{m}}^{k} . Theorem 3.2 . Let \mathcal{C} be a nondegenerate [n,k,d]_{q^{m}/q} rank-metric intersecting code with k\geq 3 . Then U_{\mathcal{C}}^{\perp^{\prime}} is a (1,2m-n-2) -evasive \mathbb{F}_{q} -subspace in \mathbb{F}_{q^{m}}^{k} . Proof. Let n=2m-r , we want to show that U_{\mathcal{C}}^{\perp^{\prime}} is a (1,r-2) -evasive \mathbb{F}_{q} -subspace. Observe that U_{\mathcal{C}}^{\perp^{\prime}} is an \mathbb{F}_{q} -subspace of \mathbb{F}_{q^{m}}^{k} with dimension km-2m+r . By contradiction suppose that U_{\mathcal{C}}^{\perp^{\prime}} is not a (1,r-2) -evasive subspace. Then there exists P=\langle v\rangle_{\mathbb{F}_{q^{m}}} with v\in\mathbb{F}_{q^{m}}^{k}\setminus\{0\} such that w_{U_{\mathcal{C}}^{\perp^{\prime}}}(P)=h\geq r-1 . Every 2 -dimensional \mathbb{F}_{q^{m}} -subspace of \mathbb{F}_{q^{m}}^{k} through P has weight at least h+2 , otherwise the intersection between such a 2 -dimensional \mathbb{F}_{q^{m}} -subspace and {U_{\mathcal{C}}^{\perp^{\prime}}} is a a q -system with complementary weights, a contradiction to Proposition 3.1 . This means that the number of nonzero vectors in {U_{\mathcal{C}}^{\perp^{\prime}}} is at least the number of vectors in \bigcup_{W\in\mathcal{L}}((W\cap{U_{\mathcal{C}}^{\perp^{\prime}}})\setminus P)\,\,\cup(P\cap U_{\mathcal{C}}^{\perp^{\prime}}), where \mathcal{L} is the set of all the 2 -dimensional \mathbb{F}_{q^{m}} -subspaces of \mathbb{F}_{q^{m}}^{k} through P . Therefore, we get q^{(k-2)m+r}-1\geq\frac{q^{m(k-1)}-1}{q^{m}-1}(q^{h+2}-q^{h})+q^{h}-1, from which we derive q^{m(k-2)+r-h}\geq 1+\frac{q^{m(k-1)}-1}{q^{m}-1}(q^{2}-1), and so (4) q^{m(k-1)}-q^{m(k-1)+r-h-2}+q^{m(k-2)+r-h-2}-q^{m(k-1)-2}+q^{m-2}-1\leq 0. Observing that q^{m(k-1)}-q^{m(k-1)+r-h-2}+q^{m(k-2)+r-h-2}-q^{m(k-1)-2}>0, since q^{m(k-2)-2}(q^{m+2}-q^{m+r-h}+q^{r-h}-q^{m})>0, as r-h<2 , and q^{m-2}-1\geq 0 , ( 4 ) gives a contradiction. ∎ We now exploit the evasiveness property of the dual subspace to derive constraints on the parameters of a rank-metric intersecting code. First, we derive a bound on the minimum distance on a rank-metric intersecting code, extending Theorem 2.6 . Corollary 3.3 . Let \mathcal{C} be a nondegenerate [n,k,d]_{q^{m}/q} rank-metric intersecting code with k\geq 3 . Then \max\{k,n-m+2\}\leq d\leq m. Proof. Let n=2m-r , then by Theorem 3.2 , we know that \mathrm{wt}_{U_{\mathcal{C}}^{\perp^{\prime}}}(\langle v\rangle_{\mathbb{F}_{q^{m}}})=\dim_{\mathbb{F}_{q}}(U_{\mathcal{C}}^{\perp^{\prime}}\cap\langle v\rangle_{\mathbb{F}_{q^{m}}})\leq r-2 for any v\in\mathbb{F}_{q^{m}}^{k}\setminus\{0\} . By applying Proposition 1.5 , we derive that \mathrm{wt}_{U_{\mathcal{C}}}(\langle v\rangle_{\mathbb{F}_{q^{m}}}^{\perp})=\mathrm{wt}_{U_{\mathcal{C}}^{\perp^{\prime}}}(\langle v\rangle_{\mathbb{F}_{q^{m}}})-mk+2m-r+m(k-1)=\mathrm{wt}_{U_{\mathcal{C}}^{\perp^{\prime}}}(\langle v\rangle_{\mathbb{F}_{q^{m}}})+m-r\leq m+2. Therefore, by Theorem 1.1 , we have that d\geq 2m-r-(m+2)=n-m-2. The statement now follows by Theorem 2.6 . ∎ As a consequence of Theorem 3.2 and the bound for evasive subspaces, we obtain the following bound on the parameters of \mathcal{C} . Theorem 3.4 . Let \mathcal{C} be a nondegenerate [n,k,d]_{q^{m}/q} rank-metric intersecting code with k\geq 3 and let n=2m-r . Then k\leq 2(r-1)-\frac{r^{2}-r}{m} and in particular n\leq 2m-\left\lfloor\frac{k+4}{2}\right\rfloor . Proof. Let U_{\mathcal{C}} be a system associated with \mathcal{C} . By Theorem 3.2 , we know that U_{\mathcal{C}}^{\perp^{\prime}} is a (1,r-2) -evasive subspace of dimension mk-2m+r in \mathbb{F}_{q^{m}}^{k} . Therefore, by Theorem 1.4 we have that mk-2m+r\leq\frac{mk(r-2)}{r-1}, from which we derive that (5) k\leq 2r-\frac{r^{2}}{m}-2+\frac{r}{m}. Now, by ( 5 ) we get the following lower bound on r r\geq\frac{(2m+1)-\sqrt{\Delta}}{2} where \Delta=(2m+1)^{2}-4m(k+2) . Since \Delta<(2m-k-1)^{2} , we get n=2m-r<2m-\frac{k+2}{2} and hence the assertion. ∎ Therefore, we have the following characterization for rank-metric intersecting code of length 2m-3 . Corollary 3.5 . Let \mathcal{C} be a nondegenerate [2m-3,k,d]_{q^{m}/q} with k\geq 3 . The code \mathcal{C} is rank-metric intersecting if and only if • k=3 ; • m\geq 6 ; • U_{\mathcal{C}}^{\perp^{\prime}} is a scattered subspace. Moreover, d=m-1 . Proof. The necessary condition follows from Proposition 3.2 and Theorem 3.4 . For the converse, observe that if {U_{\mathcal{C}}^{\perp^{\prime}}} is a scattered subspace of dimension m+3 in \mathbb{F}_{q^{m}}^{3} , then every two-dimensional subspace \ell\subseteq\mathbb{F}_{q^{m}}^{3} intersects {U_{\mathcal{C}}^{\perp^{\prime}}} in a scattered subspace of dimension at least 3 . In particular, the intersection \ell\cap{U_{\mathcal{C}}^{\perp^{\prime}}} is not complementary weight. By Theorem 3.2 , it follows that U_{\mathcal{C}} is not 2 -spannable. The claim now follows from Theorem 2.3 . For the last part, from Corollary 3.3 it follows that d\in\{m-1,m\} . The assertion then follows observing that if d=m then \mathcal{C} would be a one-weight code and so a rank-metric simplex code (see [ Ran20 , ABNR22 ] ), a contradiction with the length of \mathcal{C} . ∎ In the next section we will see some related existence results. 4. Some existence results In this section we investigate the existence of nondegenerate [2m-3,k,d]_{q^{m}/q} rank-metric intersecting codes with k\geq 3 . By Theorem 3.4 , such codes can exist only if k=3 and m\geq 6 . Therefore, by Corollary 3.5 , the existence problem of [2m-3,3,d]_{q^{m}/q} rank-metric intersecting codes with m\geq 6 can be translated into a geometric problem, namely the existence of a scattered \mathbb{F}_{q} -subspace of \mathbb{F}_{q^{m}}^{3} of dimension m+3 . We now recall some known results on the existence of scattered subspaces and derive the corresponding existence results for rank-metric intersecting codes. If m is even, scattered \mathbb{F}_{q} -subspaces of \mathbb{F}_{q^{m}}^{3} of dimension 3m/2 always exist, as proved in [ BGMP18 , CMPZ17 ] . Theorem 4.1 ( [ BGMP18 , CMPZ17 ] ) . Suppose that m is even. Then there exists a scattered \mathbb{F}_{q} -subspace of \mathbb{F}_{q^{m}}^{3} of dimension 3m/2 . Observe that for m\geq 6 we have m+3\leq 3m/2 . Hence a scattered subspace of dimension 3m/2 contains scattered subspaces of dimension m+3 , and therefore the existence of the former guarantees the existence of the latter. If m is odd, the situation is less understood. Only partial results are known; we refer the reader to [ PZ20 ] for a survey of the currently known constructions and bounds. As a direct consequence of Corollary 3.5 and Theorem 4.1 , we obtain the following existence result. Corollary 4.2 . Suppose that m\geq 6 is even. Then there exists a nondegenerate [2m-3,3,d]_{q^{m}/q} rank-metric intersecting code. This shows that the upper bound in Theorem 2.4 is achieved for k=3 and every even m\geq 6 . Example 4.3 . The [9,3,5]_{64/2} rank-metric code \mathcal{C} generated by G=\left(\begin{matrix}1&0&0&\alpha^{54}&\alpha^{18}&\alpha^{26}&\alpha^{32}&\alpha^{22}&\alpha^{19}\\ 0&1&0&\alpha^{59}&\alpha^{12}&\alpha^{50}&\alpha^{49}&\alpha^{57}&\alpha^{5}\\ 0&0&1&\alpha^{29}&\alpha^{56}&\alpha^{34}&\alpha^{61}&\alpha^{44}&\alpha^{54}\end{matrix}\right), where \alpha^{6}=\alpha^{4}+\alpha^{3}+\alpha+1 is a primitive element of \mathbb{F}_{64} , is a rank-metric intersecting of length 2\cdot 6-3 . Indeed, \mathcal{C} has as associated system a maximum scattered U \mathbb{F}_{2} -subspace of \mathbb{F}_{64}^{3} . Since U^{\perp^{\prime}} is a maximum scattered \mathbb{F}_{2} -subspace of \mathbb{F}_{64}^{3} , by Corollary 3.5 we have that \mathcal{C} is rank-metric intersecting. Remark 4.4 . The case where m is odd remains largely open, since the existence of scattered \mathbb{F}_{q} -subspaces of \mathbb{F}_{q^{m}}^{3} with large dimension is not fully understood. Any new construction of scattered subspaces of dimension at least m+3 would immediately provide new examples of rank-metric intersecting codes of length 2m-3 . Remark 4.5 . Let us remark that, for m=6 , we know the existence of [n,3]_{q^{6}/q} rank-metric intersecting codes for 5\leq n\leq 7 and for n=9 . We also know that for n<5 and n>9 , an [n,3]_{q^{6}/q} rank-metric intersecting code cannot exist. However, the existence of an [8,3]_{q^{6}/q} rank-metric intersecting code remains open. A natural question is whether such a code can be obtained by extending or puncturing (in the rank-metric sense) known constructions. However, this does not seem to be an easy task: an extensive search for [8,3]_{64/2} codes obtained by puncturing the code in Example 4.3 fails to produce a rank-metric intersecting one. 5. Non-existence of [6,3,3]_{q^{5}/q} rank-metric intersecting codes As already observed in [ BBMS25 ] , for k=2 or k=3 and m\leq 4 , the set of values of n for which a rank-metric intersecting code exists is known. The first open case occurs for k=3 and m=5 . In this case, by Theorem 2.5 and Theorem 3.4 , the only unknown value of n for the existence of a rank-metric intersecting code is n=6 . In [ BBMS25 ] , it is proved that a [6,3,3]_{32/2} rank-metric intersecting code does not exist, leaving the general case open (see Question 4.10). In the following, we show that a [6,3,3]_{q^{5}/q} rank-metric intersecting code does not exist for any value of q . To this aim we will need the projective description of a q -system; we will restrict to the case of the projective plane. Definition 5.1 . Let \Lambda=\mathrm{PG}(\mathbb{F}_{q^{m}}^{3},\mathbb{F}_{q^{m}})=\mathrm{PG}(2,q^{m}) . Consider an \mathbb{F}_{q} -subspace U of \mathbb{F}_{q^{m}}^{3} and define the set L_{U}=\{\langle{u}\rangle_{\mathbb{F}_{q^{m}}}\colon{u}\in U\setminus\{{0}\}\} as an \mathbb{F}_{q} - linear set of rank \dim_{\mathbb{F}_{q}}(U) . Denote by \tau_{j} the number of lines in \mathrm{PG}(2,q^{m}) intersecting L_{U} in j points. By a classical double counting argument, we obtain the following. Proposition 5.2 . Let L_{U} be an \mathbb{F}_{q} -linear set in \mathrm{PG}(2,q^{m}) , then \sum_{i=0}^{|L_{U}|}\tau_{i}=q^{2m}+q^{m}+1, \sum_{i=0}^{|L_{U}|}i\tau_{i}=|L_{U}|(q^{m}+1), \sum_{i=0}^{|L_{U}|}i(i-1)\tau_{i}=|L_{U}|(|L_{U}|-1). If L_{U} has rank n then the weight of any point is bounded by n . Denote by N_{i}(U) the number of points of \Lambda having weight i\in\{0,\ldots,n\} in L_{U} , the following relations hold: (6) |L_{U}|\leq\frac{q^{n}-1}{q-1}, (7) |L_{U}|=N_{1}(U)+\ldots+N_{n}(U), (8) N_{1}(U)+N_{2}(U)(q+1)+\ldots+N_{n}(U)(q^{n-1}+\ldots+q+1)=q^{n-1}+\ldots+q+1. To our aim, we need the following result, which is a consequence of the Main Theorem of [ DBVdV22 ] . Lemma 5.3 . Let L_{U} be an \mathbb{F}_{q} -linear set of rank 5 in \mathrm{PG}(1,q^{5}) with points of weight at most two. Then N_{2}(U)\leq\begin{cases}q^{2}+1,&\text{if }q>2,\\ 6,&\text{if }q=2.\end{cases} By means of a refined combinatorial argument on the linear set associated with a [6,3,3]_{q^{5}/q} rank-metric intersecting code, we prove that such a code does not exist. Theorem 5.4 . A [6,3,3]_{q^{5}/q} rank-metric intersecting codes does not exist for any prime power q . Proof. By Theorem 2.3 , there exists a [6,3,3]_{q^{5}/q} rank-metric intersecting code if and only if there exists a [6,3,3]_{q^{5}/q} system U which is not 2-spannable. Let U be a [6,3,3]_{q^{5}/q} system in \mathbb{F}_{q^{5}}^{3} , and suppose that U is not 2 -spannable. Since n=6 , d=3 and U is not 2-spannable, the following properties follow readily: 1) Each two-dimensional \mathbb{F}_{q^{5}} -subspace (we will call it a line) of \mathbb{F}_{q^{5}}^{3} has weight at least one and at most three in U . 2) For each one-dimensional \mathbb{F}_{q^{5}} -subspace (we will call it a point) of weight 0 in U , at most one line through that point has weight three in U . 3) Either U is a scattered subspace, or U has exactly one point of weight 2 , while all remaining points have weight at most 1 (otherwise we would have lines of weight larger than 3 ). Let L_{U} be the \mathbb{F}_{q} -linear set defined by U in the projective plane \mathrm{PG}(2,q^{5}) , and let \tau_{i} be the number of lines of \mathrm{PG}(2,q^{5}) intersecting L_{U} in i points. By Proposition 5.2 , for the integers \tau_{i} the following identities hold (9) \sum_{i=0}^{|L_{U}|}\tau_{i}=q^{10}+q^{5}+1 (10) \sum_{i=0}^{|L_{U}|}i\tau_{i}=|L_{U}|(q^{5}+1) (11) \sum_{i=0}^{|L_{U}|}i(i-1)\tau_{i}=|L_{U}|(|L_{U}|-1), From 1), 2) and 3) together with ( 7 ) and ( 8 ), we get that \tau_{i}=0 if i\neq 1,q+1,q^{2}+1,q^{2}+q+1 . Also, if U is a scattered subspace, then \tau_{q^{2}+1}=0 and |L_{U}|=q^{5}+q^{4}+q^{3}+q^{2}+q+1 , while if U has exactly one point of weight 2 , and all remaining points have weight at most 1 , then \tau_{q^{2}+1}=q^{3}+q^{2}+q+1 and |L_{U}|=q^{5}+q^{4}+q^{3}+q^{2}+1 . In this second case, from Equations ( 9 ), ( 10 ) and ( 11 ) we get \tau_{q+1}=\frac{-q^{11}-q^{10}+q^{9}+3q^{8}+4q^{7}+3q^{6}+q^{5}+2q^{4}+2q^{3}+2q^{2}+q}{q^{3}+2q^{2}+q+1}, which gives a negative number for every prime power q , a contradiction. Hence, U is a scattered subspace and from Equations ( 9 ), ( 10 ) and ( 11 ) we get (12) \tau_{1}=q^{10}-q^{8}-q^{7}-q^{6}+q^{5}+q^{3},\quad\tau_{q+1}=q^{8}+q^{7}+q^{6}-q^{4}-2q^{3}-2q^{2}-q,\quad\tau_{q^{2}+q+1}=q^{4}+q^{3}+2q^{2}+q+1. Let P be any point of L_{U} , so P=\langle u\rangle_{\mathbb{F}_{q^{5}}} where u\in U\setminus\{0\} . Consider the projection L_{U_{P}} of L_{U} from the point P in the quotient space \mathbb{F}_{q^{5}}^{3}/P , i.e. U_{P}=U+P in the quotient vector space \mathbb{F}_{q^{5}}^{3}/P . Then L_{U_{P}} is an \mathbb{F}_{q} -linear set of rank 5 in \mathrm{PG}(1,q^{5})=\mathrm{PG}(\mathbb{F}_{q^{5}}^{3}/P,\mathbb{F}_{q^{5}}) . By Lemma 5.3 , the linear set L_{U_{P}} admits at most q^{2}+1 points of weight 2 when q>2 and at most 6 when q=2 . This means that the number \tau_{q^{2}+q+1}^{P} of lines in \mathrm{PG}(2,q^{5}) of weight 3 in U passing through P is at most q^{2}+1 . Now, let \ell be a fixed line with weight 3 in U . By 2) , any other line with weight 3 in U intersects \ell in a point of L_{U} . Counting the number of lines of weight 3 in U , for q>2 we obtain \tau_{q^{2}+q+1}=1+\sum_{P\in\ell\cap L_{U}}(\tau_{q^{2}+q+1}^{P}-1)\leq q^{2}|\ell\cap L_{U}|+1=q^{2}(q^{2}+q+1)+1=q^{4}+q^{3}+q^{2}+1, which contradicts ( 12 ). A similar contradiction can be obtained when q=2 . ∎ The above theorem answers Question 4.10 of [ BBMS25 ] . 6. Conclusions In this paper we investigated the existence problem for linear rank-metric intersecting codes. Starting from the geometric characterization of intersecting codes in terms of q -systems introduced in [ BBMS25 ] , we analyzed the structure of the associated \mathbb{F}_{q} -subspaces and their duals. We first showed that the dual subspace associated with a rank-metric intersecting code must satisfy strong evasiveness properties. More precisely, if \mathcal{C} is a nondegenerate [n,k,d]_{q^{m}/q} rank-metric intersecting code with k\geq 3 , then the dual system U_{\mathcal{C}}^{\perp^{\prime}} is a (1,2m-n-2) -evasive subspace of \mathbb{F}_{q^{m}}^{k} . Using known bounds on evasive subspaces, this property allowed us to derive new restrictions on the parameters of rank-metric intersecting codes and to refine the upper bound on the length. More precisely, n\leq 2m-\left\lfloor\frac{k+4}{2}\right\rfloor. Moreover, we proved that the maximum possible length n=2m-3 can occur only in the case k=3 and m\geq 6 . We also obtained a geometric characterization of such codes: a nondegenerate [2m-3,k,d]_{q^{m}/q} rank-metric intersecting code exists if and only if k=3 , m\geq 6 , and the dual system U_{\mathcal{C}}^{\perp^{\prime}} is a scattered subspace of \mathbb{F}_{q^{m}}^{3} of dimension m+3 , and as a consequence d=m-1 . This characterization allows us to translate the existence problem for these codes into the existence problem for scattered \mathbb{F}_{q} -subspaces of \mathbb{F}_{q^{m}}^{3} with prescribed dimension. Using known constructions of maximum scattered subspaces, we showed that rank-metric intersecting codes with parameters [2m-3,3,d]_{q^{m}/q} exist for every even m\geq 6 , proving that the upper bound n\leq 2m-3 is tight in this case. The above argument can also be extended by considering maximum h -scattered subspaces. However, the corresponding rank-metric intersecting codes rely on the existence of such maximum h -scattered subspaces for certain parameter choices, which are currently unknown. Finally, we completed the analysis of the boundary cases by proving that [6,3,3]_{q^{5}/q} rank-metric intersecting codes do not exist for any prime power q . This result provides a complete answer to an open problem posed in [ BBMS25 ] . Several questions remain open. In particular, when m is odd the existence of scattered subspaces of dimension at least m+3 in \mathbb{F}_{q^{m}}^{3} is not fully understood, and therefore the existence of rank-metric intersecting codes of length 2m-3 in this case remains open. Moreover, as highlighted in Remark 4.5 , the existence of [2m-4,3]_{q^{m}/q} rank-metric intersecting codes is open also in the case where m is even. A more refined study on the geometric point of view on the Delsarte dual of an intersecting rank-metric code may be fruitful; see e.g. [ MNT23 , BPZ25 ] .