still have N after trimming

[silinzhong]

I am confused why there are still "N" with quality score "#" in the read. I tried different q cutoff at 3, 10, 20 for LEADING and TRAILING and the result is still the same. Is there something wrong with the other options i used?

java -jar trimmomatic.jar -version
0.39

time java -jar trimmomatic.jar PE -phred33 -threads 20
../SRR503008_1.fastq.gz ../SRR503008_2.fastq.gz -baseout test2.fastq.gz
ILLUMINACLIP:TruSeq3-PE-2.fa:2:30:10:2:True LEADING:10 TRAILING:10 MINLEN:36

zcat test2_1P.fastq.gz | head -n 1000 | grep "N" -B1 -A3
@SRR503008.365 HWI-1KL117:157:D0LYPACXX:8:1101:4180:2142 length=75
TGGTAAATGAATAGAGGTATATGATGCGGTTAAGTATAATTAAGANGAANNTTGTTTGT
+SRR503008.365 HWI-1KL117:157:D0LYPACXX:8:1101:4180:2142 length=75
CCCDFFFFHHHHHIIJJEHIJJJJJJJJJHIIIJGIJIJJJJJIJ#0?F##7?CHGGI