You can union 2 queries together to obtain the results you want.

create table Table_1 (Global_ID int);
create table Table_2 (Global_ID int, F_ID int);
create table Table_3 (Global_ID int, F_ID int);

insert into Table_1 (Global_ID) values (1);
insert into Table_2 (Global_ID, F_ID) values (1,1111),(1,22222), (1,33333);
insert into Table_3 (Global_ID, F_ID) values (1,1111),(1,44444);

select t1.Global_Id, t2.F_ID T2_F_ID, t3.F_ID T3_F_ID
from Table_1 t1
left join Table_2 t2 on t2.Global_ID = t1.Global_ID
left join Table_3 t3 on t3.Global_ID = t1.Global_ID and t2.F_ID = t3.F_ID

union all
  
select t1.Global_Id, null, t3.F_ID
from Table_1 t1
inner join Table_3 t3 on t3.Global_ID = t1.Global_ID
-- Exclude because already handled in the first part of the union
where not exists (
  select 1
  from Table_2 t2
  where t2.Global_ID = t1.Global_ID
  and t2.F_ID = t3.F_ID
)

Returns:

DBFiddle

Note: Providing the DDL+DML as shown here makes it much easier to assist.

Here is a solution that uses a FULL OUTER JOIN to join the two data tables (Table_2 and Table_3). That join (grouped using parentheses) is then joined back with the ID table (Table_1).

Note the the parentheses after the LEFT JOIN causes the outer join between the Table_2 and Table_3 to be evaluated first. This is not a subselect.

SELECT T1.Global_Id, T2.F_ID AS T2_F_ID, T3.F_ID AS T3_F_ID
FROM Table_1 T1
LEFT JOIN (
    Table_2 T2
    FULL OUTER JOIN Table_3 T3
        ON T3.Global_ID = T2.Global_ID
        AND T3.F_ID = T2.F_ID
)
    ON COALESCE(T2.Global_ID, T3.Global_ID) = T1.Global_ID
ORDER BY
    T1.Global_Id,
    COALESCE(T2.F_ID, T3.F_ID)

The problem with the above is that it doesn't properly handle duplicate values such as those that are present for Global_ID = 005. To handle duplicates, we need to assign distinct row values and add that to our logic. This can be done with some Common Table Expressions (CTEs) and a ROW_NUMBER() window function.

;WITH Table_2N AS (
     SELECT *,
         ROW_NUMBER() OVER(PARTITION BY Global_ID, F_ID ORDER BY (SELECT NULL)) AS RowNum
     FROM Table_2
),
Table_3N AS (
     SELECT *,
         ROW_NUMBER() OVER(PARTITION BY Global_ID, F_ID ORDER BY (SELECT NULL)) AS RowNum
     FROM Table_3
)
SELECT T1.Global_Id, T2.F_ID AS T2_F_ID, T3.F_ID AS T3_F_ID
FROM Table_1 T1
LEFT JOIN (
    Table_2N T2
    FULL OUTER JOIN Table_3N T3
        ON T3.Global_ID = T2.Global_ID
        AND T3.F_ID = T2.F_ID
        AND T3.RowNum = T2.RowNum
)
    ON COALESCE(T2.Global_ID, T3.Global_ID) = T1.Global_ID
ORDER BY
    T1.Global_Id,
    COALESCE(T2.F_ID, T3.F_ID),
    COALESCE(T2.RowNum, T3.RowNum)

The ORDER BY (SELECT NULL) is just a don't care placeholder to satisfy the required syntax.

Results:
| Global_Id | T2_F_ID | T3_F_ID |
| ------------ | ------------ | ------------ |
| 001 | 1111 | 1111 |
| 001 | 222222 | null |
| 001 | 33333 | 33333 |
| 001 | 444444 | 444444 |
| 002 | 1234 | null |
| 002 | 5678 | 5678 |
| 003 | null | 1111 |
| 003 | 1212 | 1212 |
| 003 | 33333 | 33333 |
| 004 | null | null |
| 005 | null | 1111 |
| 005 | null | 222222 |
| 005 | 33333 | 33333 |
| 005 | null | 33333 |
| 005 | 444444 | 444444 |

See this db<>fiddle .

It is worth noting that the table joins could also be rearranged and rewritten as:

FROM Table_2N T2
FULL OUTER JOIN Table_3N T3
    ON T3.Global_ID = T2.Global_ID
    AND T3.F_ID = T2.F_ID
    AND T3.RowNum = T2.RowNum
RIGHT JOIN Table_1 T1
    ON T1.Global_ID = COALESCE(T2.Global_ID, T3.Global_ID)

The results are identical. Which is more readable and best conveys intent is subjective.