diff --git a/isatools/isatab/dump/core.py b/isatools/isatab/dump/core.py
index 12b94712..1ea8036e 100644
--- a/isatools/isatab/dump/core.py
+++ b/isatools/isatab/dump/core.py
@@ -41,7 +41,7 @@ def dump(isa_obj, output_path,
         raise NameError('Investigation file must match pattern i_*.txt, got {}'.format(i_file_name))
 
     if path.exists(output_path):
-        fp = open(path.join(output_path, i_file_name), 'w', encoding='utf-8')
+        fp = open(path.join(output_path, i_file_name), 'wb', encoding='utf-8')
     else:
         log.debug('output_path=', i_file_name)
         raise FileNotFoundError("Can't find " + output_path)
diff --git a/isatools/isatab/dump/write.py b/isatools/isatab/dump/write.py
index d7968ead..0fcf3397 100644
--- a/isatools/isatab/dump/write.py
+++ b/isatools/isatab/dump/write.py
@@ -221,7 +221,7 @@ def flatten(current_list):
         DF = DF.replace('', nan)
         DF = DF.dropna(axis=1, how='all')
 
-        with open(path.join(output_dir, study_obj.filename), 'w') as out_fp:
+        with open(path.join(output_dir, study_obj.filename), 'wb') as out_fp:
             DF.to_csv(
                 path_or_buf=out_fp, index=False, sep='\t', encoding='utf-8')
 
@@ -477,7 +477,7 @@ def pbar(x):
             DF = DF.dropna(axis=1, how='all')
 
             with open(path.join(
-                    output_dir, assay_obj.filename), 'w') as out_fp:
+                    output_dir, assay_obj.filename), 'wb') as out_fp:
                 DF.to_csv(path_or_buf=out_fp, index=False, sep='\t',
                           encoding='utf-8')
 
diff --git a/isatools/isatab/load/core.py b/isatools/isatab/load/core.py
index 715c294f..75a605be 100644
--- a/isatools/isatab/load/core.py
+++ b/isatools/isatab/load/core.py
@@ -387,7 +387,7 @@ def merge_study_with_assay_tables(study_file_path, assay_file_path, target_file_
     log.info("Merging DataFrames...")
     merged_DF = merge(study_DF, assay_DF, on='Sample Name')
     log.info("Writing merged DataFrame to file %s", target_file_path)
-    with open(target_file_path, 'w', encoding='utf-8') as fp:
+    with open(target_file_path, 'wb', encoding='utf-8') as fp:
         merged_DF.to_csv(fp, sep='\t', index=False, header=study_DF.isatab_header + assay_DF.isatab_header[1:])
 
 
diff --git a/isatools/magetab.py b/isatools/magetab.py
index 0f3d0d6f..cd2aa60f 100644
--- a/isatools/magetab.py
+++ b/isatools/magetab.py
@@ -362,7 +362,7 @@ def write_idf_file(inv_obj, output_path):
     idf_df = idf_df.replace('', np.nan)
     with open(os.path.join(output_path, "{}.idf.txt".format(
             investigation.identifier if investigation.identifier != ""
-            else investigation.filename[2:-3])), "w",
+            else investigation.filename[2:-3])), "wb",
             encoding='utf-8') as idf_fp:
         idf_df.to_csv(
             path_or_buf=idf_fp,
diff --git a/isatools/utils.py b/isatools/utils.py
index 3fb21247..049c7bcb 100644
--- a/isatools/utils.py
+++ b/isatools/utils.py
@@ -733,7 +733,7 @@ def replace_factor_with_source_characteristic(self, factor_name):
         table_file_df.columns = self.clean_isatab_field_names(
             field_names_modified)
 
-        with open(self.path, 'w') as out_fp:
+        with open(self.path, 'wb') as out_fp:
             table_file_df.to_csv(path_or_buf=out_fp, index=False, sep='\t',
                                  encoding='utf-8')
 
@@ -856,7 +856,7 @@ def replace_factor_with_protocol_parameter_value(
 
         with open(os.path.join(
                 os.path.dirname(self.path), '{s_filename}.fix'.format(
-                    s_filename=os.path.basename(self.path))), 'w') as out_fp:
+                    s_filename=os.path.basename(self.path))), 'wb') as out_fp:
             table_file_df.to_csv(path_or_buf=out_fp, index=False, sep='\t',
                                  encoding='utf-8')