Analysis and Mathematical Physics (2022) 12:67

https://doi.org/10.1007/s13324-022-00676-8

Singular subelliptic equations and Sobolev inequalities

on Carnot groups

Prashanta Garain1 · Alexander Ukhlov1

Received: 29 October 2021 / Revised: 26 February 2022 / Accepted: 3 March 2022 /

Published online: 2 April 2022
© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2022

Abstract
In this article we study singular subelliptic p-Laplace equations and best constants
in Sobolev inequalities on Carnot groups. We prove solvability of these subelliptic
p-Laplace equations and existence of the minimizer of the corresponding variational
problem. It leads to existence of the best constant in the corresponding (q, p)-Sobolev
inequality, 0 < q < 1, 1 < p < ν.

Keywords Subelliptic operators · Carnot groups · Singular problem · Sobolev

inequality

Mathematics Subject Classification 35H20 · 22E30 · 46E35

1 Introduction

In this article, we investigate the following singular Dirichlet boundary value problem
for the subelliptic p-Laplace equation

− divH (|∇H u| p−2 ∇H u) = f (x)u −δ in , u > 0 in , u = 0 on ∂, (1.1)

where  is a bounded domain on the Carnot group G.

We assume that 1 < p < ν, 0 < δ < 1, where ν is the homogeneous dimension
of G and f ∈ L m () \ {0} is nonnegative, where m ≥ 1 to be made precise below.
More precisely, we establish existence, uniqueness and boundedness of weak solutions

B Alexander Ukhlov
ukhlov@math.bgu.ac.il
Prashanta Garain
pgarain92@gmail.com

1 Department of Mathematics, Ben-Gurion University of the Negev, PO Box 653, 8410501 Beer Sheva,
Israel

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 67 Page 2 of 18 P. Garain, A. Ukhlov

(Definition 3.1) for the problem (1.1). Further, we observe that such solutions are
associated to the following minimizing problem given by
  
μ() := inf |∇H u| p d x : |u|1−δ f d x = 1 . (1.2)
1, p
u∈W0 ()  

As a consequence, we obtain the following (q, p)-Sobolev inequality, 0 < q =

1 − δ < 1, 1 < p < ν:
  1  1
1−δ p
1, p
S |u| 1−δ
f dx ≤ |∇H u| d x
p
, ∀ u ∈ W0 (), (1.3)
 

for some constant S > 0.

By singularity, we mean the nonlinearity on the right hand side of (1.1) blow up near
the origin. Singular problems has been thoroughly studied over the last three decade
and there is a colossal amount of literature in this concern. Most of these results are
investigated in the Euclidean case and recently, some works has been done in the
Riemannian manifolds as well.
To motivate our present study, let us state some known results related to our question.
In the Euclidean case, Crandall et al. [11] proved existence of a unique positive classical
solution for the singular Laplace equation −u = u −δ , δ > 0 in a bounded smooth
domain  ⊂ R N subject to the Dirichlet boundary condition u = 0 on ∂. Lazer and
McKenna [24] established that such solutions are weak solution in W01,2 () if and
only if 0 < δ < 3. This restriction on δ has been removed by Boccardo and Orsina
1,2
[6] to investigate the existence of weak solutions in Wloc (). Indeed, they proved
existence of weak solutions in W01,2 () when 0 < δ ≤ 1 and in Wloc 1,2
() such that
δ+1
u 2 ∈ W01,2 () when δ > 1. We would like to point out that for δ > 1, the fact that
δ+1
u 2∈ W01,2 () is referred to as the Dirichlet boundary condition u = 0 on ∂. We
also refer to [2, 3] and the references therein in the semilinear context. Such results
has been further extended to the singular p-Laplace equations by Canino et al. [8], De
Cave [14], Haitao [21], Giacomoni et al. [20], Bal and Garain [5] and the references
therein.
As far as we are aware, singular problems are much less understood in the non-
Euclidean setting. In the Riemannian setting, some regularity results for singular
semilinear problems is studied by do Ó and Clemente in [15]. Wang and Wang [28]
obtained existence results for the purely singular problem and symmetry properties
of solutions to the purturbed singular problem in the Heinsenberg group for the semi-
linear case. Recently, in the quasilinear case, Garain and Kinnunen [19] established
non-existence result for the singular p-Laplace equation in a general metric measure
space which satisfies a doubling property and a Poincáre inequality.
Our main purpose in this article is to investigate singular problems in the subellip-
tic setting on Carnot groups. To this end, as mentioned at the beginning, we obtain
existence, uniqueness and boundedness of weak solutions (see Theorem 3.3) for the
problem (1.1). Moreover, we prove that such solutions are associated to the best con-
stant μ() defined in (1.2) which also gives rise to the Sobolev type inequality (1.3)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Singular subelliptic equations and Sobolev inequalities on… Page 3 of 18 67

(see Theorem 3.4). In the Euclidean case, such minimizing problems has been studied
in Anello et al. [1] for the local case, in the nonlocal case in Ercole and Pereira [16],
for the weighted and anisotropic case in Bal and Garain [4]. These are based on the
approximation technique introduced in Boccardo and Orsina [6]. We adopt this tech-
nique in the subelliptic setting to prove our main results. To the best of our knowledge,
our main results are new even for the linear case p = 2.
This article is organized as follows: In Sect. 2, we mention some preliminary results
in the subelliptic setting. In Sects. 3 and 4, we state our main results and obtain some
preliminary results respectively. Finally, in Sect. 5, we prove our main results.
Notation: We write c or C to denote a constant which may vary from line to line
or even in the same line. If c or C depends on the parameters r1 , r2 , . . . , rk , we write
c = c(r1 , r2 , . . . , rk ) or C = C(r1 , r2 , . . . , rk ) respectively. For a ∈ R, we denote by
a + = max{a, 0}, a − = min{−a, 0}. For given constants c, d, a set S and a function
u, by c ≤ u ≤ d in S, we mean c ≤ u ≤ d almost everywhere in S. If q > 1, then
the conjugate exponent of q is defined by q  = q−1 q
. The notation ·, · denotes the
standard inner product in R . N

2 Carnot groups and Sobolev spaces

Recall that a stratified homogeneous group [18], or, in another terminology, a Carnot
group [25] is a connected simply connected nilpotent Lie group G [7] whose Lie
algebra V is decomposed into the direct sum V1 ⊕ · · · ⊕ Vm of vector spaces such
that dim V1 ≥ 2, [V1 , Vi ] = Vi+1 for 1 ≤ i ≤ m − 1 and [V1 , Vm ] = {0}. Let
X 11 , . . . , X 1n 1 be left-invariant basis vector fields of V1 . Since they generate V , for
each i, 1 < i ≤ m, one can choose a basis X ik in Vi , 1 ≤ k ≤ n i = dim Vi , consisting
of commutators of order i − 1 of fields X 1k ∈ V1 . We identify elements g of G with
m
vectors x ∈ R N , N =  i=1 n i , x = (xik ), 1 ≤ i ≤ m, 1 ≤ k ≤ n i by means of
exponential map exp( xik X ik ) = g. Dilations δt defined by the formula

δt x = (t i xik )1≤i≤m, 1≤k≤n j

= (t x11 , . . . , t x1n 1 , t 2 x21 , . . . , t 2 x2n 2 , . . . , t m xm1 , . . . , t m xmn m ),

are automorphisms of G for each t > 0. Lebesgue measure d x on R N is the bi-invariant

Haar measure on G (which is generated by the Lebesgue measure m by means of the
exponential map), and d(δt x) = t ν d x, where the number ν = i=1 in i is called the
homogeneous dimension of the group G. The measure |E| of a measurable subset E
of G is defined by |E| = E d x.
Recall that a continuous map γ : [a, b] → G is called a continuous curve on G.
This continuous curve is rectifiable if

m
sup | (γ (tk ))−1 γ (tk+1 )| < ∞,
k=1

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 67 Page 4 of 18 P. Garain, A. Ukhlov

where the supremum is taken over all partitions a = t1 < t2 < · · · < tm = b of the
segment [a, b]. In [25] it was proved that any rectifiable curve is differentiable almost
everywhere and γ̇ (t) ∈ V1 : there exists measurable functions ai (t), t ∈ (a, b) such
that
n
γ̇ (t) = ai (t)X i (γ (t)) and (γ (t + τ ))−1 γ (t)ex p(γ̇ (t)τ ) = o(τ ) as τ → 0
i=1

for almost all t ∈ (a, b). The length l(γ ) of a rectifiable curve γ : [a, b] → G can be
calculated by the formula

b b n
1
1 2
l(γ ) = γ̇ (t), γ̇ (t) 2
0 dt = |ai (t)| 2
dt,
a a i=1

where ·, · 0 is the inner product on V1 . The result of [10] implies that one can connect
two arbitrary points x, y ∈ G by a rectifiable curve. The Carnot-Carathéodory distance
d(x, y) is the infimum of the lengths over all rectifiable curves with endpoints x
and y in G. The Hausdorff dimension of the metric space (G, d) coincides with the
homogeneous dimension ν of the group G.

2.1 Sobolev spaces on Carnot groups

Let G be a Carnot group with one-parameter dilatation group δt , t > 0, and a homo-
geneous norm ρ, and let E be a measurable subset of G. The Lebesgue space L p (E),
p ∈ [1, ∞], is the space of pth-power integrable functions f : E → R with the
standard norm:
 1
p
 f  L p (E) = | f (x)| d x
p
, 1 ≤ p < ∞,
E

p
and  f  L ∞ (E) = esssup E | f (x)| for p = ∞. We denote by L loc (E) the space of
functions f : E → R such that f ∈ L p (F) for each compact subset F of E.
Let  be an open set in G. The (horizontal) Sobolev space W 1, p (), 1 ≤ p ≤ ∞,
consists of the functions f :  → R which are locally integrable in , having the
weak derivatives X 1i f along the horizontal vector fields X 1i , i = 1, . . . , n 1 , and the
finite norm

 f W 1, p () =  f  L p () + ∇H f  L p () ,

where ∇H f = (X 11 f , . . . , X 1n 1 f ) is the horizontal subgradient of f . If f ∈ W 1, p (U )

for each bounded open set U such that U ⊂  then we say that f belongs to the class
1, p
Wloc ().

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Singular subelliptic equations and Sobolev inequalities on… Page 5 of 18 67

For the rest of the article, we assume that  ⊂ G is a bounded domain. The Sobolev
space W0 () is defined to be the closure of Cc∞ () under the norm
1, p

 f W 1, p () =  f  L p () + ∇H f  L p () .

0

For the following result, refer to [17, 26, 27, 29].

1, p
Lemma 2.1 The space W0 () is a real separable and uniformly convex Banach
space.

The following embedding result follows from [13,(2.8)] and [17], [22,Theorem
8.1], see also [9,Theorem 2.3].
1, p
Lemma 2.2 Let 1 < p < ν, then the space W0 () is continuously embedded in
L q () for every 1 ≤ q ≤ p ∗ where p ∗ = ν p/(ν − p). Moreover, the embedding is
compact for every 1 ≤ q < p ∗ .
1, p
Hence, for 1 < p < ν we can consider the Sobolev space W0 () with the norm

 f  :=  f W 1, p () = ∇H f  L p () .

0

The following algebraic inequality from [12,Lemma 2.1] will be useful for us.

Lemma 2.3 Let 1 < p < ∞. Then for any a, b ∈ R N , there exists a positive constant
C = C( p) such that

|a| p−2 a − |b| p−2 b, a − b ≥ C(|a| + |b|) p−2 |a − b|2 . (2.1)

3 Singular subelliptic p-Laplace equations

The system of basis vectors X 1 , X 2 , . . . , X n of the space V1 (here and throughout we

set n 1 = n and X 1i = X i , where i = 1, . . . , n) satisfies the Hörmander’s hypoellip-
ticity condition. We study the singular problem in the geometry of the vector fields
satisfying the Hörmander’s hypoellipticity condition.
1, p
Definition 3.1 (Weak solution) We say that u ∈ W0 () is a weak solution of (1.1)
if u > 0 in  and for every ω   there exists a positive constant c(ω) > 0 such that
u ≥ c(ω) > 0 in ω and for every φ ∈ Cc1 (), we have
 
|∇H u| p−2 ∇H u∇H φ d x = f (x)u −δ φ d x. (3.1)
 

1, p
Remark 3.2 First, we claim that if u ∈ W0 () is a weak solution of (1.1), then
1, p 1, p
(3.1) holds for every φ ∈ W0 (). Let ψ ∈ W0 (), then there exists a sequence

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 67 Page 6 of 18 P. Garain, A. Ukhlov

of nonnegative functions {ψn }n∈N ⊂ Cc1 () such that 0 ≤ ψn → |ψ| strongly in
1, p
W0 () as n → ∞ and pointwise almost everywhere in . We observe that
  
f (x)u −δ ψ d x ≤ f (x)u −δ |ψ| d x ≤ lim inf f (x)u −δ ψn d x
  n→∞ 

= lim inf |∇H u| p−2 ∇H u∇H ψn d x ≤ u p−1 lim ψn  ≤ u p−1 ψ.
n→∞  n→∞
(3.2)
1, p
Let φ ∈ W0 (), then there exists a sequence {φn }n∈N ⊂ Cc1 (), which converges
1, p
to φ strongly in W0 (). Choosing ψ = φn − φ in (3.2), we obtain

lim f (x)u −δ (φn − φ) d x ≤ u p−1 lim φn − φ = 0. (3.3)
n→∞  n→∞

1, p
Again, since φn → φ strongly in W0 () as n → ∞, we have

lim |∇H u| p−2 ∇H u∇H (φn − φ) d x = 0. (3.4)
n→∞ 

Hence, using (3.3) and (3.4) in (3.1) the claim follows.

Statement of the main results: Our main results in this article are stated below.
The first one is the following existence, uniqueness and regularity result.
Theorem 3.3 Let 0 < δ < 1 < p < ν and f ∈ L m () \ {0} be nonnegative, where
p∗ 
m = ( 1−δ ) . Then the problem (1.1) admits a unique positive weak solution u δ in
p∗
then u δ ∈ L ∞ ().
1, p
W0 (). Moreover, if m > p∗ − p ,

The second main result asserts that the Sobolev inequality (1.3) holds and its best
constant μ() defined above in (1.2) is associated to the problem (1.1).
Theorem 3.4 Let 0 < δ < 1 < p < ν and f ∈ L m () \ {0} be nonnegative, where
p∗  1, p
m = ( 1−δ ) . Assume that u δ ∈ W0 () is given by Theorem 3.3. Then, we have
 
    1−δ− p
μ() := inf u∈W 1, p ()  |∇H u| d x :
p
 |u|1−δ f d x = 1 =
 |∇H u δ | d x
p 1−δ .
0

(3.5)

Further, the following Sobolev inequality

  p 
1−δ
S |v| 1−δ
f dx ≤ |∇H v| p d x, (3.6)
 

1, p
holds for every v ∈ W0 (), if and only if S ≤ μ().

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Singular subelliptic equations and Sobolev inequalities on… Page 7 of 18 67

4 Preliminary results

For n ∈ N, we investigate the following approximated problem

fn
− divH (|∇H u| p−2 ∇H u) = in , u = 0 on ∂, (4.1)
(u + + n1 )δ

p ∗ 
where f n (x) = min{ f (x), n}, for f ∈ L m ()\{0} is nonnegative, where m = ( 1−δ ),
ν
provided 0 < δ < 1 < p < ν and p ∗ = ν− p . p

Lemma 4.1 Let 1 < p < ν and g ∈ L ∞ () \ {0} be nonnegative in . Then there
1, p
exists a unique solution u ∈ W0 () of the problem

− divH (|∇H u| p−2 ∇H u) = g in , u > 0 in . (4.2)

Moreover, for every ω  , there exists a constant c(ω), satisfying u ≥ c(ω) > 0 in
ω.
1, p
Proof Existence: We define the energy functional I : W0 () → R by
 
1
I (u) := |∇H u| d x −
p
gu d x.
p  

By using that g ∈ L ∞ () and Lemma 2.2, we have

u p p−1
I (u) ≥ − C|| p g L ∞ () u, (4.3)
p

where C > 0 is the Sobolev constant. Therefore, since p > 1, we have I is coercive.
Moreover, I is weakly lower semicontinuous. Indeed, we define I = I1 + I2 , where
1, p
I1 : W0 () → R defined by

1
I1 (u) := |∇H u| p d x
p 

1, p
and I2 : W0 () → R defined by

I2 (u) := − gu d x.


Note that I1 , I2 are convex and so I is convex. Also, I is a C 1 functional. Therefore, I

1, p
is weakly lower semicontinous. As a consequence I has a minimizer, say u ∈ W0 ()
which solves the equation

− divH (|∇H u| p−2 ∇H u) = g in . (4.4)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 67 Page 8 of 18 P. Garain, A. Ukhlov

Positivity: The positivity is a consequence of the strong minimum principle for non-
negative super-solutions [9, 26].
1, p
Uniqueness: Let u, v ∈ W0 () solves the problem (4.4). Therefore,
 
|∇H u| p−2
∇H u∇H φ d x = gφ d x, (4.5)
 

and
 
|∇H v| p−2
∇H v∇H uφ d x = gφ d x (4.6)
 

holds for every φ ∈ W0 (). We choose φ = (u − v)+ and then subtracting (4.5)
1, p

with (4.6), we obtain


|∇H u| p−2 ∇H u − |∇H v| p−2 ∇H v, ∇H (u − v)+ d x = 0. (4.7)


By Lemma 2.3, we get u ≤ v in . Similarly, we get v ≤ u in . Hence the uniqueness

follows. 

1, p
Lemma 4.2 For every n ∈ N, the problem (4.1) admits a solution u n ∈ W0 () which
is positive in . Moreover, u n is unique and u n+1 ≥ u n in  for every n, and for every
ω  , there exists a constant c(ω) (independent of n) such that u n ≥ c(ω) > 0 in
ω. Further, u n  ≤ c, for some positive constant c, which is independent of n.

Proof Existence: Let n ∈ N, then by Lemma 4.1, for every h ∈ L p (), there exists
1, p
a unique u ∈ W0 () such that

fn
− divH (|∇H u| p−2 ∇H u) = in . (4.8)
(h + + n1 )δ

Therefore, we define T : L p () → L p () by T (h) = I (u) = u, where u solves

1, p
(4.8) and I : W0 () → L p () is the continuous and compact inclusion mapping
from Lemma 2.2. First we observe that T is continuous. Indeed, let {h k }k∈N ⊂ L p ()
and h ∈ L p () be such that h k → h in L p (). Suppose T (h k ) = u k and T (h) = u.
Then we claim that u k → u in L p (). By the definition of the mapping T , for every
1, p
φ ∈ W0 (), we have
 
fn
|∇H u k | p−2 ∇H u k ∇H φ d x = +
φ dx (4.9)
  (h k + n1 )δ

and
 
fn
|∇H u| p−2
∇H u X φ d x = φ d x. (4.10)
  (h + + n1 )δ

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Singular subelliptic equations and Sobolev inequalities on… Page 9 of 18 67

Let
 1 −δ 1 
h k,n = (h +
k + ) − (h + + )−δ .
n n
Then, choosing φ = u k − u in (4.9) and (4.10) and then subtracting the resulting
equations and using Lemma 2.2, we get

|∇H u k | p−2 ∇H u k − |∇H u| p−2 ∇H u, ∇H (u k − u) d x

 
 + 1 −δ + 1 −δ 
= f n (h k + ) − (h + ) (u k − u) d x ≤ n |h k,n ||u k − u| d x
 n n 
≤ nh k,n  L ( p∗ ) () u k − u L p∗ () ≤ Cnh k,n  L ( p∗ ) () u k − u,
(4.11)

where C > 0 is the Sobolev constant. Thus, using Lemma 2.3 in (4.11) we have
1
u k − u ≤ Cnh k,n  t−1
( p ∗ )
, (4.12)
L ()

where t = p if p ≥ 2 and t = 2 if p < 2. Note that upto a subsequenc, h k,n → 0

as k → ∞ pointwise almost everywhere in  and |h k,n | ≤ 2n δ+1 and thus, by the
Lebsegue dominated convergence theorem, from (4.12) upto a subsequence u k −
u → 0 as k → ∞. Since the limit is independent of the choice of the subsequence,
by Lemma 2.2, u k → u strongly in L p (). Therefore, the continuity of T follows.
To prove that T is compact, choosing u as a test function in (4.8) and by Lemma 2.2,
we get
 p−1
u ≤
p
n δ+1 u d x ≤ n δ+1 || p Cu,


where C > 0 is the Sobolev constant. Thus,

u ≤ c, (4.13)

where c > 0 is a constant independent of h. Let {h k }k∈N ⊂ L p () be a bounded

sequence, then by (4.13) we have

T (h k ) ≤ c,

where c > 0 is a constant, which is independent of h k . Thus, by the compactness of

1, p
the embedding W0 () → L p () from Lemma 2.2 we deduce that the sequence
{T (h k )}k∈N extract a subsequence which converges strongly in L p (). Thus, T is
compact.
Let us set

S := {h ∈ L p () : λT (h) = h, 0 ≤ λ ≤ 1}.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 67 Page 10 of 18 P. Garain, A. Ukhlov

Then for any h 1 , h 2 ∈ S, by (4.13) we have

h 1 − h 2  L p () = λT (h 1 ) − T (h 2 ) L p () ≤ c,

for some constant c > 0, which is independent of h 1 , h 2 . Thus, by Schauder’s fixed

1, p
point theorem, there exists a fixed point u n ∈ W0 () such that T (u n ) = u n . As a
consequence, u n solves the problem (4.1). Moreover, by Lemma 4.1, we have u n > 0
in  and for every ω  , there exists a constant c(ω) such that

u 1 ≥ c(ω) > 0 in ω. (4.14)

Monotonicity and uniqueness: Choosing φ = (u n − u n+1 )+ as a test function in

(4.1) we have

J = |∇H u n | p−2 ∇H u n − |∇H u n+1 | p−2 ∇H u n+1 , ∇H (u n − u n+1 )+

  
fn f n+1 (4.15)
=  δ
−  δ
(u n − u n+1 )+ d x.
 un + n 1
u n+1 + n+11

Using the inequalities f n (x) ≤ f n+1 (x), we obtain

  
fn f n+1
J=   −   (u n − u n+1 )+ d x
1 δ 1 δ
 un + n u n+1 + n+1
  
1 1
≤ f n+1  δ
− δ
(u n − u n+1 )+ d x ≤ 0.
 un + n1
u n+1 + n+11

Noting this fact and using Lemma 2.3 in (4.15), it follows that u n+1 ≥ u n in .
Uniqueness follows similarly. From (4.14), we know that u 1 ≥ c(ω) > 0 for every
ω  . Hence using the monotonicity, for every ω  , we get u n ≥ c(ω) > 0 in ω,
for some positive constant c(ω) (independent of n).
Uniform boundedness: Choosing u n as a test function in (4.1) and by Lemma 2.2,
we get

   1   1−δ∗
 m p∗
u (1−δ)m
p
u n  p ≤ f u 1−δ
n d x ≤  f  L m () n dx =  f  L m () un d x
  
≤ c f  L m () u n 1−δ ,
(4.16)

for some constant c > 0, independent of n. Therefore, we have u n  ≤ c, for some

positive constant c (independent of n). 


Remark 4.3 As a consequence of Lemma 4.2 we define the pointwise limit u δ of u n

in . Note that by the monotonicity of u n in Lemma 4.2 we have u δ ≥ u n in  for
every n ∈ N.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Singular subelliptic equations and Sobolev inequalities on… Page 11 of 18 67

∗
Lemma 4.4 Suppose that f ∈ L q () \ {0} is nonnegative, where q > p∗p− p . Then
u n  L ∞ () ≤ C, for some positive constant C independent of n, where {u n }n∈N is the
sequence of solutions of the problem (4.1) given by Lemma 4.2.

Proof For k ≥ 1, we define A(k) = {x ∈  : u n (x) ≥ k}. Choosing φk (x) =

(u n − k)+ as a test function in (4.1), first using Hölder’s inequality with the exponents

p ∗ , p ∗ and then, by Young’s inequality with exponents p and p  , we obtain

    1   1
fn  p∗
 p∗
p∗ p∗
φk  =
p
  φk d x ≤ f φk d x ≤ f dx φk dx
1 δ
 un + n
A(k) A(k) 
  1

∗ p∗
≤C f p dx φk  (4.17)
A(k)
 
 p

p∗ p∗
≤ φk  + C() p
f dx ,
A(k)

where C is the Sobolev constant and C() > 0 is some constant depending on  ∈
∗ 
(0, 1) but independent of n. Note that q > p∗p− p gives q > p ∗ . Therefore, fixing
 
 ∈ (0, 1) and again using Hölder’s inequality with exponents q∗ and q∗ , for
p p
some constant C > 0 which is independent of n, we obtain

p
  p   p   1

  p∗ q
p∗ p∗ q

φk  ≤ C
p
f dx ≤C q
f dx |A(k)| p∗
A(k)  (4.18)
p
  1

p∗ q

≤ C|A(k)| p∗ .

Let h > 0 be such that 1 ≤ k < h. Then, A(h) ⊂ A(k) and for any x ∈ A(h), we have
u n (x) ≥ h. So, u n (x) − k ≥ h − k in A(h). Combining these facts along with (4.18)
and again using Lemma 2.2 for some constant C > 0 (independent of n), we arrive at

  p   p
p p∗ p∗
p∗ p∗ p∗
(h − k) |A(h)|
p
≤ (u n − k) dx ≤ (u n − k) dx
A(h) A(k)
p
  1

p∗ q

≤ Cφk  ≤ C|A(k)|
p p∗ .

Thus, for some constant C > 0 (independent of n), we have

C α
|A(h)| ≤ ∗ |A(k)| ,
(h − k) p

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 67 Page 12 of 18 P. Garain, A. Ukhlov

where

p∗ p 1
α=   .
pp ∗ q∗ 
p

∗
Due to the assumption, q > p∗p− p , we have α > 1. Hence, by [23,Lemma B.1], we
have u n  L ∞ () ≤ C, for some positive constant C > 0 independent of n. 


Lemma 4.5 Let {u n }n∈N be the sequence of solutions of the problem (4.1) given by
Lemma 4.2. Then, for every n ∈ N, we have

(u n − φ) 1, p
u n  p ≤ φ p + p f n d x, ∀φ ∈ W0 (). (4.19)
 (u n + n1 )δ

Moreover, for every n ∈ N, we have

u n  ≤ u n+1 . (4.20)

1, p
Proof By Lemma 4.1, for any h ∈ W0 (), there exists a unique solution v ∈
1, p
W0 () of the problem

f n (x)
− divH (|∇H u| p−2 ∇H v) = , v > 0 in , v = 0 on ∂. (4.21)
(h + n1 )δ
+

1, p
Note that v is also a minimizer of the functional J : W0 () → R defined by

1 fn
J (φ) := φ p − φ d x.
p  (h + + n1 )δ

1, p
Thus, J (v) ≤ J (φ), for every φ ∈ W0 () and we obtain
 
1 fn 1 fn
v p − v dx ≤ φ p − φ d x. (4.22)
p  (h + + 1 δ
n)
p  (h + + n1 )δ

Setting v = h = u n (which is positive in ) in (4.22), the estimate (4.19) follows.

By Lemma 4.2, we know u n ≤ u n+1 for every n ∈ N. Thus, choosing φ = u n+1 in
(4.19) we obtain u n  ≤ u n+1 . Hence the result follows. 


Lemma 4.6 Let {u n }n∈N be the sequence of solutions of the problem (4.1) given by
Lemma 4.2 and suppose u δ is the pointwise limit of u n given by Remark 4.3. Then,
upto a subsequence

1, p
u n → u δ strongly in W0 (). (4.23)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Singular subelliptic equations and Sobolev inequalities on… Page 13 of 18 67

1, p
Moreover, u δ is a minimizer of the energy functional Jδ : W0 () → R defined by

1 1
Jδ (v) := v p − (v + )1−δ f d x. (4.24)
p 1−δ 

Proof By Remark 4.3, we know that u n ≤ u δ in  for every n ∈ N. Thus, by choosing

φ = u δ in (4.19) we get u n  ≤ u δ  for every n ∈ N. Hence, using the norm
monotonicity u n  ≤ u n+1  from (4.20), we have

lim u n  ≤ u δ . (4.25)
n→∞

1, p
Moreover, since u n u δ weakly in W0 (), we get

u δ  ≤ lim inf u n . (4.26)

n→∞

1, p
Thus from (4.25) and (4.26) along with the uniform convexity of W0 () from
Lemma 2.1, the convergence in (4.23) follows.
1, p
To prove the second result, it is enough to show that, for all v ∈ W0 (), we have

Jδ (u δ ) ≤ Jδ (v). (4.27)

To this end, for any n ∈ N and recalling that f n (x) = min{ f (x), n}, we define
1, p
Jn : W0 () → R by

1
Jn (v) := v p − Hn (v) f n d x,
p 

where
1  + 1 1−δ  1 −δ −
Hn (t) := t + − t .
1−δ n n

Then we observe that Jn is C 1 , bounded below and coercive, and therefore, Jn has a
minimizer, say vn ∈ W0 (). This gives, Jn (vn ) ≤ Jn (vn+ ), from which it follows
1, p

that vn ≥ 0 in . Noting that Jn  (vn ), φ = 0 for all φ ∈ W0 (), we conclude

1, p

that vn solves (4.1). By the uniqueness result from Lemma 4.2, we get u n = vn in .
Hence u n is a minimizer of Jn and we have

Jn (u n ) ≤ Jn (v + ), ∀ v ∈ W0 ().
1, p
(4.28)

Below we pass to the limit as n → ∞ in (4.28) to prove the claim (4.27). Indeed, by
Remark 4.3 we know that u n ≤ u δ in , which along with the Lebesgue dominated
convergence theorem gives us
 
1
lim Hn (u n ) f n d x = (u δ )1−δ f d x. (4.29)
n→∞  1−δ 

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 67 Page 14 of 18 P. Garain, A. Ukhlov

Also, by the strong convergence from (4.23), we have

lim u n  = u δ . (4.30)
n→∞

Hence, by (4.29) and (4.30), we have

lim Jn (u n ) = Jδ (u δ ). (4.31)
n→∞

Again,
 
+ 1
lim Hn (v ) f n d x = (v + )1−δ f d x, (4.32)
n→∞  1−δ 

1, p
for any v ∈ W0 (). Now, letting n → ∞ in (4.28) and then using the estimates
(4.31), (4.32) along with v +  ≤ v, the inequality (4.27) holds. Hence the result
follows. 


5 Proof of the main results

1, p
Proof of Theorem 3.3 Uniqueness: Suppose u, v ∈ W0 () are weak solutions of
(1.1). Then by Remark 3.2 choosing φ = (u − v)+ ∈ W0 () as a test function in
1, p

(3.1) we have 
+
|∇H u| p−2
∇H u∇H (u − v) d x = f u −δ (u − v)+ d x, (5.1)
 
 
|∇H u| p−2 ∇H u∇H (u − v)+ d x = f v −δ (u − v)+ d x. (5.2)
 

Subtracting (5.1) and (5.2), we have

 
|∇H u| p−2 ∇H u − |∇H v| p−2 ∇H v, ∇H (u − v)+ d x = f (u −δ − v −δ )(u − v)+ d x ≤ 0.
 

Therefore, by Lemma 2.3, we obtain u ≤ v in . In a similar way, we have v ≤ u in

. Hence the result follows.
1, p
Existence: For every n ∈ N, using Lemma 4.2, there exists u n ∈ W0 () such that
 
fn
|∇H u n | p−2 ∇H u n ∇H φ d x = φ d x, ∀ φ ∈ Cc1 (). (5.3)
  (u n + n1 )δ

Limit pass: By the strong convergence (4.23) in Lemma 4.6, upto a subsequence, we
have ∇H u n → ∇H u δ pointwise almost everywhere in  as n → ∞. Hence, for every
φ ∈ Cc1 (), we have

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Singular subelliptic equations and Sobolev inequalities on… Page 15 of 18 67

 
lim |∇H u n | p−2
∇H u n ∇H φ d x = |∇H u δ | p−2 ∇H u δ ∇H φ d x. (5.4)
n→∞  

Let supp φ = ω   and so by Lemma 4.2, there exists a constant c(ω) > 0, which
is independent of n such that u n ≥ c(ω) > 0 in ω. Thus, u δ ≥ c(ω) > 0 in ω and also
we have

fn f
δ
φ≤ φ L ∞ () ∈ L 1 ().
un c(ω)δ

By Remark 4.3, using the pointwise convergence u n → u δ almost everywhere in 

and the Lebesgue dominated convergence theorem, we have
 
fn f
lim φ dx = φ d x. (5.5)
n→∞  (u n + 1 δ
n)  u δδ

Combining the estimates (5.4) and (5.5) in (5.3), we obtain u δ is a weak solution of
(1.1).
Boundedness: Using Lemma 4.4 we have u δ ∈ L ∞ (). 


Proof of Theorem 3.4 First, we prove (3.5). Let us set

  
1, p
Sδ := v ∈ W0 () : |v| 1−δ
f dx = 1 .


Thus it is enough to obtain

p(1−δ− p)
μ() := inf v p = u δ  1−δ .
v∈Sδ

We observe that Uδ = θδ u δ ∈ Sδ , where

 − 1−δ
1

θδ = u 1−δ
δ f dx .


1, p
By Remark 3.2, choosing φ = u δ ∈ W0 () as a test function in (3.1), we have
 
|∇H u δ | p d x = u δ  p = u 1−δ
δ f d x. (5.6)
 

Therefore, we have
 
p p(1−δ− p)
Uδ  p = |∇H Uδ | p d x = θδ |∇H u δ | p d x = u δ  1−δ . (5.7)
 

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 67 Page 16 of 18 P. Garain, A. Ukhlov

p
−
Let v ∈ Sδ and define by μ = v p+δ−1 . Then by Lemma 4.6, since u δ is a minimizer
of the functional Jδ given by (4.24), we have

Jδ (u δ ) ≤ Jδ (μ|v|). (5.8)

Using (5.6), we have

 1
1 1 1 
Jδ (u δ ) = u δ  p − u 1−δ
δ f dx = − u δ  p . (5.9)
p 1−δ  p 1−δ

Again, since v ∈ Sδ , we have

  p(δ−1)
μp μ1−δ μp μ1−δ
Jδ (μ|v|) = p |v|
p − 1−δ ≤ p v
p − 1−δ = 1
p − 1
1−δ v δ+ p−1 .
(5.10)

Since v ∈ Sδ is arbitrary, using (5.9) and (5.10) in (5.8), we obtain

p(1−δ− p)
u δ  1−δ ≤ inf v p . (5.11)
v∈Sδ

Using (5.7) and (5.11), we obtain

p(1−δ− p)
Uδ  p = u δ  1−δ ≤ inf v p . (5.12)
v∈Sδ

Since Uδ ∈ Sδ , from (5.12), we obtain (3.5).

To prove the second part, let (3.6) holds. If S > μ(), then from (3.5) and (5.7)
above, we obtain

  p 
1−δ
S Uδ1−δ f dx > |∇H Uδ | p d x. (5.13)
 

1, p
Since Uδ ∈ W0 (), (5.13) violates the hypothesis (3.6). Conversely, let

S ≤ μ() = inf v p ≤ U  p ,

v∈Sδ

for all U ∈ Sδ . We observe that the claim directly follows if v = 0. So we consider

1, p
the case of v ∈ W0 () \ {0}. This gives

 − 1−δ
1

U= |v| 1−δ
f dx v ∈ Sδ .


Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Singular subelliptic equations and Sobolev inequalities on… Page 17 of 18 67

Therefore, we have

 − 1−δ
p

S≤ |v| 1−δ
f dx v p .


Hence, the result follows. 



Acknowledgements The authors thank the reviewers for careful reading of the paper and really valuable
comments.

Data availability Data sharing not applicable to this article as no datasets were generated or analysed during
the current study.

Declarations
Conflict of Interest The authors have no conflicts of interest to declare that are relevant to the content of
this article.

References
1. Anello, G., Faraci, F., Iannizzotto, A.: On a problem of Huang concerning best constants in Sobolev
embeddings. Ann. Mat. Pura Appl. 194(3), 767–779 (2015)
2. Arcoya, D., Boccardo, L.: Multiplicity of solutions for a Dirichlet problem with a singular and a
supercritical nonlinearities. Differ. Integr. Equ. 26(1–2), 119–128 (2013)
3. Arcoya, D., Moreno-Mérida, L.: Multiplicity of solutions for a Dirichlet problem with a strongly
singular nonlinearity. Nonlinear Anal. 95, 281–291 (2014)
4. Bal, K., Garain, P.: Multiplicity of solution for a quasilinear equation with singular nonlinearity,
Mediterr. J. Math. 17(3), Paper No. 91, 20 (2020)
5. Bal, K., Garain, P.: Weighted and anisotropic Sobolev inequality with extremal. Manuscripta Math.
(2021)
6. Boccardo, L., Orsina, L.: Semilinear elliptic equations with singular nonlinearities. Calc. Var. Partial
Differ. Equ. 37(3–4), 363–380 (2010)
7. Bonfiglioli, A., Lanconelli, E., Uguzzoni, F.: Stratified Lie Groups and Potential Theory for their
Sub-Laplacians. Springer Monograph in Mathematics (2007)
8. Canino, A., Sciunzi, B., Trombetta, A.: Existence and uniqueness for p-Laplace equations involving
singular nonlinearities. Nonlinear Differ. Equ. Appl. 23(2), Art. 8, 18 (2016)
9. Capogna, L., Danielli, D., Garofalo, N.: An embedding theorem and the Harnack inequality for non-
linear subelliptic equations. Commun. Partial Differ. Equ. 18, 1765–1794 (1993)
10. Chow, W.L.: Systeme von linearen partiellen differential gleichungen erster ordnung. Math. Ann. 117,
98–105 (1939)
11. Crandall, M.G., Rabinowitz, P.H., Tartar, L.: On a Dirichlet problem with a singular nonlinearity.
Commun. Partial Differ. Equ. 2(2), 193–222 (1977)
12. Damascelli, L.: Comparison theorems for some quasilinear degenerate elliptic operators and applica-
tions to symmetry and monotonicity results. Ann. Inst. H. Poincaré Anal. Non Linéaire 15, 493–516
(1998)
13. Danielli, D.: Regularity at the boundary for solutions of nonlinear subelliptic equations. Indiana Univ.
Math. J. 44, 269–286 (1995)
14. De Cave, L.M.: Nonlinear elliptic equations with singular nonlinearities. Asymptot. Anal. 84(3–4),
181–195 (2013)
15. do Ó, Clemente, R.G.: Some elliptic problems with singular nonlinearity and advection for Riemannian
manifolds. J. Math. Anal. Appl. 460(2), 582–609 (2018)
16. Ercole, G., Pereira, G.A.: Fractional Sobolev inequalities associated with singular problems. Math.
Nachr. 291(11–12), 1666–1685 (2018)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 67 Page 18 of 18 P. Garain, A. Ukhlov

17. Folland, G.B.: Subelliptic estimates and function spaces on nilpotent Lie groups. Ark. Math. 13, 161–
207 (1975)
18. Folland, G.B., Stein, E.M.: Hardy Spaces on Homogeneous Group. Princeton University Press, Prince-
ton (1982)
19. Garain, P., Kinnunen, J.: Nonexistence of variational minimizers related to a quasilinear singular
problem in metric measure spaces. Proc. Am. Math. Soc. 149(8), 3407–3416 (2021)
20. Giacomoni, J., Schindler, I., Takáč, P.: Sobolev versus Hölder local minimizers and existence of multiple
solutions for a singular quasilinear equation. Ann. Sc. Norm. Super. Pisa Cl. Sci. (5) 6(1), 117–158
(2007)
21. Haitao, Y.: Multiplicity and asymptotic behavior of positive solutions for a singular semilinear elliptic
problem. J. Differ. Equ. 189(2), 487–512 (2003)
22. Hajlasz, P., Koskela, P.: Sobolev met Poincare. Mem. Am. Math. Soc. 688 (2000)
23. Kinderlehrer, D., Stampacchia, G.: An Introduction to Variational Inequalities and Their Applications,
volume 88 of Pure and Applied Mathematics. Academic Press, Inc. [Harcourt Brace Jovanovich,
Publishers], New York-London (1980)
24. Lazer, A.C., McKenna, P.J.: On a singular nonlinear elliptic boundary-value problem. Proc. Am. Math.
Soc. 111(3), 721–730 (1991)
25. Pansu, P.: Métriques de Carnot–Carathéodory et quasiisométries des espaces symétriques de rang un.
Ann. Math. 1–60 (1989)
26. Vodop’yanov, S.K.: Weighted Sobolev spaces and the boundary behavior of solutions of degenerate
hypoelliptic equations. Sib. Math. J. 36, 278–300 (1995)
27. Vodop’yanov, S.K., Chernikov, V.M.: Sobolev spaces and hypoelliptic equations. Trudy Inst. Mat.
7–62 (1995)
28. Wang, X., Wang, Y.: Subelliptic equations with singular nonlinearities on the Heisenberg group. Bound.
Value Probl. 22, 7 (2018)
29. Xu, Ch.J.: Subelliptic variational problems. Bull. Soc. Math. France 118, 147–169 (1990)

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center
GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers
and authorised users (“Users”), for small-scale personal, non-commercial use provided that all
copyright, trade and service marks and other proprietary notices are maintained. By accessing,
sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of
use (“Terms”). For these purposes, Springer Nature considers academic use (by researchers and
students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and
conditions, a relevant site licence or a personal subscription. These Terms will prevail over any
conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription (to
the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of
the Creative Commons license used will apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may
also use these personal data internally within ResearchGate and Springer Nature and as agreed share
it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not otherwise
disclose your personal data outside the ResearchGate or the Springer Nature group of companies
unless we have your permission as detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial
use, it is important to note that Users may not:

1. use such content for the purpose of providing other users with access on a regular or large scale
basis or as a means to circumvent access control;
2. use such content where to do so would be considered a criminal or statutory offence in any
jurisdiction, or gives rise to civil liability, or is otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association
unless explicitly agreed to by Springer Nature in writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a
systematic database of Springer Nature journal content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a
product or service that creates revenue, royalties, rent or income from our content or its inclusion as
part of a paid for service or for other commercial gain. Springer Nature journal content cannot be
used for inter-library loans and librarians may not upload Springer Nature journal content on a large
scale into their, or any other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not
obligated to publish any information or content on this website and may remove it or features or
functionality at our sole discretion, at any time with or without notice. Springer Nature may revoke
this licence to you at any time and remove access to any copies of the Springer Nature journal content
which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or
guarantees to Users, either express or implied with respect to the Springer nature journal content and
all parties disclaim and waive any implied warranties or warranties imposed by law, including
merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published
by Springer Nature that may be licensed from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a
regular basis or in any other manner not expressly permitted by these Terms, please contact Springer
Nature at

onlineservice@springernature.com